“Nothing truly great ever came from a comfort zone.”

– Anonymous  –

When studying fraud in its various forms and methods you quickly learn that fraudsters, by nature, don’t let themselves get comfortable. They can’t afford to remain in their comfort zone, as they know they’re being pursued: they’re constantly on the move, looking for the next gap or loophole to exploit.

Advertisers, however, have a tendency to get comfortable. Whether knowingly or unknowingly, many advertisers find themselves putting too much trust in outdated methods, which at the time of implementation may have indeed provided some level of improved performance, protection or assurance. 

The Evolution of Mobile Advertising Models

When mobile app advertising first hit the market, advertisers were introduced to the Cost Per Install (CPI) model – rewarding publishers with a payment per each install they managed to generate.

The advertisers’ approach at the time was, “get me the install and I’ll take care of retention”. This worked well for a while.

As fraudsters started targeting CPI campaigns and generating fake users or hijacking real ones, user quality started deteriorating and the focus shifted towards LTV-centric campaigns, birthing in-app event measurement and the CPA model – Cost Per Action.

CPA-based campaigns were the next step in the evolution of app promotion, with some advertisers even going as far as abandoning CPI promotions altogether, focusing solely on in-app events. 

By keeping track on in-app events an advertiser could differentiate quality users from less quality ones by measuring engagement, progress and in-app purchases. 

Advertisers were now optimizing not just by their media partner’s CTR but also by their ability to deliver quality users.

As user quality started improving the common belief was that running with cost per action campaigns actually protected advertisers from fraud.

CPA campaigns were believed (and rightfully so) to produce higher-quality, more engaged users, especially when compared to users coming from CPI-only campaigns. 

Well… sorry to be the one to burst that bubble, but this couldn’t be further from the truth.

Fraudsters Can Have a Field Day With CPA-Based Campaigns

Going back to the way fraudsters operate, this blind faith in the power of CPA as a fraud preventing method is exactly where fraudsters want you to be, as they’ve already caught up.

Utilizing sophisticated bots for their activity, fraudsters have managed to go far beyond the install, faking in-app events and purchases deep within the app.

In an extensive fraud data study conducted this year, we found that the average fraudulent install generated around 0.9 in-app events on average in Q4 of 2018. During Q2 of 2019, this figure tripled to 2.7 events per every fraudulent install, showing a clear path of where fraudsters are focusing their efforts.

Gaming apps, which rely heavily on measurable in-app events, are the ones who suffer the worst from this fraud. This issue, however, is hardly exclusive to the gaming vertical.

In-app purchases, which many verticals rely on in one form or another, are also becoming a target for sophisticated fraud.  A record 2% of all in-app purchases in Q2 of 2019 were identified as fraudulent – 10x more than Q1 of 2019.

It’s all About the $$$

While the average CPI stands at about $2.89, cost per action rates can be as high as $4.58 for a registration event (beginning of user journey), and up to $40 or $87 for purchase or subscription events, respectively. 

The potential reward for fraud, based on these numbers, is therefore significantly higher. Even though CPA-based campaigns are not as common as CPI-based ones, the reward for a successful infiltration beyond the installation point, going under the radar of standard fraud protection tools, would mean a highly rewarding payday from CPA events.

 The fake sense of safety some marketers feel with CPA campaigns along with the high rewards involved make this a win-win situation for fraudsters.

“The only thing necessary for the triumph of evil is for good men to do nothing.”

– Edmund Burke –

Fraud is not taking a break. If there’s money to be made, fraudsters will surely try and find a way to get a piece of it. We must always examine the current situation and evaluate our next steps, staying put will give bad actors the small advantage they seek.

When looking at the evolution of online advertising models, fraud actually has an integral part in the industry’s development, as it comes up with creative methods to eliminate fraud, improving its positioning and performance in the process.

It’s now time to move ahead once again by protecting our investments, by looking for fraud where it’s uncomfortable to look, beyond the install and attribution point. 

Post-attribution fraud detection is an integral part of AppsFlyer’s fraud protection suite, Protect360, as we look to identify sophisticated fraud which is not identified in real time, but it doesn’t stop there.

As we look forward and develop our existing and future methods of fraud protection we take a look at events, post install behavior and user biometric data as we go beyond our comfort zone and uncover fraud where we once thought it was safe.

The post Why CPA Campaigns Don’t Protect You From Fraud appeared first on AppsFlyer.

AppsFlyer is always on the lookout to grow our extensive community of integrated partners. With the advancements in Audiences, we are thrilled to announce our newest integrated partner – TikTok Ads. 

According to AppsFlyer’s latest Performance Index, TikTok Ads is poised to become a major global player, taking the #1 spot in our growth index. Therefore, it comes as no surprise that TikTok’s explosive growth as one of the most downloaded apps of all time is only providing advertisers with more opportunities to expand into new markets, create brand new audiences, and test new ad formats.

AppsFlyer’s powerful Audiences management solution experienced massive innovation and growth this year, delivering updates in precise targeting, one-click connection, intuitive data visualization, and incrementality testing, to name just a few. We are thrilled to launch this impactful partnership and further provide our mutual clients with seamless audience creation and management capabilities.

Providing advertisers with a way to connect to an ecosystem is at the core of AppsFlyer’s values and allows our clients to benefit from a more efficient marketing experience.

We’re excited to have been working closely with TikTok Ads and look forward to more exciting updates in the future.

The post Announcing First-to-Market Audience Integration with TikTok Ads appeared first on AppsFlyer.

The post Top 5 Data Trends that Shaped Mobile App Marketing in 2019 appeared first on AppsFlyer.

In part I of our incrementality series, we covered the complexities of measuring incremental lift and the true impact of re-engagement activity. By testing different influencing factors, advertisers can gain true insight into the value of their re-engagement dollars. This is done primarily through very carefully crafted A/B tests.

A/B testing is not only a scientific notion but has also become a vital tool for digital marketers in the 21st century. It is one of the most powerful ways to gain a competitive advantage and priceless learning from your retargeting activity. Though insights gained from a successful A/B test are a strategic way for marketers to make informed budgeting and optimization decisions, they can also put them at risk of costly miscalculations if carried out inaccurately.

When it comes to A/B testing for incremental lift, getting the process right is critical. For starters, we recommend using an intuitive tool such as AppsFlyer’s Audiences to ensure success rather than building an A/B test from scratch.

The complexities of data management are hard to comb through and can potentially drain your resources and budget. To get started, consider the below points based on our extensive experience working with top brands across all verticals.

1. Creating Test and Control Groups Devoid of “Noise”

In today’s ecosystem, each one of your customers is subject to many marketing signals that impact his or her user behavior. This can include in-app messaging, push notifications, emails, offline and online marketing messages, to name a few. The challenge arises once you want to specifically measure the impact of one re-engagement campaign on top of an already existing program. In order to do so, you need to make sure that both the test and the control environments are not cluttered with external influencers tainting your test group (or, alternatively, that both are influenced evenly by the same exact factors… more on that later).

Generally, marketers can’t stop-play their entire program to test just one factor and its impact alone. So, what are they to do?

Say an eCommerce brand identified an audience for their test (people who registered but didn’t purchase in the last 30 days). This audience will be exposed to a retargeting campaign on Facebook (A).

Using AppsFlyer Audience’s Venn diagram, the eCommerce brand saw that the audience selected also overlaps with two other audience segments that are currently being targeted in other retargeting campaigns:

• B (targeting users who Installed in the last 90 days)
• C (targeting users who were dormant in the last 7 days)

Let’s focus on the two potential options below:

EXCLUDE OVERLAPPING AUDIENCES

Option 1: The brand can exclude users who exist in other audiences connected to retargeting campaigns (B and C) from their target test group A (Facebook retargeting campaign). That way they ensure that both test & control groups are excluded from all other retargeting campaigns running in conjunction at this time.

INCLUDE TEST AUDIENCE IN OVERLAPPING AUDIENCES

Option 2: Alternatively, they can make sure that the test & control audience used for the test on Facebook (A) will also be targeted by the same two retargeting campaigns (B+C).

In this scenario, the eCommerce brand could attribute the lift to the Facebook retargeting campaign, assuming that the results yielded a lift.

Audience overlap: identify the right audience for your test

With AppsFlyer’s split testing feature, you can create two or more randomized groups (test & control) at a click of a button; and by using AppsFlyer’s Venn diagram for visualization, you can see the overlap of your target audiences. This will help you create an ideal test scenario in no time!

2. Deciding on the Optimal Audience Volume and Duration for your Experiment

For each business, the volume of users differs. It is thus difficult to agree on which segment size would be best to test for optimal results and whether those results are statistically significant enough for further analysis. There is an important tradeoff between how confident you would be with the test results and the high cost of maintaining a lengthy test period.

Remember, marketers need to identify an audience segment that is qualified for retargeting. If the target group is refined to the point where it is very small, then keep in mind that the control group will be even smaller, rending the results insignificant.

If the daily volume of app users is limited, the duration of the test must be longer in order to capture more and more engagement, and best understand the average user behavior.

How much time needs to pass for the incremental test to result any meaningful insights? Would a daily comparison do? Not likely… as the user funnel lengthens and the touchpoints broaden it may take more time to understand the full picture.

For some business cases, conversions may trickle in after the end of the campaign. This changes the amount of users that may now qualify as conversions over time. Therefore, longer test periods may be necessary depending on the nature and time it takes for those users to fully convert.

To provide a good indication on the ideal volume and length for your test, you can measure the performance of your audience before the start of the test to understand the current behaviour of your audience in a pretest state. This baseline together with your expectations of the lift by the test campaign could help you estimate what conversion volume will be needed to make the test statistically significant.

3. Excluding Outliers

There will always be outliers in your data, and it’s important to detect and understand how to deal with them using the right method based on your KPIs. These edge cases can render misleading numbers that can cause incorrect calculation of your test results.

As mentioned above, volume is also important when considering what to do with outliers in your data. The larger your audience pool is, the less impactful your outliers are. Think about it, If you had one individual that made a purchase of $1000 the average revenue per user will be impacted more if you are looking at 20 purchases alone vs. 2000 purchases.

