COVID-19: Mobile Ad Traffic & ROAS Up, CPMs Down, Opportunity To Scale

The effects of the Covid-19 pandemic, while undeniably tragic, have resulted in an increased global need to stay connected online. This need to connect, to distract, and to occupy time while isolated at home has resulted in opportunities for mobile games and mobile apps across the online advertising ecosystem.

The “pause” created by people staying at home is allowing us to spend more time with close family and loved ones while setting aside time to stay connected online to our communities and interests. Some experts have even predicted that the current situation could lead to a baby boom early next year.

What’s happening to the stock market, travel, retail, and services businesses have been difficult, to say the least. But as some of those businesses have pulled back ad spend, it’s opened the ad marketplace for other advertisers to step up.

Notably, mobile app advertisers have a unique opportunity.

We’ve been watching trends for a couple of weeks, and it’s no hiccup: CPMs for mobile games and apps have decreased markedly.

We can all understand why: People are spending more time playing games, and advertising has reduced for COVID-19 impacted advertisers. This has created a significant buying opportunity for online companies.


COVID-Impact To Mobile App Ad Spend


We’ve seen this across multiple mobile app niches, and geographies and it’s been steady since early March starting in Italy.

To illustrate the scope of this opportunity, below is data collected by during the period between February 21 – March 23, 2020, it is a representative snapshot from over $100 million in global ad spend from top mobile ad networks compiled from 50 mobile app games and apps.

Data varies slightly from country to country, of course, but our data show countries that have been required to shelter in place, like Italy, appear to be temporarily less competitive for mobile apps and gaming advertisers than they were a month ago. We seem able to leverage learnings from the Italian mobile ad marketplace to forecast CPM reductions about to occur in other geographies where we expect a new shelter in place restrictions to occur in the future.  In the U.S. shelter in place is being rolled out on a state-by-state basis, that allows for additional geographic segmentation and time offsets.

Company Grouping #1: Italy, UK, US. CPM reductions vs Italy.

COVID-19 CPMs vs Spend

Company Grouping #2: Italy, Germany, Spain, UK US. CPM reductions vs Italy.

Representative ROAS vs CPM: CPMs decrease, ROAS increases.

Covid-19 ROAS vs CPM

This makes sense: When required to shelter in place, people seek a distraction with Netflix, YouTube TV, Hulu and of course, they play games and they spend a bit of money on entertainment.  Between Saturday, March 14, 2020, and Monday, March 16, 2020, the number of Disney+ signups more than tripled compared to the same period from the week prior, according to data provided exclusively to Forbes from streaming analytics firm Antenna.

And because one in three Americans now have a stay-at-home or shelter-in-place mandate as of this writing, the US ad market appears ripe for mobile app and gaming advertisers.

Ad spend and CPM patterns like the ones shown above do not happen often. In fact, we typically only see CPM drops of this magnitude for the 2-4 weeks following Christmas. As we’ve never seen the impact of a pandemic on CPMs, these cost fluctuations may not last more than a few weeks from the start of shelter in place in each geography (state or country).

However, if the current “stay at home” and similar protocols last for 10 or more weeks, these reduced CPMs may last all the way into early summer.

Or maybe not. With the wildcard of the economy and job insecurity, the ad spend crystal ball gets murkier. If severe financial distress sets in broadly across states and countries, some people may cut back expenses, including fun distractions.

No one really knows how long this opportunity will last. Because of that, if you can, we recommend an increase to your mobile ad spend to close out Q1 and front load as much ad spend as possible into early Q2. If CPMs normalize, at least you’ll have captured market share and net profit while it was available.


Suggestions for Advertising During COVID-19


As you know, one of the best practices for advertising during the holidays is holiday-themed creative. It’s also a best practice to hook ad creative into cultural themes and trends.

