2025 Predictions Retrospective:
What We Got Right (and Wrong)

As we wrap up 2025, it’s time to look back at our 2025 Predictions, Premonitions, and Prophecies and see how they panned out. Last year (i.e. 2024) we were 90% correct. And given that our coming 2026 predictions write up is about to be released soon – it’s probably important for most of our readers to know how well we got it this year. 

So let’s review each prediction and assess the outcome with evidence from the industry.

Modeled Conversions Go Mainstream - Correct

Our prediction was that more and more ad platforms will deploy modeled conversion products to deliver results without relying on last-click attribution postbacks. We expected companies like Moloco, Liftoff, IronSource, and Smadex to follow this trend of using machine learning to model conversions and optimize campaigns.

Outcome: This prediction was spot on!. In 2025, the use of modeled conversions became widespread across the advertising ecosystem. Evidence? AppLovin share price and AXON2.0. The AppLovin system estimates conversions enabling a much better campaign evaluation and optimization. Meta’s approach to the iOS signal loss has also leaned heavily on modeled data.

The success of early adopters (Google’s AI-driven campaigns, Meta’s Advantage+ and Conversions API, etc.) demonstrated that modeling works. Advertisers have grown more comfortable trusting AI to estimate conversions (not that they have much choice…) and are shifting towards AI attribution engines to measure incremnetality instead of relying on last-touch attribution.

Optimization Based on Incremental Value - Correct

We predicted that some Ad platforms would start optimizing for incrementality - rather than chasing raw postback numbers. We foresaw platforms incorporating incremental lift metrics into their algorithms, moving beyond last-click attribution.

Outcome: This has indeed happened, with the most notable example being Meta’s new Incremental Attribution setting. In 2025 Meta began rolling out an option in Ads Manager to report and optimize on incremental conversions – conversions that would not have happened without the ad campaign. This is a fundamental shift towards efficiency and real value.

Advertisers are quickly adopting incrementality concepts. Over half of performance marketers now use incrementality testing, and 36% plan to invest more in incrementality measurement next year!
The inclusion of incrementality signals in platform algorithms confirms our prophecy. Even Meta’s own blog post acknowledges that this is “a step in the right direction” and a sign of where attribution is headed. We expect (ahhm…know!) Google, TikTok and others to introduce similar incrementality-based optimization features in the near future, validating that incremental value is the north star for campaign optimization.

“MMPs” Becoming “MPs” - Correct

Prediction: We predicted that Mobile Measurement Partners (MMPs) would dramatically pivot their offering. 

Outcome: This has played out vividly in 2025. The most striking evidence is AppsFlyer’s fall 2025 transformation. In November, AppsFlyer announced a major product expansion with eight new products and positioned itself not as a mobile attribution company but as a “Modern Marketing Cloud” built for the AI era (i.e. MMC – we were close).

During 2025, both Appsflyer and Singular went and did a quiet U-turn on MMM.

Kochava is launching an AI workspace for productivity, Adjust, one of the pioneering MMPs, suffered major layoffs under AppLovin’s ownership. Competitors openly noted that the status quo was falling apart and that to survive, attribution companies must change. 
In other words, MMPs can no longer just count installs; they need to prove value in new ways.

MMM “Back to the Attic” - Mixed Outcome

Prediction: We predicted that the hype around Media Mix Modeling (MMM) as a one-size-fits-all solution would fade.

Outcome: This one was partially correct. Indeed, MMM did lose its momentum as a fad. Many companies that rushed to adopt vendor-packaged MMM in 2022-23 found it cumbersome and not actionable for day-to-day decisioning. 

However, Marketers became more realistic about MMM’s limitations. It’s clear now that MMM requires strong analytics expertise, years of data, and frequent updates – it’s not a plug-and-play tool for quick optimizations. 

We are not anti-MMM, but MMM becomes a pain in our asses when INCRMTNAL gets compared to MMM solutions. INCRMNTAL is to MMM what a GPS is to a Compass.

Google’s Privacy Sandbox Delayed - Correct

Prediction: We were skeptical that Google would fully launch its Privacy Sandbox.

Outcome: Not only was this prediction correct, but reality went even further. Google not only delayed the Privacy Sandbox - it officially shut the project down in 2025. Google announced in October 2025 that it was killing off Privacy Sandbox entirely and halting development of its proposed tracking/targeting APIs. 

The end of Privacy Sandbox meant that, for now, the status quo of cookies continues almost uninterrupted.

iOS App Store Redesign - Half Correct

I predicted something based on no hard data points or insights - saying Apple’s App Store would get a major redesign, given it had been over 8 years since the last big redesign. We had no leaked info or rumors – just a hunch.

Outcome: Apple proved us kind of right by delivering an App Store redesign in November 2025 – not within the iPhone app, but on the web interface. Apple rolled out a completely new web App Store experience, making it far more functional for discovery and browsing. For the first time, users on the web can navigate the App Store by categories and device, essentially replicating much of the native App Store’s capabilities in a browser.

Final Thoughts

Overall, our 2025 prophecies held up well – the majority were proven correct by the events of this year. We accurately foresaw the industry’s shifts toward modeled conversions and incrementality, the evolution of the MMPs, and even Apple and Google’s platform moves. A few nuances were different (MMM didn’t disappear…YET!), yet the core directions we predicted were on target.

If nothing else, this retrospective reinforces that the only constant in adtech is change. Betting on fundamental trends served us well in 2025. As we look to 2026, we’ll carry forward the lessons from this year. 

Here’s to staying ahead of the curve in the ever-evolving marketing landscape.