Though it is common to exclude top and bottom outliers from your data, is not always the right thing to do. For example, if your KPI is revenue per user, you cannot exclude the bottom 5% of your data because those ‘outliers’ include users who did not make a purchase vs. those who made a purchase: ($100 + $0)/2 users vs, ($100)/1 user.

Another thing to note, you may need to choose the right method to “normalize” your data according to your KPI. One example is using an average of averages approach in order to eliminate the inflation caused by high volume segments. When simply calculating your result’s average, you are not taking into account the impact of the size or volume of your segment.

Let’s say we only have purchase orders in two states, New York and Pennsylvania. The total order amount in NY is $1000 for 10 orders and $100 for 2 orders in PA.

With an average calculation, the average purchase order is $1100/12 = $91.67. This means that NY influenced the results here due to the higher-order volume and purchase amount. Therefore, this is not a good representative example of PA average order amount.

With an average of averages ($1000/10 +$100/2)/2 = $75, which is a $16 difference from the averaged calculation above.

Having the right method to deal with edge cases in your data can have a serious impact on the calculation of your test results. That’s why there is no “one method fits all” but rather a carefully crafted calculation based on each brand’s business model.

4. Pay Attention to Seasonality

Choosing the right time to start running your test is crucial. When you’re looking at your KPIs, take into consideration the normal trends of behavior for your users. One example that comes to mind is if you decided to launch your test around the holiday season, particularly when it falls on Black Friday and Cyber Monday. The results of your campaign will output a larger lift than the control group in this case because it’s not the right indicator of how the campaign would have performed had there not been any special events during that time. Keep in mind other major events or market trends happening when you want to run a test to prevent shifting the data unnecessarily.

5. Establishing the Right Process to Ingest and Review the Data from your Test

All of your efforts could go to waste if you are not able to analyze the results properly and better yet, take immediate action. Properly analyzing and reporting on your test results is a long and expensive process and can really drain your resources (particularly for your data team). Since some networks provide partial raw data or none at all, AppsFlyer closes the gap by providing an easy way to ingest the raw data you need to make informed decisions.

And we don’t stop there. Given AppsFlyer’s extensive experience in data visualization, we are working on providing some powerful visuals for your A/B testing and reporting needs.

Final Words

Leading companies will agree on the importance of a test and learn mindset. With retargeting budgets growing and advertisers experimenting with different placements, networks and ad formats, understanding the true impact of your re-targeting efforts is crucial in making budgeting decisions and justifying scale. More and more brands are developing testing methods to optimize their marketing program, and AppsFlyer is keeping up with the demand to provide high-end solutions for advanced marketing needs.

The post 5 Challenges to Consider When Measuring the Incremental Lift of Your Retargeting Campaigns appeared first on AppsFlyer.

Over the last decade, social media has emerged as a crucial driving force for lead generation, installs, and business.

Social media apps generate billions of app installs annually, enabling marketers and app developers to grow their user base and engage with existing customers. And apps provided by Facebook, Instagram, Snapchat, Twitter, and other social platforms offer versatility in the ways they help brands achieve goals across the funnel stages – from awareness to acquisition to engagement and beyond.

Given the importance of social media in any brand’s marketing mix, it goes without saying that any user experience spanning the social media-to-app journey must be exceptional and that all pre- and post-install activities must be measurable.

But, alas, things aren’t always so simple; mobile marketers, product managers, and app developers aiming to move prospects and customers from their social media properties to app experiences — especially those seeking to route (or deep link) into the app itself – run into obstacles.

When Links Break, so Does Attribution

Why is this happening? It’s simple and even logical. Social media platforms work hard to ensure that users stay in – and don’t stray from – their apps. So in some cases, when users try to leave, they face a less-than-ideal user experience, typically in the form of broken journeys and friction points.

This is what frequently happens when social media posts include deep links, especially when Universal Links (Apple’s deep linking standard deployed on iOS 9+) are involved. When users tap on these links, they often break; this leads to points of friction in a customer’s journey, in the form of error messages, dialog boxes, and users finding themselves in unexpected destinations. In turn, this friction sets off a chain reaction of frustration, abandonment, lost opportunities, and lower conversion rates.

This situation is obviously unacceptable for users. But it is equally undesirable for a brand since it cannot measure attribution accurately; broken journeys result in misattribution and non-attribution.

AppsFlyer’s Landing Pages to the Rescue

What is someone focused on acquisition and growth to do? This is where AppsFlyer’s new Landing Pages feature comes into play.

With Landing Pages, app marketers have a simple way to ensure excellent user experiences by providing a smooth path for customers from social media apps into a brand’s app.

The Landing Pages feature seamlessly transports users that tap on a deep link from social media posts and / or stories in Facebook, Instagram, Pinterest, Twitter, and WeChat directly to a landing page.

This mobile web landing page can display a “preview” — for example, product discounts, holiday sales, etc. — of in-app content. Showing in-app content is a proven way to deepen engagement with customers by giving them a “preview” or taste of what they will see in the app. Why is this so? When people know what to expect, they are more likely to engage with content when it reappears later in a customer journey.

Landing Pages include a OneLink deep link which routes users on the appropriate path:

• Users with the app will be directed to in-app content.
• Users without the app will be directed to the App Store.

With Landing Pages, it’s win-win-win: happy users, happy customers, happy marketers.

What else can Landing Pages do?

As shown above, Landing Pages help marketers provide a smooth user experience as well as accurate attribution metrics. But they can do much more:

• Design creatives for experimentation agility: A/B tests are predicated upon agility and iterations. Landing Pages empower marketers and product managers to rapidly design and release creatives to test and hone based on performance.
• Compare social media platform performance: Creating different links for each social media platform enables marketers to understand and contrast the performance of different social media apps, landing pages, and campaigns.
• Customize and personalize: Marketers can rapidly create distinct landing pages for different sources, segments, and campaigns and can then easily monitor the performance of the various landing pages.

How Do Landing Pages Work?

With the WYSIWYG editor tool, designing and customizing landing pages is a cinch. Marketers and app developers can gain newfound agility, deploying landing pages rapidly without having to request development resources to design and launch content.

Here’s an example showing how easy it is to design and customize a landing page for a Black Friday campaign:

Voila! You now have a landing page, and you can use deep links, including Universal Links, in your social posts without risking poor experience.

But there’s more. By design, OneLink is an integral component of the overall AppsFlyer solution. That means that metrics such as CTR, installs, LTV, and more are provided in clear visualizations for these OneLink-powered Landing Pages.

The post Announcing AppsFlyer Landing Pages: Removing Friction from Social Media-to-App Journeys appeared first on AppsFlyer.

GDPR was a turning point in the world of data privacy. Not just due to the sheer magnitude of the regulation and how broad its effects were, but because of how it made data privacy a talking point in offices and at dinner tables around the world. It made privacy mainstream. The irritating inbox ambushes about “updated terms and services” from long-forgotten brands inspired everything from funny memes and tweets to a famous Spotify playlist. In the weeks leading up to May 25, 2018, everyone knew what GDPR was—whether they were in the affected area or not.

On the business side, GDPR also forced a shift in the global approach to data privacy. Many organizations that had barely been touched by previous regulations were forced to rethink how they collect and manage sensitive user data. In the year and a half since GDPR went into effect, some of the world’s largest corporations have been heavily fined, the number of in-house DPOs at European corporations has skyrocketed, but perhaps—most importantly—overall awareness among the public about data privacy and what their inherent rights are, has grown substantially.

GDPR was just the beginning

Aside from creating a global shift, GDPR served as one of the gateway regulations for data privacy laws in the digital marketing sphere. Similar bills and laws have since been proposed and accepted, one of the most recent and significant of which is the California Consumer Protection Act (CCPA).

Going into effect on January 1st of 2020, the CCPA will provide similar coverage for the residents of California, holding organizations accountable for collection and management of data from users in that region. While this is not the first data privacy protection law to be passed in the US, it covers a broad range of organizations, requiring many to take measures towards compliance that they may have not had to previously.

Many believe that while the GDPR marked the beginning of significant data privacy legislation, CCPA will mark the beginning of real compliance enforcement; it won’t just be the big players anymore that will be held accountable and fined for violation. This is one of the reasons CCPA and similar legislation have been causing a great deal of worry among management professionals across almost every industry.

A recent survey from Gartner shows that “accelerating privacy regulation” has become the #1 concern for senior executives worldwide, across industries and countries. It has surpassed many other topics, such as sourcing high-quality talent. Accelerating privacy regulation was furthermore named a risk of “very rapid velocity,” due to the severe negative impact it could have on an organization in a short matter of time. 

Despite this concern, alarmingly, very few companies report being CCPA-ready. A survey by TrustArc found that in March of this year, only 14% of respondents reported that they are CCPA-compliant. Of those questioned, 83% replied that they believe they can leverage at least some of the preparations put in place for GDPR compliance, and they’re not wrong.

Organizations that have already taken the necessary steps to meet the GDPR standards are on the right path to CCPA compliance, but there are still differences between the two regulations that need to be addressed. One of these differences, for example, concerns the scope and frequency in which data subjects need to be updated about how their personal information is collected or disclosed to business partners. With that said, however, many of the processes put in place ahead of GDPR can be leveraged towards CCPA compliance, including steps taken in regards to: 

1. Data mapping 
2. Processes to receive and handle data subject requests 
3. Methods to delete personal information 
4. Methods to provide access to personal information in readily useable formats
5. Technical and organizational measures used to protect personal information
6. Privacy notices

Putting privacy at the forefront

CCPA is just a few short weeks away from taking effect, and other similar privacy regulations are expected to follow in 2020 and beyond. New bills have already been passed or proposed in Nevada, Washington and New York, some of which are more extensive and severe than the CCPA. The rest of the world is also well on its way.

At AppsFlyer, we’re beyond pleased at the increased awareness globally to the importance of user privacy. It is a critical piece of the evolving technology ecosystem, and failure to embrace this movement is a severe disservice to all players — from service providers to end users (and everyone in between). Privacy is a fundamental human right, and should be maintained and honored through every extension of your presence –mobile and digital devices included.