You should avoid those overt messages with COVID-19. Maybe there’s a joke or two to be had around the toilet paper phenomena, but otherwise, people are probably playing games as an escape from what’s going on. So, consider softening your creative strategy according to these guidelines:

  • Stay upbeat, positive, and respectful.
  • Showcase your game or app as providing stress relief during the current situation (perhaps as an alternative to the news?)
  • Avoid direct mentions of the pandemic in general.
  • Don’t offer medical advice, guidance, or recommendations of any kind.
  • Keep things light and whimsical; avoid sardonic humor.
  • Everyone is experiencing COVID-19 in their own unique way. Refrain from creating ads that some people might consider offensive.
  • Fake news bulletins, breaking news imagery, or aggressive creative shouldn’t be used.
  • Make sure to tune up your creative testing best practices by reviewing our post, Why is the Control Video so Hard to Beat?

There is always opportunity in change and challenges. If you are in the mobile app and mobile gaming space, we believe the opportunity to scale during the current situation is one you should strongly consider.

Please reach out if we can be of help with creative services or user acquisition services.


COVID-19 Tracking Resources


Focus on What Matters: Here are the 3 Drivers of UA Campaign Performance

There are dozens of ways to improve campaign performance. Everything from the color on a call-to-action button to testing a new platform can give you better results.

But that doesn’t mean every UA optimization tactic you’ll run across is worth doing.

This is especially true if you’ve got limited resources. If you are on a small team, or you’ve got budget constraints or time constraints, those limitations will preclude you from trying every optimization trick in the book.

Even if you are the exception, and you have got all the resources you need, there’s always the issue of focus.

The focus may actually be our most precious commodity. Amid all the noise of day-to-day campaign management, choosing the right thing to focus on makes all the difference. There’s no point in clogging up your to-do list with optimization tactics that won’t make a significant difference.


UA Campaign Performance

Fortunately, it’s not hard to see which areas of focus are worthwhile. After managing over $3 billion in ad spend, we’ve seen what really makes a difference, and what doesn’t. And these are, irrefutably, the three biggest drivers of UA campaign performance right now:

  1. Creative optimization
  2. Budget
  3. Targeting

Once creative, targeting, and budget are working and aligned, your campaigns’ ROAS will be healthy enough that you won’t have to chase after every optimization technique you hear about for barely-noticeable improvements. 

Let’s start with the biggest game-changer:


Creative Optimization


Creative optimization is hands down the most effective way to boost ROAS, period. It crushes any other optimization strategy, and honestly, we see it delivering better results than any other business activity in any other department. 

But we’re not talking about just running a few split-tests. To be effective, creative optimization has to be strategic, efficient, and ongoing. 

We’ve developed an entire methodology around creative optimization called Quantitative Creative Testing. The fundamentals of it are:

Campaign Performance

Only a tiny percentage of the ads you create ever perform.

Usually, only 5% of ads ever actually beat the control.

But that’s what you need, isn’t it – not just another ad, but an ad good enough to run, and to run profitably. The performance gap between winners and losers is massive, as you can see below. The chart shows ad spend variations across 600 different creative assets, and we allocate spend strictly on performance. Only a handful of those 600 ads really performed.

campaign performance

We develop and test two core types of creative: Concepts and Variations.

80% of what we test is a variation on a winning ad. This gives us incremental wins while allowing us to minimize losses. But we also test concepts – big, bold new ideas – 20% of the time. Concepts often tank, but occasionally they do perform. Then sometimes, they get breakout results that reinvent our creative approach for months. The scale of those wins justifies the losses.

campaign performance


We don’t play by the standard rules of statistical significance in A/B testing.

In classic A/B testing, you need at least a 90-95% confidence level to achieve statistical significance. But (and this is critical), typical testing looks for tiny, incremental gains, such as a 3% lift.

We don’t test for 3% lifts. We’re looking for at least a 20% lift or better. Because we’re looking for an improvement that big, and because of the way statistics works, we can run tests in a fraction of that traditional a/b testing would require. 

