Mobile Audience Segmentation Explained

Sponsored by Start.io

U.S. mobile ad spending crossed $200 billion last year. That’s more than half (51.2 percent) of total U.S. media dollars—and an impressive two-thirds (66 percent) of digital ad dollars. Of this spending, the vast majority is happening in apps.

That makes sense. After all, that’s where people are spending their time. But as with any advertising opportunity, success isn’t guaranteed. Mobile advertising isn’t just about showing up—it’s about showing up with relevance. And that’s where mobile audience segmentation comes in. 

The AdTech Explained Team developed this explainer, in partnership with Start.io, to help you better understand the current state of audience segmentation, particularly in the mobile realm.

Let’s start at the top: audience segmentation. What are the predominant methods being employed today?

The specific tactics that companies employ when segmenting audiences today are as nuanced as the audiences themselves. However, even as methods evolve alongside privacy regulations and platform policies, it’s still useful to think of audience segmentation approaches as existing in two buckets: deterministic and probabilistic.

Deterministic data is pretty straightforward: A company observes something specific and measurable—such as someone's geographic location or keyboard language—and builds an audience segment based on verifiable observations, such as "people who live in Anaheim, CA" or "English speakers in Southern California."

On the other hand, probabilistic data involves educated guesses. In other words, a company observes some data signals and uses them to make a guess about who a person might be or what they might like. For example, if a company can see that you live in a major city and you’re playing mobile games at 11 p.m. on a Saturday, it might guess that you’d be interested in ordering some snacks and serve you an ad for food delivery services. Or, if the company knows you frequently visit a local gym and have downloaded a fitness app on your phone, it might guess that you like to exercise and serve you an ad for workout clothes. These kinds of inferences can be used to build audience segments. 

There are a lot of companies out there using first-party data (i.e., data they gather themselves) for audience segmentation. That data might include demographic, behavioral, psychographic, geographic, technographic, and firmographic data, as well as interests, life events, purchase history, and buyer journey segmentation. The specific type of data you employ matters. But it helps to start with the question of deterministic vs. probabilistic: Do you know? Or are you guessing?

OK, let’s get more specific. What is mobile audience segmentation? 

Mobile audience segmentation, sometimes referred to as device-based audience segmentation, is based on—you guessed it—data coming from mobile devices. Mobile audience segmentation can be both deterministic and probabilistic, depending on the data available and a brand’s campaign goals. Both methods deliver unique value. But without a doubt, the strongest mobile audience segmentation starts when you’re working with a robust set of first-party data. 

Why is the “mobile” part so important?

Just think about how your relationship with your own mobile device, and you’ll probably see why focusing on mobile audiences and data signals is so powerful when it comes to helping advertisers reach specific groups of people. Mobile devices throw off a tremendous number of signals, and those signals are typically tied to a specific individual in ways that other data signals aren’t. 

Mobile first-party data is especially valuable because it provides a surprisingly accurate picture of who you are and what you're interested in. Where you go, when you’re on your phone, which apps you own, your keyboard language—these signals communicate a lot of information about you without overtly divulging your identity. Used wisely and responsibly, mobile data is the key to delivering the relevant ads people want while protecting their privacy. 

What are the most common ways to segment mobile audiences? 

There are countless ways to segment an audience using mobile data. Broadly speaking, these are the most common: 

Geographic: A brand might segment by country, state, city, or down to a specific ZIP code. This is made possible by geographic mobile data users have provided in apps (e.g., Facebook users indicating they live in San Francisco). 

Location-based: These mobile segments are based on the device’s GPS signal, indicating where a user is at the time an ad can be served. For example, a brand might want to serve ads only when a user is within a specific radius of a storefront. Or, location can be used to create audience segments of business travelers who are commonly in roaming mode. 

Demographic: As with geography, information fed into mobile apps can inform demographic segmentation by gender, age, income, and more. 

Psychographics: Brands often want to reach people based on their values and interests, which can be inferred by mobile behaviors. For example, users who routinely install and engage with news apps or content tend to value knowledge and participate more in politics. 

Technographic: Sometimes the specifics of a person’s device matter in terms of the type of ads you use to engage them. Mobile audience segmentation can be applied to at the device tech level (e.g., building a segment of iPhone 16 users with devices that have at least 8 GB RAM).

Behavioral: Mobile audiences built on behavior—what users do on their phones, for how long, and with what results—are particularly powerful (e.g., mobile gamers who tend to engage with games for 20 minutes or more in a session and have made at least two in-app purchases).

What else should marketers know about mobile audience segmentation?

Segments can be incredibly granular, depending on the platform you’re using. For example, Start.io offers thousands of audience segments, from the very large (for example, 279 million social media users in the United States) to the very small (for example, 1,500 cricket fans in Anaheim, CA).

With mobile audience segmentation, data is the gift that keeps on giving. With every subgroup an advertiser creates, the brand can test to find the personalized messages that work best and continually iterate to find ones that work even better.

Even as certain mobile identifiers disappear from the landscape, there’s still a tremendous amount that can be done when you combine the power of deterministic mobile signals with AI to infer patterns and bring scale to the insights. 

What is Start.io’s role in the mobile audience segmentation space? 

Start.io is a mobile-driven SSP that specializes in publisher monetization solutions. Start.io’s platform delivers hundreds of millions of ads every day and is integrated with more than 500,000 active apps. Every time an ad request comes in, Start.io receives a small amount of additional, privacy-compliant data that's attached to the ad request to help the company find the best ad for the user. This can include data like location, device types, mobile operating system, keyboard language, and more. On the average day, Start.io processes billions of first-party mobile data signals. This information is useful in helping mobile apps deliver high-performing targeted ads based on accurate audience segments.

Learn more about mobile audience segmentation and how Start.io can help here.