When data is your business, privacy needs to be embedded into the fiber of your technology. A fundamental piece of the privacy puzzle is keeping data, well, private. It is therefore our commitment to our customers and to ourselves that we will never make selling data our business. Each customer is in full ownership of their own data, it is never sold or shared with other customers; not via data collectives or shared persona graphs. Our extensive security and privacy program is the reason why we are trusted by some of the world’s most sensitive brands, which include financial and insurance institutions. 

We have taken and will continue to take extensive measures to ensure that the company, products and services we offer are compliant with global and regional privacy regulations. We’ve spent the past several months working to finalize the organizational preparations for the CCPA. Under the CCPA, AppsFlyer is considered a service provider. The act of disclosing personal information to an entity that processes it on behalf of the business for business purposes (as defined by the CCPA) is permitted and will not be deemed as “selling” data. In addition, AppsFlyer is compliant with numerous global and regional privacy certifications, including:

• ePrivacy seal
• TRUSTe
• ISO 27018
• EU-US and Swiss-US Privacy Shield Framework

Customers working with us and potential future customers can rest assured that working with AppsFlyer means working with a partner that is privacy-forward and compliant. 

Putting privacy at the forefront does not just involve our own compliance, however. AppsFlyer is dedicated to assisting our customers in preparing for this (and other) regulations, through education, transparency and support.

The post CCPA Ushers in a New Era of Data Privacy appeared first on AppsFlyer.

If you are a performance marketer, you’re probably looking to get more from your data with one goal in mind: to seamlessly & accurately funnel data from multiple sources to one place, for holistic, coherent and reliable analysis. The trouble is, with all the data fragmentation coming from different sources, macros, varying URL formats, and campaign structures, many of us struggle to get insights from our data. Even pulling the data in the first place poses its set of challenges.

For us earthlings generating data continues to expand, making extracting the right insight at the right time nearly impossible. Most marketers analyze after the fact, too late to make any impactful, critical iterations on the fly. When marketers dish out top dollar for their efforts, they are expecting deep insight in return and at the very least, the ability to accurately calculate true ROI to improve their marketing strategy. Without it, how can they understand whether their strategy is effective or whether they should consider changing course entirely?

For most performance marketers, ROI is the single most important KPI; but calculating ROI across your entire marketing program is not a simple task. How can you imagine making strategic decisions when this metric is inaccurate, unavailable or completely submerged from your analysis? While you pull revenue data from different sources, without full coverage and granularity of our marketing spend, there is no way to precisely calculate ROI; and therefore, you continue to remain in the dark.

Announcing Ad Spend Ingestion

At AppsFlyer, we are committed to providing solutions to some of the most complex problems that performance marketers face. Covering 100% of your ad spend data to extract meaningful insights is a major goal of ours. That’s why we are excited to announce a series of new features, one of which is the Ad Spend Ingestion Tool. This will allow advertisers to upload spend data from any source, even those that don’t have an existing API integration already.

With the Ad Spend Ingestion tool, users can easily upload spend data from anywhere, including owned and earned media, influencers, and even out-of-home (OOH) like TV/radio/billboard and more! Simply send your file via an email attachment or upload directly to the UI and let AppsFlyer do the rest.

Using the ad spend ingestion tool, you can even send an optional Geo dimension in your file for deeper analysis by a specific segment; and files can accommodate data from multiple apps.

The CSV validation logic will verify that all mandatory fields are provided and validate whether they are in the right format. Advertisers can send the file via email and will be provided with automated feedback.

Data Visibility and Partner Transparency

Providing partner networks the proper access they need to ensure that the data passed from their source is accurate, timely and true is essential to partnership success. We understand how hard it is for advertisers to manage this process on their own and manually. That is why we provided a solution for partners to take ownership on behalf of the advertiser to easily and seamlessly pass spend data to AppsFlyer.

Starting today, our partner networks have the ability to upload spend data on their advertiser’s behalf and access can be controlled from a dedicated permissions tab.

Ad Spend Ingestion Page

Full Control and Visibility

With managing ad spend, advertisers gain full visibility into each and every file uploaded. For example, what various apps in the file, the date-range sent, last uploaded by, and more! All in an effort to maximize efficiency and transparency between advertisers and partners.

It is crucial to have the full view and control of your data at all times, therefore we provided access to a full file management dashboard so that both partners and advertisers have the ability to troubleshoot gaps in the data, understand data integrity through match percentages and the total ingested spend. Advertisers can also revert incorrect files to ensure accuracy. 

Partners who are provided with permitted access can see their specific network’s files, while the advertisers can see all files being ingested across the board. This allows for optimal collaboration directly from one central dashboard.

A Growing List of API Integrated Partners

Some time ago, AppsFlyer announced the first automated ad spend/cost and ROI reporting for Facebook, Google, Pinterest, and others. Today, AppsFlyer has over 185 integrations for cost and ROI reporting for both click and API based integrations, with new partnerships continually announced.

We recently added Liftoff to the list of API-based integrations; which already includes the likes of Facebook, Google, Pinterest, Tapjoy, AppLovin, AdColony, and many more. Through this integration, clients can now measure their performance and ROI down to the ad and geo level for Liftoff, one of the top trafficked networks in AppsFlyer.

These advanced integrations provide instant access to campaign data like impressions, clicks, and cost, alongside post-install activity such as in-app engagement metrics, ARPU, etc.

Integration Dashboard – Cost API Status

We also released an API status page to ensure visibility into data re-freshness, and connection validity, so that advertisers gain full confidence and visibility into their ad spend data flow in AppsFlyer.

Integration Dashboard

Using the most innovative solution makes or breaks even the most refined marketing strategy. Funneling your ad spend data into one centrally-managed dashboard gives you unprecedented data visibility, analysis capabilities and the exponential ability to scale! And remember, smart decisions only come to those who are empowered by smart analysis.

We are so proud of the solutions we roll out to our customers and hope to keep delivering what is expected of the most innovative attribution platform in the market.

For more information on AppsFlyer’s Ad Spend feature, reach out to your Success Manager today.

The post Calculating Your True ROI with Powerful Ad Spend Enhancements appeared first on AppsFlyer.

In the era of  “data is the new oil,” attribution data has become the fuel for decision-makers across marketing, product, data insights, R&D and even finance departments.

While navigating the complexities of the emerging MarTech landscape, many AppsFlyer customers look to their agency partners to help them effectively deploy and use the multiple disciplines where data makes an impact on their budgets and business decisions.

AppsFlyer’s mission is to deliver value to our end-clients around the globe —the brands. As the global market leader in attribution and marketing analytics, AppsFlyer understands it is in the unique position to support agency partners in driving positive performance and unlocking new sources of growth for mutual brand clients, with regards to mobile and cross-device activation, data analysis, ad spend optimization, creative, planning, media buying and more; through in-house or external teams.

AppsFlyer has long invested in providing agency professionals with education around best-practices so that they are effectively equipped to consult their brand clients on developing data-driven digital strategies, on leveraging platform capabilities, and on  advanced data modeling to optimize ROAS.

“For 360i, going from micro tech implementation to macro strategic thinking and learning how to guide clients when it comes to the attribution that informs larger marketing decisions is mission critical. This program provides us the ability to quickly transform technical knowledge into strategic business concepts that help clients capitalize on change.”

Vladimir Golinder, VP, Head of Media Technology

Today, AppsFlyer is excited to announce the official global launch of  Solutions Partner Program for Agencies (SPP), a dedicated training and certification program designed to help agency partners from across regions and markets enhance delivery for their clients. 

SPP Launch Partners

AppsFlyer is proud to highlight the following launch partners. Certified Solutions Partners have developed in-depth expertise, met thorough standards and include agency partners that have qualified as proficient platform users and mobile experts:

• 360i (USA)
• Adways (Japan)
• Bamboo (USA)
• CyberAgent (Japan)
• Hearts & Science (USA)
• i-Cherry (Brazil)
• M&C Saatchi Performance (India, Singapore, United Kingdom, USA)
• MediaCom (United Kingdom)
• Merkle (United Kingdom)
• Mindshare (United Kingdom & USA)
• Omega Media (Vietnam)
• PMAX (Vietnam)
• Reprise (South Africa)
• Septeni (Japan)
• Yodel Mobile (United Kingdom)

“Our digital media agency, recently awarded ‘Agency of the Decade’, is always looking for ways to remain agile in the industry, build deep expertise and deliver more growth for our clients. As an AppsFlyer Global Alliance Solutions Partner, our teams are now fully trained and certified, providing us with an even stronger foundation to scale delivery, optimise performance and increase efficiencies for our clients.”

Dan Rosen, Global COO

Empowering Agency Partners to Stay Ahead of the Curve

In 2019 alone, AppsFlyer customers made $20 billion worth of marketing decisions leveraging AppsFlyer technology. 

AppsFlyer’s Solutions Partner Program for Agencies provides a means for agency professionals to learn and validate knowledge and expertise in mobile and cross-channel marketing, as well as AppsFlyer capabilities. 

In addition, AppsFlyer Solutions Partners are able to:

• Gain exclusive partner access to new AppsFlyer’s products for their clients
• Access dedicated best-practice resources
• Create and manage joint marketing partnership with AppsFlyer
• Get dedicated sales & technical support
• Receive pre and post-activation consultation

In preparation for the launch of this program, the AppsFlyer Agency Alliances team has been working at full-speed to provide long-term value for agency partners by serving as a trusted extension of their team — to deliver training and certification to agency teams around the world. 

SPP is a commitment to our agency partners and to our brand clients: we’re here to empower you to succeed.

For additional details on AppsFlyer’s Solutions Partner Program for Agencies, visit: https://www.appsflyer.com/appsflyer-solutions-partner-program/

The post Announcing AppsFlyer’s Solution Partner Program for Agencies appeared first on AppsFlyer.

It’s December, winter is here and you can just feel Christmas right around the corner. 

December, along with its many holiday traditions, is usually when you’ll run into almost every company’s yearly summary, showing how everything is bigger and better, throwing big numbers around in an attempt to showcase how awesome they are.

While that’s all fine and well, when conducting our research for this post we ran into another interesting trend that caught our attention.

