To fight back against fake installs and dodgy attribution, advertisers need to ask better questions

• With install fraud so high companies need to stop relying on simple metrics.
• AI offers bad actors the chance to generate fake inventory more easily.
• 'Grey attribution' is the abuse of view-through attribution (VTA) and click-through attribution (CTA).

With mobile gaming revenues returning to growth after a few years of struggle, 2026 is poised to be a positive one.

It’s also going to be a big year for programmatic advertising, with global spend set to pass $200 billion for the first time, with mobile gaming continuing to make up the biggest portion. 

But it’s not all good news. Inevitably, whenever opportunities to make money arise, some individuals will seek to take advantage, whichis why ad fraud remains a persistent problem for our industry. 

It’s an open secret that digital ad fraud is endemic. Digital ad fraud is relatively easy to perpetrate and delivers high returns, and efforts to police it arerelatively weak.

All of this is likely to be compounded in the coming year by the growth of AI-generated apps, which offer numerous opportunities for bad actors to create multiple ‘slop’ apps solely as a means of generating (often fictitious) inventory to sell, as well as the ability to harness armies of AI bots for click fraud on a massive scale.

The combination of invalid traffic (IVT), fake attribution, and ad fraud is impacting programmatic advertising globally. Ad fraud intelligence company Pixalate’s Q3 2025 Global IVT Benchmarks Report found that a third of all mobile ad traffic is either invalid or fraudulent, based on its analysis of 106 billion ad impressions in Q4 2025.

I can’t think of many other industries where a third of your marketing budget is basically stolen with impunity.

An area of abuse that the team at Bigabid is especially focused on is the abuse of view-through attribution (VTA) and click-through attribution (CTA), which we call ‘grey attribution’ because, while the tactics used can’t be said to be deliberately fraudulent, they certainly don’t follow best practices.

VTA abuse happens when a DSP attempts to show an ad impression just before an organic install or return, which means that action is attributed to them despite the fact that they never influenced the user.

CTA abuse happens when a DSP records a click-based engagement even with ad formats which are view-only. Because clicks have a longer lookback window and higher attribution priority than other forms of engagement, it’s a direct way to ‘steal’ attribution from other vendors and/or organic installs.

There are other forms of grey attribution, and they are all quite complicated and hard to spot, even if you know what to look for.

So what can advertisers and publishers do? After all, the rise of programmatic advertising has hugely increased the complexity of how ad campaigns are delivered.

According to advertising giant Dentsu, over 71% of total ad spend will be algorithmically driven by 2026, rising to 76% by 2028. This means that app marketers need to be savvy about how their ad budgets are being spent beyond just ROAS.

Think about metrics beyond ROAS to understand where your installs are really coming from.

If the sources of ad fraud were easy to identify, then obviously advertisers would stop giving these bad actors a chunk of their budget. The problem is the difficulty in truly verifying attribution and clicks.

The positive for advertisers is that there are measurement metrics beyond ROAS, which can more accurately show you are getting bang for your buck in the short-term, but don’t really help advertisers with whether their campaigns are delivering the right users over the longer term. 

A lot of campaign measurement currently focuses on the scale cost of UA. This leads to what I call a spray and pray approach - you just spray your ads at placements that offer high volume at a low cost, and you pray that it brings you new players.

The problem is, this approach doesn’t really bring value to the advertiser, as it doesn’t discriminate between genuine new players and players that would have organically returned anyway.

Instead, I’d encourage advertisers to look at metrics that help to show user behaviour, to complement cost-driven metrics. So, looking at installs for a thousand impressions (IPM) or the conversion rate from click to install (CTI), or the time window between click and install (CTIT), are all metrics that can show whether the right users are seeing your campaigns and are acting on them.

If advertisers don’t ask the right questions, it’s an invitation for vendors not to care as much about quality installs as they may, perpetuating a cycle of wasted spend and poor return.

Another important metric is the number of returning users coming from deep links. When retargeting lapsed players, you expect them to click the ad and go straight to the app, as it’s likely to already be installed. That's called a deep link. A returning user is more likely to play and spend in-game, so the higher the number of deep links being attributed, the higher proportion of those quality players your campaign is reaching

Finally, make sure you are looking at performance metrics country by country, rather than grouping regions together. There can be a big difference in traffic cost and the value of acquired users, so it’s like comparing apples with oranges.

Different nationalities may react differently to ad campaigns for a whole host of reasons, including cultural, economic, behavioural and more. Plus, there is a huge variation in CPI cost between different countries and regions, with installs in Europe and the US easily costing four times as much as those in Latin America.

The more you understand the motivations of your target audience, the better your metrics will help you understand how your ad campaigns are really performing beyond just CPI and ROAS.

Choose your partners carefully, and try to prioritise quality over quantity

So, should advertisers simply be more selective of the vendors they work with and focus more on quality over quantity? Well, yes, but it’s not that simple. Cutting out the least effective and cheapest inventory pushes up CPI and lowers ROAS, but it’s unlikely to change the actual number of new or returning players. 

That’s because every campaign will end up attributing users who would have organically installed the game regardless, and changing the vendor or the ad creative won’t affect that.

I believe that it’s our responsibility, the mobile advertising ecosystem, to be upfront with advertisers about the shadier things that go on with ad campaigns.

At Bigabid, we’re providing advertisers with tools to detect when things are too good to be true. By doing that, the advertisers we work with understand the implications of chasing the lowest cost installs, so they can make informed choices.

After all, we all know that if something looks too good to be true, it usually is - and that applies just as much to mobile advertising as it does to anything else in life.

If we do collectively want to fight fake attribution and ad fraud, vendors need to increase efforts, and advertisers need to be asking more questions about where their installs are coming from.

 There will be more technologies that can help, such as AppsFlyer’s Protect360, but with AI making it so easy to create fake apps and fake users, we all need to be more vigilant if we are to stop the fraud juggernaut from taking a massive slice of the mobile advertising pie.