This approach saves our clients a ton of money and gets us actionable results far faster. That, in turn, allows us to iterate far more rapidly than our competitors. We can optimize creative in dramatically less time and with less money than traditional, old-school a/b testing would allow.

Be flexible about brand guidelines.

Branding is critical. We get it. But sometimes brand requirements stifle performance. So, we test. The tests we run that bend brand compliance guidelines don’t run long, so very few people see them, and so there’s minimal damage to brand consistency. We also do everything possible to adjust creative as quickly as possible, so it complies with brand guidelines while still preserving performance. 

campaign performance

Those are the key points of our current methodology around creative testing. Our approach is constantly evolving – we test and challenge our testing methodology almost as much as the creative we run through it. For a deeper explanation of exactly how we develop and test 100x ads, see our recent blog post, Facebook Creatives: How to Produce and Deploy Mobile Ad Creative at Scale, or our white paper, Creative Drives Performance in Facebook Advertising.


Why It’s Time to Rethink Creative as the Primary Driver of Campaign Performance


Naming creative as the #1 way to improve performance is unconventional in UA and digital advertising, at least among people who have been doing it for a while. 

For years, when a UA manager used the word optimization, they meant making changes to budget allocations and audience targeting. Due to the limits of the technology we’ve had up until fairly recently, we simply didn’t get campaign performance data fast enough to act on it and make a difference during a campaign. 

Those days are over. Now, we get real-time or near real-time performance data from campaigns. So, while targeting and budget manipulations are powerful ways to improve performance, which you need to use with creative testing, we know creative testing beats the pants off both of them.

Google itself has acknowledged this, citing a study that found  “on average, media placements only account for about 30% of a brand campaign’s success while the creative drives 70%.”


But that’s not the only reason to get laser-focused about optimizing creative. Possibly, the best reason to focus on creative is that the other two main drivers of UA performance – budget and targeting – are becoming increasingly automated. The algorithms at Google Ads and Facebook have taken over much of what used to be a UA manager’s daily tasks. 

This has several powerful consequences, including that it levels the playing field to a large extent. So, any UA manager who had been getting an advantage thanks to third-party ad tech is basically out of luck. Their competitors now have access to the same technology. 

That means tougher competition, but more importantly, it means we’re shifting towards a world where creative is the only real competitive advantage left. 

All that said, there are still significant performance wins to be had with better targeting and budgeting. They may not have the same potential impact as creative, but they have to be dialed in or your creative won’t perform as it should.




Once you find the right person to advertise to, then half the battle is won. And thanks to fantastic tools like lookalike audiences (now available from both Facebook and Google), we can do incredibly detailed audience segmentation. We can break audiences out by:

  • “Stacking” or combining lookalike audiences
  • Isolating by country
  • “Nesting” audiences, where we take a 2% audience, identify the 1% members inside of it, then subtract the 1%ers out so we’re left with a pure 2% audience

These sorts of super-targeted audiences allow us to optimize campaign performance at a level most other advertisers can’t do, but they also allow us to avoid audience fatigue for far longer than we would otherwise be able to do. It’s an essential strategy for maximum performance. 

We do so much audience segmentation and targeting work that we built a tool to make it easier. Audience Builder Express lets us create hundreds of lookalike audiences with incredibly granular targeting in seconds. It also allows us to alter the value of certain audiences just enough so that Facebook can better target the super-high value prospects.

[su_vimeo url=”″ width=”640″ height=”560″ title=”Audience Builder Express: Building Facebook Audiences in a Flash”]

While all this sophisticated audience targeting helps performance, it has one other benefit: It lets us keep creative alive and performing well for much longer than without our advanced targeting. The longer we can keep creative alive and performing well, the better. 




We’ve come a long way from bid edits at the ad set or the keyword level. With campaign budget optimization, AEO bidding, value bidding, and other tools, now we can simply tell the algorithm which types of conversions we want, and it will go get them for us. 