A Bit of Background: 

In July of 2019 we introduced a new feature into our fraud prevention suite, Protect360, called Post-Attribution Fraud Protection.

While a new feature release is not usually a big deal, this one relies on a concept that went against the industry’s core belief – that all fraud can be blocked or at least identified in real time. 

Post-attribution is an additional layer of protection on top of our existing real time protection suite, meant to identify new fraud patterns as they materialize in our ecosystem and block them, even after the attribution takes place.

Fraudsters who are well educated about existing fraud prevention mechanisms, focusing their efforts on going through the install phase without being identified, are now less likely to go unnoticed.

Why is this important you ask?

Quite simply, it means breaking preconceptions about fraud patterns, where and when they originate and exposing yet another loophole that allowed fraudsters to thrive. 

Which takes us to this declining trend. 

Wait, Isn’t a Declining Trend a Bad Thing?

This declining trend is one we’re actually very proud of. 

Since the product relaunch in July, the number of AppsFlyer customers using Protect360 increased by 28%, however the number of attempted fraudulent attacks on AppsFlyer’s clients (presented in the graph) decreased by 33%!

The number of fraud attacks reflects an aggregation of actually blocked fraud and identified attempts, meaning that since July AppsFlyer managed to decrease its entire ecosystem’s exposure to fraud by 33%   

This figure is amplified by the fact that Appsflyer’s monthly measured install count grew by 20% in that time period. 

Many fraud prevention vendors take pride in the large number of installs they manage to block, however, the question that should be asked is: 

Why are they going through so many attacks in the first place?

The answer is simple.

Fraudsters are no longer working randomly, they do their due diligence and choose their targets wisely; fraudsters aim for the injured gazelles, the easy prey, where their attacks are most likely to go under the radar, an experienced fraudster would already know which protection mechanisms are easier to infiltrate and which are less likely to damage their ROI by identifying their activity as fraud. 

A weaker fraud prevention solution can easily be identified by fraudsters mapping their target’s servers calls during their BI process. This can be done using different tracking and web-sniffing tools available online.

Same goes for the attribution provider’s SDK, with some SDKs more vulnerable than others for hacking or reverse engineering; open source SDKs posing a much more appealing target than closed source ones, with their code exposed for anyone to see.

At the end of the day, fraudsters aim to achieve a positive ROI. They set their sights on specific targets that either don’t have active anti-fraud solutions in place, or ones using solutions they know how to bypass, steering away from protection solutions that are more likely to block them.

Simply put, AppsFlyer clients are significantly more protected against potential fraud attacks simply because they’ve chosen AppsFlyer in the first place.

Working On the Cure 

Several months ago, we discussed AppsFlyer’s commitment to not only treat the fraud disease but actually cure the industry from it.

The decline in fraud attacks attempted on our ecosystem shows we’re headed in the right direction. As we’re looking to truly solve the fraud issue rather than simply treating it, this fraud prevention trend fits into our vision of a fraud-free environment.  

But we’re not done yet.

If there’s anything to be learned from our experience in ad fraud prevention is that fraud is usually lurking in the corners where we assume it’s not.
No solution can ever be considered as bulletproof. This mentality is crucial to not only keep up with fraud, but possibly beat it to the punch, and to do that we can’t afford to rest on our laurels. The problems we don’t yet know about are the ones keeping us alert, as they’re the ones not reflected in the graph above.

2020 and the Challenges Ahead

A review of 2019 and the trends leading up to it will help us take a sneak peak into the challenges ahead in 2020. Trends and numbers can tell a story and lay out upcoming targets as we attempt to eradicate mobile ad fraud from our ecosystem.

Evolution of Fraud Methods

New and improved methods are very likely to be introduced moving forward, targeting loopholes across different stages in the user journey, using technology we may not be familiar with yet. 

The evolution of fraud through the years has shown that fraudsters are not only adaptive but also very innovative. 

Constant analysis, modification and improvement of existing identification and blocking rules and patterns, as well as development of new preventative technology is imperative in order to stay ahead.

Going Above and Beyond (App Install Fraud)

In-app fraud has long been a growing focus point for fraudsters. Going deeper into the user journey presents a double benefit for fraudsters:

1. Whitewashing fraudulent app installs by falsifying user engagement beyond the app install point.
2. Gaining CPA rewards for engagement measuring events or even in-app purchases. 

As more marketers adopt in-app event engagement measurement and reward partners using CPA models, we don’t expect fraudster focus will go away anytime soon. 

A Global Epidemic 

Each country and region pose different behavior patterns, technology adoption and market regulation. Some regions are slower to raise awareness to fraud, adopt anti-fraud tools or lack the regulatory ability to properly treat it, all of which could affect the entire ecosystem. 

Increasing global adoption of anti-fraud measurements, KPIs and awareness is required.

What’s Next

Vigilance and transparency are key for identifying new fraud patterns and protecting our ecosystem from those who wish to exploit it. The fraud methods we know and are able to block are very likely to evolve, apply better technology and find new ways to make financial sense for their operators. The four fraud types mentioned above might break these operations into specified methods that are easy to understand and study, however reality shows that a combination of methods could be the dominant trend moving forward. 

The biggest concern at any given moment is of the risk we’re not yet aware of, these new types of fraud will require yet further steps forward in both perception and technology. While you are reading this article fraudsters are likely already hard at work on this next generation of ad fraud. 

As we look ahead to 2020 we aim for more declining fraud trends. It’s our responsibility as ad fraud prevention vendors to make fraud operations a bad investment for the ones operating them, make their job as difficult as possible and aim to lower their ROI – pushing them away from our network. 

The post 2019 Fraud Trends Uncover Fascinating Results appeared first on AppsFlyer.

Marketers rely on AppsFlyer as their centralized attribution platform and the single source of truth for attribution, measurement, and reporting. Since the beginning, it’s been part of AppsFlyer’s core mission to work with all the global partners, serving as an unbiased mobile attribution platform. As such, it has always been a top priority for our team to provide advertisers with the tools they need to deep dive into advanced data management. For the past three years, AppsFlyer has had extensive integration with monetization networks providing ad revenue attribution based on aggregate revenue data. It has become clearer that the Industry must take a step forward toward providing more accurate data for advanced marketers. That’s why we are ecstatic to announce the NEW granular, ad revenue integrations with IronSource and MoPub (a Twitter company), together with additional integrated partners for Ad Revenue Attribution via API Integration: TikTok Ads, Bytedance, Mintegral, and Voodoo Ads. 

No matter what monetization strategy advertisers choose: in-app ads, in-app purchases, subscription model, freemium or premium options; they must be able to measure the performance coming from all of their different revenue streams.

In today’s highly competitive world, where marketers are battling for eyeballs, and ad real-estate is increasingly expensive, app owners are looking to expand ways of generating revenue. One way to do so is by tapping into other forms of ad real estate, for example by running ads within their app!

The challenge, however, is that mediation networks hold different data points regarding how users interact with these in-app ads. Further complicating things, some networks pass revenue data on the averaged, aggregated total, by simply dividing the total revenue generated on a specific ad placement by the number of users that clicked or saw that ad. With that being the case, the true LTV or ARPU can be calculated only to a certain extent.

Though in-app ads could potentially be a major source of revenue for an app, without understanding the engagement behind the revenue, how could advertisers optimize performance and scale?

Considering that the number of advertisers who are generating revenue from in-app ads is only growing, the task of measuring granular revenue is more important than ever.

With that in mind, mediation platforms realized the enormous value of sharing granular data and are now providing accurate user-level data down to the device ID, in the case of IronSource and impression-level revenue data in the case of MoPub, allowing attribution providers to tie revenue data back accurately to the attributed source. That makes understanding user-level behavior and user-value possible, which in turn allows for better optimization, and for advertisers to tap into potentially lucrative revenue streams for their business – a win/win for all!

What’s the solution?

We are happy to see that more partners are adopting the industry-accepted approach for exposing both user-level (IronSource) and impression-level (MoPub) revenue data by the leading mediation networks.  By fully integrating with mediation platforms such as MoPub and IronSource, AppsFlyer is able to collect revenue data segmented by different attributes and also report on a granular level, so that true LTV, ARPU and ROI metrics can be confidently calculated.

“The next evolution of mobile growth requires closing the monetization and marketing loop to gain a holistic picture of app business health and achieve complete optimization,” said Yevgeny Peres, VP Growth, Developer Solutions at IronSource. “As one of the pioneers in making ad revenue measurement data as granular, accessible and actionable as IAP data, we’re excited to see this integration with AppsFlyer enable our mutual partners to better understand the true cohorted ARPU and ROAS of their campaigns and ultimately accelerate app growth.”

Ad Revenue Attribution: Now with New Connected Partners

In addition to the user-level and impression-level integrated partners mentioned above, we are also providing additional ad revenue monetization partners to the already extensive list. Part of integrating with partners means that our users are able to receive the related revenue amount, all seamlessly and automatically via the API connection. 

AppsFlyer’s Ad Revenue Attribution feature allows advertisers to understand the true performance of their user acquisition efforts by correctly tying the ad revenue generated back to the user acquisition source. This leads to a more accurate and complete view of their ROI.

Today advertisers can head over to the overview dashboard to analyze their revenue generated from in-app ads. In addition, they can further understand their ROI in the cohort report where they can separate their in-app ads revenue from other revenue streams, such as in-app purchases.

“Acquiring new users has never been more competitive — and providing publishers with revenue data at the impression-level, in real-time, is critical to optimizing user acquisition and monetization. For publishers who want to grow ad revenue, it is practically running blind without having insights on the most basic ad event – the impression. Making impression-level revenue data available to attribution providers like AppsFlyer will provide greater granularity to our mutual customers, allowing them to turn this data into actionable insights and offer the holistic view needed to propel their business to the next level.”  – David Gregson, Product Manager at MoPub.

Data is only useful if you have a good way to view it, analyze it and use it. Being able to collect your ad revenue across different sources, at any granularity, and funneled it into a coherent, centralized view is AppsFlyer’s ultimate goal, and we have more in store to ensure that becomes a reality for our advertisers.

