How App Tracking Enables More Powerful Lookalike Modeling and Targeting


For the last several weeks, we’ve spent a lot of time talking about how app tracking and mobile apps can help marketers segment and monetize current app users. Today we’d like to take a slightly different tack, and discuss how app tracking, and insights from thorough audience analysis, can drive more effective user acquisition.

There are a handful of innovations that have made a huge difference in digital marketing effectiveness over the past decade, and few have had more impact than lookalike modeling and targeting. Thousands of brands and tens of thousands of campaigns use lookalike modeling every day, whether or not the brand teams paying for the campaigns are even aware of it.

With lookalike targeting, a media provider models an acquisition target based upon the demographic, behavioral and other measurable characteristics of your current brand customers or advertising/site interactors. The provider uses salient characteristics of your customers, or characteristics that are overdeveloped in your target audiences, and scours the web to find more such individuals, under the (usually correct) assumption that such people will be more likely to interact with your brand in fruitful ways. It’s not a 100% perfect science, but it has been shown again and again to be rather effective as a means of cost effectively acquiring new customers – both inside and outside the mobile app space.

Different media companies often focus on different measures for lookalike modeling. For example:

  • Using social networks and relationships with customers as an indicator that an individual is more likely to respond and install or otherwise take a desired action. 
  • Focusing on demographic characteristics to predict likely brand affinities
  • Putting Android, iOS or Windows Mobile location data and consumer movement patterns to work identifying likely responders
  • Leveraging topical interests demonstrated via similar web content consumption habits as a means of estimating likely responsiveness to a value proposition. 

So, if the media companies are already using lookalikes, why do you need to implement app tracking to provide lookalike targeting insights? In other words, why track your apps – why pay for something – that media companies are probably already doing?

Well, for at least 6 reasons:

1. Because they’re your customers. It’s your data. As a marketer, do you want to know your customers, or cede that knowledge to others? Would you like rich insights into who installs your app, or to have partners keep that data in a black box? This is about fundamental fiduciary responsibility. You need to understand the people who use your Android, iPhone and Windows apps. You need to track what they do, inside and outside your apps. And an aside: if you aren’t hypercurious about who your customers are and what makes them tick, one might suggest that you are in the wrong field.

2. Because media companies are not incentivized to share app marketing insights. When you rely on media companies to implement advanced targeting strategies without understanding the specifics of what they are doing, you need to recognize that it is in their interests to keep the most effective insights to themselves. If you were a media company that had found some sort of targeting Holy Grail, how likely would you be to share it? Most of us don’t choose mobile app marketing media vendors by which are the most altrusitic but rather on which drive the best results per dollar spent on mobile app acquisition.

3. So you can base lookalike modeling on ALL your data and installers, for greater precision. Most media providers see only a small fraction of your total mobile app install acquisitions. Given that, they can base their modeling on only a tiny portion of the insights that could be available to them if you united all your data and extracted maximum information from it. Similarly, the ability to “see” more of the people who performed mobile app installs makes it possible for you to identify more statistically significant audience segments for lookalike modeling.

4. So you can model your most profitable iPhone and Android app users. When you control your customer data, you can draw distinctions between average installs and high performance installs, thereby providing the foundation for better lookalike modeling. It’s unlikely that a media company would do this, or that it would be in your interest to offer that sort of unlimited access to your data, no matter how important the partner to your business.

5. Because you can prudently limit vendor access to your customer data. By placing primary responsibility for targeting insights inside your own walls, you can limit the amount and type of data you need to share with media partners. Your data will likely reveal insights that can create competitive advantage – do you really want media partners to be able to offer those insights to all your competitors, or to use them to improve marketing performance for all competitors? Do you want learnings from your iOS or Android app to power marketing for the b*stards across the street?

6. So you can create cross-device app customer profiles and base your lookalike modeling on cross-device behavioral insights. By deploying a data management solution across your mobile applications — one that includes user-to-device or cross-device ID and cookie matching, your app tracking information can be enriched and enhanced with other insights that enhance lookalike modeling accuracy. You can know more about each individual, and based your lookalike modeling on that richer insight.

Best of all, the costs for mobile app measurement and attribution are almost always a teeny tiny fraction of your total media and marketing expenditure. By spending just a few percent of your budget on app tracking, you can radically improve the effectiveness of your customer acquisition efforts. To say nothing of how  mobile app attribution data can improve your ability to drive increased marketing effectiveness and to monetize the app users you already have.

Low cost. High benefit. That’s math any marketer can get behind.

Share Button