There is still an art to budgeting, though. Per Facebook’s Structure for Scale best practices, while UA managers do need to step back from close control of their budgets, they do have one level of control left. That’s to shift which phase of the purchase cycle they want to target. 

If a UA manager needed to get more conversions so that the Facebook algorithm could perform better, they can move the event they’re optimizing for closer to the top of the funnel – to app installs, for example. Then, as the data accrue and they have enough conversions to optimize for a more specific, less frequent event (like in-app purchases), then they can change their conversion event target to something further down the funnel. 

This is still budgeting, in the sense that it’s managing spend, but it’s managing spend at a strategic level. Now that the algorithms run so much of this side of UA management, we humans are left to figure out a strategy, not individual bids. 


Bottom Line, Creative is King! UA Campaign Performance is NOT a Three-Legged Stool


Each of these primary drivers is critical to campaign performance, but it’s not until you use them in concert with Creative that they really start to stoke ROAS. They are all interconnected. Ignore one, and suddenly the other two won’t hold you up. 

This is a big part of the art of campaign performance and management right now – bringing creative, targeting, and budgeting together in just the right way. The exact execution of this varies from industry to industry, client to client, and even week to week. But that’s the challenge of great user acquisition management right now. For some of us, it’s a lot of fun. 

Player Profiles Are Key to Dominating Facebook Ads in 2020

Player profiles and motivation. Why do people play mobile games? What motivates them to choose one app over another, or to click on one ad over another? What motivates them to continue playing a game rather than just playing it once?

The answer lies in a concept called “horizontal segmentation.” It’s an idea that remade the food industry a few decades ago and has remade several other industries since then. We believe it’s about to remake user acquisition and Facebook ads, too.


Pasta sauce, pickles, and Howard Moskowitz


The first champion of horizontal segmentation in consumer marketing was a guy named Howard Moskowitz. Moskowitz, a food researcher, and psychophysicist had been asked to find the perfect sweetness level for a new type of Pepsi. But after he dug into the problem, Moskowitz discovered there was no perfect level of sweetness for the new type of Pepsi. There were only perfect levels of sweetness.

Or, as Malcolm Gladwell explains in his TEDTalk on the subject, there was no perfect Pepsi – there were only perfect Pepsis.

Pepsi, unfortunately, wasn’t sold on this concept. Moskowitz had to keep pitching it far and wide, with tepid reception, until a pickle company gave his idea a spin. And Howard, true to form, found them not the perfect pickle, but their perfect pickles – zesty, classic, etc.

That was a good start, but the revolution of horizontal segmentation didn’t take off until Prego, a pasta sauce company, hired Moskowitz. Moskowitz, once again, went out and did a ton of research about how people felt about pasta sauce. He returned with not the perfect pasta sauce, but with the perfect pasta sauces.

Most famously, Howard returned with a recommendation to make chunky pasta sauce. No other company had been selling chunky sauce, and pasta sauce makers weren’t aware that people even wanted chunky sauce. But because Prego trusted Howard and the research that pointed to this unmet need, Prego launched a line of chunky pasta sauce. And went on to sell $600 million of it over the next few years.


Horizontal Segmentation for Facebook User Acquisition


So, what does all this sauce talk have to do with user acquisition? Everything, actually. Because in essence, Moscowitz had unearthed the following ideas:

1. People are not all the same.

2. While people are not all the same, if you study them in large groups you will find they tend to cluster around a certain set of profiles or preferences. Personas, if you will.

3. With enough data and data analysis, these preference clusters can be found.

4. If you develop products expressly tailored for each of these preference clusters, you can sell more stuff. A lot more stuff.


Gladwell explains in his TEDTalk that, for example, if his audience was to get together and make one perfect brew of coffee together, they would individually give that brew a collective average score of about 60 on a scale of zero to 100.

But if the group was allowed to break up into coffee clusters based on certain dominant preferences and was able to make coffee expressly designed for each of those preferences, the satisfaction score for those preference-driven brews would rise to about 75 to 78.