Subscribe to AppsFlyer content

The latest mobile marketing tips & trends delivered to your inbox.

The post Measure Granular App Revenue with AppsFlyer Ad Revenue Attribution appeared first on AppsFlyer.

In 2010, building an app’s user base was easy. First, the app was developed, then uploaded to the App Store, and finally, traffic began flowing in. Fast forward 10 years and now the world of apps is highly sophisticated, driven by data to inform decision-making and powered by robust tech solutions. The app space is also extremely competitive, with millions of apps vying for users’ attention.

Given the lightning fast evolution of the app space and how apps are marketed, let’s take a step back and explore what made app marketing the massive industry it is today.

Monetization: Premium > Freemium > IAA > Subscription

In the early days of the App Store, installing apps cost money, but after Apple released in-app purchases (IAP) in 2010, developers were able to focus on widespread exposure by offering a free, lighter version. Users could then decide whether or not to pay for full or extra features. This was a fundamental and rapid shift because it allowed users to adopt a more exploratory approach to diverse app discovery. Clearly, Freemium and IAPs dominated throughout the decade:

But in-app ads (IAA), especially display banners, were always there, even while additional ad units like interstitials, rewarded videos and offerwalls emerged. Burst campaigns and incentivized advertising were common and were used, for better or for worse, in various UA strategies.

In 2018, ‘Hyper Casual’ games — simple, highly addictive games with a very short shelf life — introduced IAA at scale, which first spread across Gaming genres (e.g. 34% increase in Core games, according to AppsFlyer data), and later to other verticals. Hyper Casual games ignored the common monetization model completely, monetizing solely on ads. Suddenly, non-paying users (the vast majority) could generate incremental revenue without harming the user experience.

In recent years, there has been an increasing adoption of subscription models (7x increase in the share of apps measuring subscription, according to our data), historically associated with verticals such as Entertainment and Music (think Netflix and Spotify). Unlike Gaming’s individual role in the shift to in-app advertising, subscription models are becoming more common across a variety of verticals, such as Gaming, Lifestyle, and Dating.

The Decade’s Growth Engine

Apps provide value to users, who, in return, generate income for app developers. To maintain this ideal flow, apps need to attract and engage a significant volume of users for as long as possible. While the beginning of the decade saw increased reliance on the Apple App Store (and Google Play after 2012), the boom in the sheer number of apps made doing so nearly impossible. By mid 2010, the App Store had 200K apps; in 2020, there will be nearly 7M, according to Statista.com.

As organic discovery became largely broken, apps were no longer able to rely on organic installs at scale. Additionally, marketers started running more cost-per-install (CPI) campaigns. Despite higher media costs, this guaranteed an install at the end of the funnel.

Today, apps focus on buying users and leveraging deep data insights to drive growth, rather than using paid traffic as a means to drive organic lift. In reality, getting organic users became so difficult that marketers had no choice but to focus on improving their UA methods and scaling their budgets. By putting their faith in data, they had the confidence to ensure their financial investment generated returns.

While marketers previously had to rely on relatively superficial app-level data (e.g. platform, geo), today, the sheer richness of data points can make your head spin: from LTV and ROAS to building predictive models to connect preliminary behavioral signals to lifetime value.

In the past couple of years, we’ve seen AI and automation take over, digesting data at scale and making crucial decisions behind the scenes in real-time.

Needless to say, ‘shooting in the dark’ is a thing of the past. The playing field is illuminated, allowing marketers to set and prioritize their goals, find the optimal price for every campaign and every user, divide budgets between sources, test multiple scenarios simultaneously and more.

The Duopoly and the Rest

While Facebook and Google were around first, inevitably, rapid growth attracted other ad networks to seek a piece of the remaining pie. Advertisers who couldn’t optimize via Facebook and Google due to budget or qualification reasons, or wanted to drive reach and diversify their portfolio, became a perfect match for numerous ad networks offering high reach at a lower price.

However, unlike the duopoly, most ad networks served as mediators between advertisers and publishers. This complexity limited transparency, making it nearly impossible to validate the authenticity of each side.

The growing interest in mobile apps and the inflation among ad networks, attracted fraudsters who have been active on the web for years. Now, they had an opportunity to attack a relatively vulnerable industry, polluting the entire funnel from an advertiser’s wasted budget to a publisher’s irrelevant ads, via the ad network.

App install fraud became a huge pain point for app marketers, creating a crisis of trust between marketers and ad networks. Billions were lost, and this forced the industry to heavily invest in developing protection solutions, which have seen success. However, as fraudsters evolve, new threats like in-app fraud are emerging in an increasingly high stakes game of cat and mouse.

The mobile app industry today has zero-tolerance to fraud, with marketers minimizing the number of media sources, and following strict protection guidelines.

The Evolution of Attribution

10 years ago, mobile app attribution did not even exist. Mobile was a fragmented collection of multi-platform pieces, making measurement extremely difficult. But as the industry exploded, and the range of sources increased, marketers were in need of systematic campaign evaluation and optimized budget distribution.

The evolution of attribution technologies which made mobile apps the most measurable ecosystem ever created was a significant milestone in this decade. It is unlikely now that an app marketer cannot imagine his day-to-day work without attribution data.

Today, mobile app attribution solutions are extending beyond mobile as the quest towards holistic or people-based attribution continues. Because mobile is the most frequently used device, its role in full circle connectivity across online and offline touchpoints is key to success.

What will the next decade bring? AR, VR, AI, 5G… But who knows what the future holds? One thing’s for sure, data will continue to push the industry forward!

The post In Data We Trust: A Look Back at an Explosive Decade in App Marketing appeared first on AppsFlyer.

The post We Just Raised $210M, and We Couldn’t Have Done It Without You appeared first on AppsFlyer.

AppsFlyer is always on the lookout to grow our extensive community of integrated partners. With the advancements in Audiences, we are thrilled to announce our newest integrated partner – TikTok Ads. 

According to AppsFlyer’s latest Performance Index, TikTok Ads is poised to become a major global player, taking the #1 spot in our growth index. Therefore, it comes as no surprise that TikTok’s explosive growth as one of the most downloaded apps of all time is only providing advertisers with more opportunities to expand into new markets, create brand new audiences, and test new ad formats.

AppsFlyer’s powerful Audiences management solution experienced massive innovation and growth this year, delivering updates in precise targeting, one-click connection, intuitive data visualization, and incrementality testing, to name just a few. We are thrilled to launch this impactful partnership and further provide our mutual clients with seamless audience creation and management capabilities.

Providing advertisers with a way to connect to an ecosystem is at the core of AppsFlyer’s values and allows our clients to benefit from a more efficient marketing experience.

We’re excited to have been working closely with TikTok Ads and look forward to more exciting updates in the future.

The post Announcing First-to-Market Audience Integration with TikTok Ads appeared first on AppsFlyer.

I am excited to share with you that we have secured $210 million in a Series D funding led by General Atlantic with participation from our existing investors. At a $1.6B valuation, AppsFlyer has become the first and only unicorn in the attribution category. Our long-term unbiased and independent positioning combined with our customer-obsessed mindset have allowed us to build something far beyond our wildest dreams. And we are just 1% done.

Since the previous funding round three years ago, AppsFlyer has grown its team 4X to 850 employees across 18 global offices. Our ARR has grown 5X, exceeding $150 million in 2019. This followed a five-year growth in ARR from $1 million to $100 million that made AppsFlyer one of the fastest-growing SaaS companies. Last year alone, our customers made $28B worth of decisions using AppsFlyer’s attribution platform.

In the last couple of years, we’ve adopted a very long-term vision, measuring the quality of our decisions in a 5-10 year time frame. I am confident that partnering with General Atlantic will allow us to push our long-term mindset forward and have a significant impact on our company, our customers, and on the entire market.

Exciting times like these are an excellent opportunity to thank everyone who has been a part of our journey:

Our Customers

• Thank you for choosing AppsFlyer to represent you – the marketer – and your interests in this complex ecosystem. We see ourselves as the marketers’ best friend, and hope you do too.
• Thank you for recognizing our unbiased, independent, and transparent approach in a market full of conflicts of interest.
• Thank you for trusting us with your most valued asset – your customers’ data. We will continue to adopt and lead the highest standards when it comes to security, data privacy, and compliance.
• Thank you for valuing our efforts to create an outstanding customer experience, and thank you for your understanding, patience, and loyalty when, at times, our delivery fell short.
• Thank you for your feedback along this journey. It has allowed us to build this robust attribution platform, which is now becoming a core part of the modern marketing stack.
• Thank you for letting us be a part of your journey and growth.

Our Customers’ end users

While we don’t have a direct relationship, you are our north star in every decision we make. In order to do what’s right for our customers, we must first do what’s right for you. This is also why we ask every AppsFlyer employee to have their consumer hat on every day. From enabling a great web-to-app user experience, to leading the highest security and privacy standards. Thank you!

Our Partners and ecosystem

• Thank you for seeing the value in transparent measurement and for taking special care of AppsFlyer’s customers. We will continue working, together, to make this complex ecosystem safer and fraud-free. In an effort to solve the fraud prisoner dilemma and the bleeding cash cycle we hold our partners to the same high standards we hold ourselves, for the benefit of compliant partners, our customers, and the entire market.
• Thank you for investing in our partnership. We take great pride in the strong relationships we’ve created together and in the deep technical integrations and products we’ve launched.

Our Competitors

• Attribution is extremely hard, complex, and sensitive. Many companies have tried to build attribution platforms. Most have failed. I have a lot of appreciation for what you’ve managed to build.
• We thank you for challenging us and pushing us to deliver an outstanding experience and unparalleled value to our customers. It crystallizes our unique brand and positioning in the market, as well as our long-term vision.

Investors

• In the early days, back in 2011-2012, more than 40 investors turned us down. While it wasn’t a pleasant experience, I’d like to thank you nonetheless. It made us stronger.
• Thank you Magma Venture Partners, Pitango Venture Capital, Eight Roads, Goldman Sachs, Qumra Capital, and DTCP (Deutsche Telekom Capital Partners). Thank you for believing that we can grow both personally and professionally to take AppsFlyer to the next stage and build a meaningful business in a complex market.
• Thank you all for accepting our “Revenue is Secondary” mindset. A framework that maximizes long-term value for customers and for the market.
• And thank you General Atlantic, for believing in our vision to democratize marketing. We promise to always do our best to make you proud to be a part of AppsFlyer. We are thrilled to be partnering with great people and a special company that has such an amazing track record and philanthropic heritage.

The incredible AppsFlyer team

• In our last internal engagement survey, 97% of you mentioned that you are very proud to be part of AppsFlyer. The fact that you take such pride in your work is what defines our success as an organization. I feel very fortunate to work with such a diverse group of enthusiastic, smart, and empathic people. I learn so much from you every day. Thank you for that!
• In the very same survey, you gave hundreds of suggestions for improvement. Your willingness to support each other, challenge the status quo (myself included), and embrace change, never ceases to amaze me.
• Thank you for seeing challenges as opportunities, for your empathy, your honesty, and your constructive feedback.
• Thank you for your ALL-IN attitude – for acting in the best interest of the company, our customers, and their end users. You demonstrate our customer-obsessed culture day in and day out, and we see the results every day through incredible feedback from happy customers.
• Thank you for understanding how fortunate we are, and taking the time and effort to give back to your communities through AppsFlyer Cares.
• Thank you for allowing yourselves to make mistakes and learn. Seeing our people grow and evolve personally and professionally is one of the highlights of my job.

Our friends and families

• You play a massive part in what we have created. Thank you for letting us dream, for believing in us, and for providing us with much-needed support.
• Thank you for taking part in our family activities, and for visiting us in the office every once in a while. Our kids.
• Thank you for reminding us to keep experimenting, make mistakes, learn, and evolve.

Don’t cue the music yet, I’m almost done

As attribution becomes a crucial element of every marketers day-to-day, this funding will allow us to continue to execute on our vision of providing marketers with a robust, unbiased, and independent attribution platform. The natural progression for us is to maintain an open platform for partners and third-party developers. This will enable brands to innovate in ways that are almost unimaginable today. We’re thrilled to have General Atlantic’s support on our journey towards democratizing marketing.

1% Done. I hope you’re ready for the next episode!

The post We Just Raised $210M, and We Couldn’t Have Done It Without You appeared first on AppsFlyer.

Lifetime value (LTV) forecasting is essential for mobile app developers trying to understand the total value of their user base in quantitative terms. 

Once calculated, predictive LTV has a plethora of use cases, including ROI-driven marketing automation, long term accounting forecasts, and user re-engagement (CRM). This article discusses why predicting LTV can be challenging and provides some insight on how it can be done effectively.

Challenge #1: Standard machine learning doesn’t fully solve the problem

Popular machine learning (ML) algorithms such as random forests, gradient boosted trees, and neural networks are tried and true approaches for finding complex patterns in data. 

Practical LTV forecasting pushes the limitations of these approaches, which require extensive training, large test sets, and long historical records. To outright predict day 365 LTV, a model would require access to a large number of users with known (actual) day 365 LTVs available, i.e., users who began using a product at least a year ago. 

Due to market changes, app updates, and changing UA strategies, the fundamental behavior of these older users can be different from new ones. Thus, a “pure” machine learning model by definition is always training on stale (years old) data. 

Figure 1: ML-based forecasts for long term (1 year+) LTV forecasts must leverage data from users who are many months old. Historical data may include product redesigns, pricing upgrades and promotions. And thus historical cohorts rarely represent the current behavior.

This may work for very stable products. However, most modern products experience frequent app updates, pricing changes, and market fluctuations. Thus, out-of-the-box machine learning alone doesn’t solve the problem. 

Additionally, many products yield significant revenue from users who convert days or months after install (i.e. mobile games), or are “whale” driven with a small percentage of rare users driving the bulk of the cohort’s revenue.  

ML models, in the context of LTV, also require backtesting to validate (i.e. training a model on a historical cohort, using it to predict a future cohort, and then checking against that cohort’s actuals). This procedure can be a lot more time consuming than typical cross validation within a single training set, and also makes the assumption that past performance is indicative of current performance. 

The graphic below describes the multiple submodels of our in-app purchases (IAP) forecasting system, which predict LTV for paying vs. non-paying users, and for the short term vs. long term. The individual outputs are then combined to produce a single LTV number for a cohort at a given horizon (number of days since that cohort began using the app).

Figure 2: An optimal approach to LTV modeling

Each submodel can have multiple theoretical foundations based on their applications, accuracy requirements, and scope. We’ve achieved success in LTV modeling with the following principles:

• ML: We use ML models when one can access a large quantity of relatively recent training data; for example, predicting subscription to a product at the end of a free trial. Recency is key because products undergo frequent updates and changes to monetization. Any model that relied on year old or older data to train would not be robust to this.
• In-app events: Measuring in-app engagements can help predict LTV, especially for non-paying users and young cohorts. Knowing how often someone is engaging with an app, or whether they completed a certain series of events, gives models otherwise unknown insight into their purchase conversion probability.
• Model training data: Larger data sets are usually good for ML. However, we also dynamically and automatically choose what data to ignore. For example, holidays and promotions are automatically removed from training sets if those cohorts are significantly different.
• Bayesian methods: Parametric models may not be as sophisticated as their ML counterparts, but provide more extrapolation power. Bayesian approaches explicitly model heterogeneous or evolving user behavior and better quantify uncertainty. Additionally, they allow the flexibility to supplement predictions for newer apps with “priors” from comparable products.
• Secondary models: We use cohort level LTV models that have a different set of inputs and assumptions to validate our aggregated user level projections. When both models line up, we can be more confident in the results; when they don’t, it’s indicative of possibly incorrect assumptions being made, or changes in dynamics that weren’t captured.

Under-scoping the data science (and data engineering) resources and skills sets required to build a functional LTV system seems to be one of the major pitfalls for companies trying to build this tool in-house. If it is built, the challenge becomes validation and adoption by users, and the most common sticking point is forecast accuracy.

Challenge #2: Everyone wants 100% accuracy all the time

Accuracy is hard! It is useful to break down the problem into two parts: 

• Systematic bias: A systematic over or under prediction in a category (i.e. iOS vs Android users)
• Variance: Sample noise remedied by larger samples

Figure 3: Example of (left) unbiased vs (right) biased model. Both models have the same variance due to underlying data noise. However, the biased model undervalues iOS and overvalues Android due to modeling error. 

It is hard to avoid variance, especially in behavioral data (i.e., whales, product updates, data issues). Bias, however, can be reduced by identifying and addressing modeling flaws that remain present over a large sample size. To address bias, consider how your modeling system approaches: 

• Categorical differences: Overvaluing a country platform or source
• Temporal changes: Overvaluing older installs by ignoring changing monetization or conversion dynamics
• Age: Overvaluing younger users compared with older ones 

A good LTV forecasting system will have three key performance goals:

• Globally accurate: We aim for less than 10% error at day 365 when up to a month’s worth of users are aggregated to support confidence in high-level ROAS forecasts.

  These forecasts are used to make the most consequential business decisions (i.e. deciding next month’s marketing budget) and so require the highest level of accuracy to ensure the best decision is being made.

• Unbiased over country/platform/channel dimensions: Marketers run campaigns across these dimensions; therefore, it’s important that our models correctly capture any LTV differences between them. If not, it can lead to biased decision making, where UA spend is incorrectly distributed across platforms or countries because LTV predictions do not fully account for the heterogeneity across them.
• Directionally correct at the campaign level: We want projections at the campaign level to drive intelligent decision making in the long term.  Thus, accurate forecasts from early user behavior and low volume are necessary.

Figure 4: (Left) A histogram of LTV error for a hypothetical model with 10% global all accuracy. The orange slice represents a sub-set (for example tier 1 users only) which may have  17% average bias. (Right) Some models may predict a campaign score which increases with increasing LTV.

Posing the right question on accuracy in the right way (model bias and sample variance) is almost as hard as building a model to answer it. When assessing accuracy, think about your goal:

• For Return on Ad Spend (ROAS) forecasting or marketing automation, low sample size will cause large margins of error for each campaign’s ROAS. Therefore, think about an aggregate portfolio of campaigns where the overall ROAS has some bias and risk. Are you making the right portfolio decision on average over your entire spend?
• CRM also leverages user level forecasts, but the absolute number is not as important as the relative rank. When working with this goal in mind, also consider relative or directional accuracy. Is the true stack rank of day 365 revenue well predicted by LTV?
• Corporate financial models consider entire geos or sources and thus have low sample variance due to high volume. Here, being unbiased toward a specific category is most important. Are you over predicting your Tier 1 users?

Knowing the internal client’s needs and risk tolerance is essential for correctly directing R&D efforts, model improvements, and validation. Poor collaboration between end-users and the data science team can leave a partially or fully developed data science product mired in “accuracy” land. The next section provides a foundation for starting the conversation.

Challenge #3: Building models to satisfy many use cases

Once a user level LTV prediction is made, it can be used by a variety of corporate stakeholders. Each audience requires different levels of accuracy and granularity. It is critical to understand the final use case when scoping and designing an LTV model. While it is hard to satisfy every business need, ensuring the system is human interpretable may help development and adoption of an LTV forecasting system:

• Stability: What is a reasonable forecast update frequency? Frequent changes may signal inaccuracy. However, ignoring real surprising behavior that strongly changes LTV is also undesirable.
• Accuracy: What is an acceptable level of variance (sample size) and bias (model sophistication)?
• Temporal granularity: How soon after install does the forecast come? Is a cohort a week or a month? Different teams care about different temporal granularity and responsiveness. 
• Cohort granularity: Does your end user care about individual users, all users, or a country? 

Table 1 illustrates some common use cases and requirements for LTV predictions as they relate to stability, accuracy, and granularity.  

Table 1: Use cases of an LTV model and the requirements for each

Summary

Predicting user-level LTV is challenging and requires months, or even years, of dedicated data science and engineering efforts to set up robust, accurate production systems. However, the business value provided by user-level LTV predictions is undeniably significant, offering solutions to a variety of use cases. Once an LTV forecasting system built upon solid methodology is put in place, it becomes a highly beneficial tool that can be used to quickly judge the success of marketing efforts, and to automate intelligent marketing decisions.