As Gladwell explains, “The difference between coffee at 60 and coffee at 78 is the difference between coffee that makes you wince and coffee that makes you deliriously happy.” (Gladwell 17:16).

So, to paraphrase what Moskowitz and Gladwell have explained and to bring it into the context of Facebook and Google user acquisition:

  • There is no perfect game. There is no perfect way to advertise or marketing for that game, either.
  • But there are perfect games. There are perfect ads.
  • Finding the persona clusters or player profiles in gaming for these games and ads (and more specifically, finding the motivations of these persona clusters) will let us create ads that make people dramatically happier and take the action we want them to take (click, download, purchase, or view ads).

Revolution or Evolution?

This is clearly an evolution – if not a revolution – in how most UA and design teams have been doing creative for user acquisition.

We’ve all been doing reasonably well by focusing on creative testing, but we’ve been largely focused on the creative elements of those ads that have the largest appeal (though good competitive analysis does take emotional hooks and the tone of ads into account). But most UA and design teams have been approaching their advertising and creative strategy with the assumption that there was a perfect game, a perfect app, a perfect ad. “A platonic ideal,” as scholars say.

What if that’s not so? If there are only perfect apps, perfect games, perfect ads.

What if we could talk to people based on what motivates them to play the game?

Maybe, we could take a video ad – the exact same video – and then adjust the copy or the call to action based on what motivates different users to want to play. By tailoring messages according to what those player profiles will most respond to, could we increase ROAS by… 20%? 30%? Or could we expand the game into a whole new audience segment?


Game Theory and User Behavior Models


If you know your way around game theory, you’ll know we already have quite a few detailed studies about different gaming profiles and user personas.

So, we actually have quite a lot of information about what motivates people to play games.

You can look to “Fogg’s Behavior Model,” or Scott Rigby’s Player Experience of Need Satisfaction (PENS) model. Or you can view your users through something simpler, like “The Big Five,” or the “Five-Factor Model.” Paula Neves breaks that model out well in her article Looking at Player Motivation Models:

  • Openness to Experience: If one is inventive and curious or consistent and cautious.
  • Conscientiousness: If one is efficient and organized or easy-going and careless.
  • Extraversion: If one is outgoing and energetic or timid and reserved.
  • Agreeableness: If one is friendly and compassionate or challenging and detached.
  • Neuroticism: If one is sensitive and nervous or secure and confident.

The five traits, under the acronym OCEAN, are therefore treated on a spectrum where you can be open or closed to experience, conscientious or unconscientious and so on.”

Bartle’s Player Types, in particular, might be the model that the gaming industry embraces the most.

Bartle’s Player Types

bartle;s player profile types
Image Credit: 2014

In fact, given the names of the player types and their descriptions in Bartle’s model, it almost seems like some of these player types have already filtered down into ad creative.

Ad Creative

player profile ad creative
Image Credit: 2020

Or maybe behavior models aren’t quite your thing. Maybe you want to view things through the lens of which types of emotional hooks appeal to different player types.

Emotional Hooks

gaming emotional hooks
Image Credit: 2020

As a UA manager or designer (or even a game designer, but that’s another article), you can start seeing your user base through any of these different models. And you can start customizing your advertising accordingly.


Examples of Player Profiles


Cutting-edge UA and design teams are already exploring these different models and applying them to their user base. When we polled our clients, we learned that about 10% of them have these sort of persona profiles in place. Typically, these models are being generated in marketing departments through the use of user surveys.

So, here’s what an actual player profile might look like. This is actually fake data, but it gives you an idea of what information gets included and how detailed these profiles tend to be.

player profiles
Image Credit: 2020


Note how the graphic shows what this particular player profile most cares about in the game. The Demographic info is included, but what will really shape our creative is in the left column. Demographics might help a little, but user motivations are the driver.