To read more about predictive modeling for apps, check out AppsFlyer’s guide:

The post Pitfalls of Modeling LTV and How to Overcome Them appeared first on AppsFlyer.

Last July, we announced the new Cohort Report and, as promised, we continued developing new features and enabling new insights to give you a deeper look into how your users engage with your app.
We’re now excited to bring you two new updates that will help you make the most of Cohort Reports.

All of your performance and ad spend data in one place

It often happens that marketers spend money on acquiring new users and end up targeting existing ones (and vice versa!). This overlap can cause quite a few issues when trying to analyze campaign performance. In order to accurately analyze cohorts, marketers need a tool that gives them enough flexibility and granularity to tease out the right insights and help them make more informed decisions about their campaign optimizations.

With the new Unified view, both UA and retargeting managers can now get an accurate breakdown of their cost and campaign performance data for both new users and existing users, all in one place.

Having a unified view of all the critical data about campaign performance in one place gives marketers the ability to see their true campaign ROI and make informed decisions every step of the way.

Optimize your cohorts by specific KPIs

Marketers across verticals and industries can benefit from Cohort Reports with a vast selection of groupings and KPIs. Whether you’re checking the performance of a weekend shopping campaign or assessing retention according to booked rides, AppsFlyer’s Cohort Reports has the tools you need to understand how your campaigns are performing over time.

But, what if you want to drill down and find out how a specific KPI performs over time?

With the new KPI by Attribution Time view (what we call “trend type” in the AppsFlyer dashboard), users are grouped by conversion time (date).This enables you to evaluate campaign performance over time and to view campaign KPIs relative to the conversion time (date).

Here’s an example:

An advertiser wants to compare the performance of two email retargeting campaigns. Emails are sent on Mondays and Fridays. In order to analyze the performance of these two campaigns, let’s say on day 3, the advertiser can now select the new view to see how much revenue was generated on that day on the y-axis and how both campaigns performed over time.

The result concluded that the ‘Friday’ campaign had a stronger start but soon both campaigns matched performance and displayed a similar reduction towards the end of the month.

Having a new view on how campaigns perform over time enables new insights and allows marketers to optimize their campaigns according to a specific KPI.

Stay tuned for more updates!

Take me to AppsFlyer Cohort

Not an AppsFlyer user yet? Get started today!

The post Do more with AppsFlyer’s Cohort Reports: New insights on performance and campaign ROI appeared first on AppsFlyer.

Your MWC 2020 networking plans start here!

Hola, que tal?

Mobile World Congress is almost here! Are you ready? MWC is the Superbowl of conferences. The World Cup of the mobile world. The Tour de France of tech innovation. You get the point. This isn’t the kind of event you just show up to. You need a game plan for MWC that includes both the show and the networking events, since that’s where a lot of the connections and deals are made.

I’ll leave the show plan to you, but it obviously should include a stop by the AppsFlyer booth – just look for the lighthouse (yes, lighthouse). For the networking part, in what has become my yearly tradition, I’m here to give you a breakdown of the best events and parties MWC has to offer.

Take a look, hit up as many as you can, and don’t forget that according to extensive (personal) research the best cure for a hangover is more alcohol.

Sunday, February 23

18:00-22:00 – Mobile Sunday 2020 with Tech.eu

Antiga Fàbrica Estrella Damm, 515 Carrer del Rosselló

5 times in a row! Mobile Sunday is back, once again. The traditional event takes place the day before MWC kicks off. As always, it’s hosted 4YFN (this time with xside). If you’re in town already, this should cover your first night in Barcelona.

Get your ticket here.

Monday, February 24

17:00-19:30 – MEF Connects

MWC Venue, FIRA, Hall 8 Auditorium – Theatre D

This event takes place at the end of MEF’s Summit and is conveniently located inside the MWC venue at Hall 8. Many of the Future of Mobile Summit attendees will be there.

Reserve a spot here.

18:00-21:00 – Startup Grind’s Corporate Innovation Summit

Antiga Fàbrica Estrella Damm, 515 Carrer del Rosselló

Startup Grind’s first of three events during MWC. This one will feature professional major brands like Asics, FC Barcelona, and Vueling, who will discuss the development of innovation in a corporate environment. Two panels will be followed by dinner and a networking session.

Get your ticket here.

19:00-22:00 – IoT Stars MWC 2020

Antiga Fàbrica Estrella Damm, 515 Carrer del Rosselló

IoT keeps growing and so does this event, held for the 6th annual time. Great networking for anyone from the IoT or connectivity sector. Full of professionals and experts.

Get your ticket here.

19:00-21:00 – The App Marketing Afterwork 2020

Jaleo Barcelona – Tuset Street, 5 Barcelona

This great meetup will host some of the leading app marketing experts for an evening of networking, food, and drinks. If you’re in the ad tech or the mobile marketing industry, this event hosted by PickASO, TheTool, Smadex, Splitmetrics, SearchAdsHQ and Interceptd – is a must.

Request your invite here.

19:00-22:00 – Transparent, AdTech party

Secret Venue

Tappx and Google Cloud partnered to bring you an AdTech cocktail party. If you’re a mobile marketer, this may be a great pre-party for the main event of this night (yes, the Mobile Monsters party. Keep reading). The bar hosting this party was featured as one of the best of 50 bars in the world.

Get your ticket here.

21:00- 01:00 – MWC Mobile Monsters

The Sutton Club, 13 Carrer de Tuset St.

By now, you probably already have friends who’ve been to our parties. This year it’s going to get crazier than ever. Just wait and see what we have in store for you. We partnered with App Annie, Mixpanel, Remerge, and Airship, to bring you a mind-blowing, monster savvy, dance party that will keep you buzzing long after.I’ll be there, with bells on.

Request an invite here.

Tuesday, February 25

12:30-14:00 South of France networking cocktail

MWC20 French Pavillon Hall 5, Stand 5b61

If you want to take a nice break in the middle of a hectic MWC day, this is a fine option. head over to the French Pavilion at Hall 5. and have some drinks with tech companies from the southern region of France.

Request an invite here.

17:00-18:30 AppsFlyer’s Happy Hour

MWC20 Hall 8.1 Booth #F41

This is another one of our traditions at MWC. If you haven’t managed to find time to visit our booth during the day, stop by at the end-of-the-day and hang out with us. This may or may not have music, free drinks and the majority of the people at Hall 8.1.

OK, it may.

19:00-01:00 – VC Night After Mobile World Congress

Antiga Fàbrica Estrella Damm, 515 Carrer del Rosselló

The second Startup Grind event is already a tradition at MWC, hosting many VCs and featuring an open mic session. A great networking event, good food, and wine guaranteed.

Get your ticket here.

Wednesday, February 26

17:00-18:30 – AppsFlyer’s Happy Hour

MWC19 Hall 8.1 Booth #F41

You made it this far, so come grab a drink and chill with us to celebrate the end of day 3. Friends, music, drinks, love.

18:00-20:30 – Digital Sapiens Unicorns & Stars 4YFN Party

Carrer dels Comtes de Bell-Loc 161 Les Corts

The 4YFN end-of-show networking event with media professionals, investors, entrepreneurs, and recruiters to mingle and network. This event enables participants to make valuable connections.

Space is limited so, hurry up and request a ticket.

18:30-22:00 – Startup Grind – Chiara Massironi fireside chat

ITNIG, 100 Carrer de Pujades

The final networking event by Startup Grind will have a fireside chat with Chiara Massironi, who is heading Twilio startups in East US and EMEA. This event will offer an open-mic session with a nice networking event to cool-off after a busy third day at MWC. As always, with Startup Grind’s events, you will enjoy a very good networking and dining experience.

Get your ticket here.

19:00-22:00 – VR/AR Association Executive Dinner

Hotel H10 Casa Mimosa, Carrer de Pau Claris, 179

If you want to mingle with some VR & AR professionals, this is the event for you. This VIP dinner has already confirmed attendees from Microsoft, Nestlé, Nielsen, Pico-Interactive, Visyon, Mediapro, and VRARA.

Buy your ticket here.

20:00-23:00 – Quobis Tapas Party – MWC

Carlitos Restaurante, 50 Calvet

Quobis annual party. If you want to enjoy the end of MWC with a glass of wine and some good tapas – this is the place.

Register here.

MWC is going to be super packed this year, just like any other year. With all these long walks from hall 1 to hall 8.1, planning your time is critical. Set the bar high, come energized, set your calendar in advance, and have fun along the way. Go with whatever’s worth your professional time, and personal party preferences.

Hasta luego, amigos!

The post Navigate your way to the best places to eat, drink, and party in Barcelona appeared first on AppsFlyer.

Unsurprisingly, the market today is facing a serious challenge around the standardization of data. Without being able to report properly and holistically on key performance indicators (KPI’s), marketers are making serious budget and optimization decisions based on incorrect or partial information.

Cost reporting is a zero-sum game. Obtaining inaccurate and partial data is equal to not having the data to begin with. Without it, marketers struggle with assessing the true ROI of their campaigns, optimizing efficiently and deciding how to allocate their budget. This is not an acceptable status quo. Every performance marketer should be able to have complete and accurate cost data side-by-side with their trusted attribution data.

In an ideal scenario, any partner or network could report accurate cost data regularly, and at the same granularity and dimension throughout, so that both sides (networks and advertisers) could benefit from an optimized marketing plan. In reality, extracting, gathering, funneling, and analyzing the data from every source and at any dimension is a daunting challenge, to say the least.

Luckily, we like challenges here at AppsFlyer. We have set out to solve this major pain point and are proud to provide a single-vendor solution.

Extracting Your Cost Data From Any Source

First and foremost, there is a need to extract data from hundreds of sources. The second part involves the method; there are multiple different methods of extraction cost, and all must be supported. No two sources report data in the same way, at the same level of granularity and frequency; which means extracting what you need at the right time, is no simple task.

As such, there is no accepted standard for campaign naming conventions, click URL structures or types of dimensions; therefore, comparing hundreds of networks’ performance and cost side-by-side is not easy. For example, each network has different types of metrics associated with cost: twitter has tweets, Facebook has page likes, Snapchat has swipe ups, and so on.

Building upon thousands of existing partnerships and integrations allows AppsFlyer to provide a strong starting point for any advertiser to extract the data they need using various methods such as click, API or our Ad Spend Ingestion feature.

Cost data extraction can be further complicated if campaigns are reported on different dimensions that cannot be processed by your system. For example, if your campaign is reporting cost based on geographic location, but you are unable to process data based on that dimension, the insight becomes irrelevant to you.

In addition, some networks report cost data on the campaign level, while others provide it at any structural level, which is why a good cost reporting solution should be flexible. AppsFlyer is able to support cost data at any dimension or level made available by the partner network

Using the Ad Spend Ingestion feature, advertisers can reprocess incomplete or incorrect data and also have the flexibility to ingest additional sources of data (influencer channels, email marketing, push notifications, etc.).

Our main mission is to ensure that as the preferred single-vendor solution, our clients will be able to view the data from any source, at any granularity, tied to AppsFlyer attribution even if data was missing, lagging or inconsistent in the past.

Owning the Data Management Flow

AppsFlyer allows advertisers to own every element of their data pass, no matter the source or method of extraction. Relying on other networks, partners, or channels to pass data regularly, and in time for advertisers to make critical decisions is nearly impossible today. Even if the process is improving over time, the advertisers benefit from owning not only the permissions to access the data, but also benefit from joint collaboration between the network, publisher, and measurement provider.

With AppsFlyer’s dedicated partner access permissions, teams can collaborate on identifying gaps, lags, and inconsistencies, easily allowing them to correct the data on the go.

Take for example a scenario where a network has issued a rebate, and an advertiser wants to now reflect the true cost for that channel in reports and ROI calculation for a given period of time. Now, think about how much time is lost to communicate the gap, then process the data correction. With AppsFlyer’s cost reporting solution, in one easy step, the true numbers can be updated and reflected in the UI instantly.

ROI Optimization: Going From Insights to Actions

Being able to rely on your data ushers in much-needed peace of mind for marketers because now they can calculate the true ROI of their performance and make strategic optimization decisions they can trust.

AppsFlyer provides ROI reporting to help app marketers measure the effectiveness of their app campaigns. By combining in-app activity and lifetime value data with ad cost and other campaign details, AppsFlyer can deliver real-time ROI reports on app installs.

This data is also available in various places including the Overview Dashboard, Cohort Dashboard, Pull API, Pivot and Custom Dashboards to name a few…

In Conclusion

The solution is two-fold, to build the most reliable app attribution platform that can collect, organize, and standardize all cost reporting using different methods of collection; and to do so under one roof.

The second is to leverage powerful insights and take immediate action, ensuring that you are always ahead of the game when it comes to performance optimization.

Since AppsFlyer is the leader in attribution with an unprecedented number of connected partners, we provide a unique position to glue the missing pieces together and report on ROI tied to attribution as a single vendor solution.

If calculating true ROI is the holy grail for marketers, then we are proud to say that we have come one giant step closer to achieving just that.

The post AppsFlyer’s Cost Reporting Solution: Solving One of the Biggest Digital Marketing Problems appeared first on AppsFlyer.

It is no surprise that Mobile World Congress is one of, if not my absolute favorite time of year (I mean, I even wrote a blog about it!). 

From the insane booths we are able to create year after year to the growing and more culturally diverse group of people we have joining us every time: MWC is pure magic for me and everyone at AppsFlyer. 

Mobile World Congress is the largest event in the global telecoms industry and, as such, we were looking forward to welcoming many thousands of visitors to our exhibition stand and social events during the week. As a people-obsessed company that places the best interests of its customers, partners, and people at the heart of everything we do, the potential risk of exposure to the coronavirus for attendees is not one the company is going to take. It is for these reasons that we have decided to withdraw from this year’s MWC. 

It is important for me to say here, that even within this extremely tough decision, I have learned so much about our team and even more about AppsFlyer’s “People-Obsessed” mantra than I have during years when we were able to bring this to the forefront in Barcelona. I’ve seen team members not get upset about missing the event itself, but about missing a chance to catch up with their customers and prospects. I’ve seen them check-in and offer help to the Marketing Department as we take a step back and withdraw from this year’s conference. It is one thing for me to say we put people first, but it is another thing for me to see it day in and day out. 

Last year I described this magic, and the joy I get to experience as a CMO as a roller coaster. Our journey as marketers is a roller coaster of exciting creative peaks as well as ‘boring’, day-to-day tasks. In this journey, it’s important to keep the balance, be true to what you really believe in and execute really really well on those creative peak moments. Even when those creative peak moments come at a time when you think there’s no way that plan A will happen. 

I still believe in that, even in times when we have to change our plans unexpectedly. 

The decision to cancel our attendance was made with a heavy heart, but with the welfare of many people in mind. Not just the fifty-plus members of staff who were scheduled to represent us in Barcelona and their many colleagues across the world, but our clients, customers and friends in the industry as well. While it is a tough decision to make, we know we’re doing what’s right given the available information. We look forward to coming back again in full force next year. 

The post What Taking A Step Back Taught Me About Going ‘All In’ appeared first on AppsFlyer.

Back in 2017, Cisco estimated that the average consumer in North America had 8 different connected devices. This number may sound high at first, but just think about it for a minute: this includes mobile phones, tablets, laptops, smart watches, smart TVs, virtual home assistants, and the whole world of IoT. It is expected that this number will grow to 13 devices by 2022.

While consumers have been enjoying the benefits of hyper-connectivity, marketers have been struggling with this new reality. When a single user has multiple devices and uses multiple platforms, the user’s identities are fractioned and siloed throughout. Brands are faced with only partial transparency into the user and his journey, making it far more challenging to create precise, targeted brand messaging along the funnel. 

In an attempt to solve this, marketers are faced with two options: creating  broad, repetitive messaging to cast as wide a net as possible; or employing a cross-device, cross-platform attribution solution that can connect the dots to reveal a single user behind them and their conversion journey. 

While there’s really no question that the second option—a people-based attribution solution—is the better one, this isn’t the end of the story. It’s not a beautiful golden-brick road all the way to personalized, well-timed messaging from here on out. The reason for this is that the technology lying underneath some of these solutions can be a privacy nightmare.

I’ll show you mine if you show me yours

Cross-device, cross-platform, cross-channel attribution technology is not an easy puzzle to solve. There’s a reason why there are very few providers on the market offering this. You might think that the biggest challenge is threading together the fragmented data pieces to create an accurate picture of the user behind them. Accuracy is crucial for people-based attribution, sure, but the true challenge here is reaching this accuracy without stomping all over user privacy.  

There’s an easy way out: pool the data. The more data at hand, the easier it is to connect the dots. If Brand A has one piece of data about a user, Brand B has another piece of data about the same user, and Brand C can help complete the picture of the user identity, one can just grab all these data points from three different brands and connect them. In return for their data contribution, Brands A, B and C get full visibility into the completed puzzle, essentially giving them access to insights based on each other’s data. 

Some companies do exactly that: create a shared persona graph, where data is tied together across brands to create a clearer image of the people behind the devices. By pooling the data together from their entire customer base, they can use John’s data to answer Mary’s questions. 

The temptation is clear and the benefits are obvious; this data crowdsourcing approach undoubtedly produces accurate identities and an accurate attribution funnel. The process of pooling and sharing data as described above, however, is invasive and presents significant privacy challenges especially in meeting the requirements of the GDPR and other global data protection regulations. 

If it walks like a duck, quacks like a duck…

In the GDPR realm of data privacy roles, there are two very different entities: data controllers and data processors. In short, the data controller determines the purposes and the means by which personal data is processed. A company collecting and  processing personal data is considered a data controller. A data processor, on the other hand, processes data on behalf of the controller.

The CCPA has similar roles, where data controllers are referred to as “businesses” and data processors are “service providers”. In both cases, one party determines how data is collected and carries out the collection process, whereas the other party is only entitled to process or analyze the collected data.

In the realm of digital marketing, the brands are effectively data controllers (or businesses), in the sense that they are gathering data across their digital properties. Their attribution solution, a third-party data processing solution, should be just that — a data processor.

Brands paying an attribution provider to pool their data into a shared persona graph, that other brands are also paying into, are effectively selling end user data to third parties. This is where things get uncomfortably murky; the minute an attribution provider becomes a vendor of data, they’re shifting their data privacy role under GDPR and CCPA. And in most cases, they’re doing it under the radar, on the margins of legality, without being truthful to their customers about it.

So how does a company that defines itself as GDPR-compliant, as a processing third party that doesn’t sell data, suddenly make such a shift

You might note that the providers that offer customers pooled data services are sorely lacking on their company-level privacy efforts, such as privacy compliance programs, external audits and certification. A company that disregards privacy on such a fundamental level is not one you want to be in business with, and definitely not one you want to be giving data access to. 

There are other ways

Your data is your data. It shouldn’t be pooled, sold or otherwise shared outside of your brand. If you are choosing to work with a vendor who does this, make sure they’re not misrepresenting themselves and their data privacy practices.

Shared persona graphs are not the only way to tie together user identities, they’re the easy way out for attribution companies with limited technology or a blatant disregard to privacy. 

Private persona graphs achieve the same level of accuracy, without compromising the inherent need for privacy. With private persona graphs, data is not commingled among customers, but tied together within the brand’s marketing properties. By employing multiple methods for connecting pieces of the user identity, private persona graphs provide marketers the desired effect without selling or sharing data outside of the brand.

Of course, privacy isn’t just about the graph. As mentioned above, you need to dig deeper to research the company’s general approach to privacy. This is true for any vendor you work with, not just data management platforms.

When working with private graphs, you can rest assured that your attribution provider (and you, by extension) are respectful and proactive about protecting your users’ privacy. You can also sleep better at night, with your conscience and business integrity intact.

The post On Persona Graphs, Privacy and Responsibility appeared first on AppsFlyer.