Competitive Analysis

So, we can take the motivations from these different player clusters and look at what’s unique about them. We can also see which other titles they’re playing. This helps us come up with new creative ideas and a new list of competitors, which in turn lets us do better competitive analyses.

Armed with this information, almost any designer would immediately understand that they need to create different ads for these different player segments. These are different audiences. Even if the ads are ultimately meant to sell the same game to all the different profiles, ads tailored to each profile are going to perform dramatically better than one ad designed to try to please them all.

Even being able to swap out different bullet points for each player persona could result in tremendous improvements to ROAS.


Media Buying for Player Profiles


There are also media buying implications to a player profile methodology. Each of the player profiles above, for example, tends to favor somewhat different websites, apps, and even YouTube channels. So if a user acquisition manager wanted to squeeze every possible drop of ROAS from an audience, they would segment that audience based on different player profiles.

If you’re acquainted with Facebook’s new Structure for Scale recommendations, you may have frowned a bit at that last suggestion. Per the Structure for Scale recommendations, Facebook would rather we not get too granular with audience targeting. Because the smaller a dataset we’re using for our ads, the less effective the algorithm will be.

But that’s where the art of UA management intersects with the science of the ad tools we have now. It’s up to each UA manager to find the sweet spot between audience targeting and audience automation.

It will take some experimentation to figure out whether it’s appropriate to segment audiences for player profiles. Or, if it’s better to just make different ads for the different personas and to let the algorithm figure out optimal performance.


People Can’t Always Tell You What They Really Want


Before you try to apply all this by launching a user survey or research study, know this: People may not be able to tell you what they really want. Because they don’t know.

Sometimes, you don’t know what you really want until you happen to stumble upon it.

For example, Moskowitz, the pasta sauce researcher, discovered that many people wanted chunky pasta sauce, not the perfectly smooth pasta sauce that was only available at the time. Only with some unusually clever research was Moskowitz able to unearth this desire.

Gladwell describes this elegantly by talking about coffee. As he explains, if you ask the typical American what type of coffee they like, they’ll say “I want a dark, rich, hearty roast.” (10:40). But in actuality, most of us don’t really want a dark, rich, hearty roast. Only about 25-27% of us actually like a dark, rich, hearty roast. Most of us like milky, weak coffee. But we’ll never, ever say that.

Product Research

All this, of course, turns standard product research on its head. It means all those surveys, focus groups and other tactics we’ve been using to figure out what people want from our games are not the full stories.

But this is exactly the principle that put Apple on the map. Steve Jobs didn’t go find out what people wanted from a better cassette player. He built them something beyond their imagination: The iPod.

Henry Ford is often quoted with a similar sentiment: “If I had asked them what they wanted, they would have said faster horses.”

So what we may be aiming for here is a quantum leap, not an iteration. Whoever can figure it out, systematize it, and apply it to game design and creative strategy could well become a billionaire. Or another Howard Moskowitz.


Player Profiles as UA Creative Strategy 2.0


For the past two years, we’ve watched Facebook and Google move towards fully automating UA advertising. And because the algorithms of those two ad platforms have been getting better and easier to use all the time, much of the qualitative side of UA management is now best done by machines.

Creative, however, is still best managed by humans. This includes all aspects of creative – creative development, creative strategy, and creative testing. Creative is our best competitive advantage. Especially, now that the ad platforms have removed the competitive advantage third-party adtech used to deliver.

But now, with player profile theory, there’s a whole new dimension to creative. It takes everything to another level.

Once we can see what’s motivating people to play games, we’ve discovered a whole new level of creative theory. It makes getting overly focused on button colors look downright shallow.

All the creative elements we’ve been focused on up until now (colors, sounds, even ad copy) need to be realigned. This will serve these customer motivations.

With player profiles theory and player motivations, we have crossed into Creative Strategy 2.0.



Read Our

Creative & UA Best Practices For Facebook, Google, TikTok & Snap ads.

    Please prove you are human by selecting the Tree.