eCPMs: Why first-party data is such a big problem.

Everyone who wants to increase revenue from in-app advertising speaks about first-party data as if it is the holy grail, as it is the only real solution to Apple's ATT.

As great as saying "we have to use first-party data" is - there are a lot of barriers which make it incredibly difficult to use.

What did Apple & ATT do that makes first-party data so important?

ATT gives users the ability to opt out of tracking, and tracking is defined as:

"linking data collected from your app about a particular end-user or device, such as a user ID, device ID, or profile, with Third-Party Data for targeted advertising or advertising measurement purposes, or sharing data collected from your app about a particular end-user or device with a data broker. “Third-Party Data” refers to any data about a particular end-user or device collected from apps, websites, or offline properties not owned by you."

Users who opt out of ATT don't have an identifier returned, which effectively means you know nothing about that user. All of the tracking goes away, but you can use your own first-party data to show personalized ads.

Also, if 50-84% of your users opt out of ATT (and these are some of your most valuable users) then there is a very clear incentive to figure out how to use your first-party data to get your monetisation of in-app advertising revenue back on track.

So, this poses a really interesting and big question;

"If everyone says first-party data is the solution, and the problem is massive, then why isn't it being used"?

The problem with first-party data

You'd think that first-party data would be easy to do something with as you own it and have access to it. Unfortunately, it's actually really hard to use.

Let's cover just three initial blockers.

Your data wasn't intended for this use.

The first big problem is that your data is a byproduct of your app being used. It was never intended, designed, or structured to be used to the extent of being able to improve your advertising revenue.

Yes, you may have a tonne of data - but being able to process it and use it for an unintended purpose is a monumental task. Multiply this by the number of apps you have, and the difficulty of actually using your data in an effective and efficient way becomes near impossible.

You aren't allowed to combine or mix data

The second big problem is that you can't combine any of your first-party data with any other source to enrich it or make it more useful. Most of what your game generates isn't that insightful or helpful to deliver a relevant or personalized ad.

Additionally, no existing ad network can compliantly use your first-party data and return an ad. If they combine your first-party data with any data not owned by you, it becomes non-compliant with Apple and can't be used.

Furthermore, if the data contains any PII (personally identifiable information) it becomes non-compliant for T1 & T2 users in the EU Economic region who opt out of TCF 2.2 (which is a big consideration as of January 16th 2024 when Google rolls out their big CMP (compliance management platform) requirement changes.

So you've generated all of this data, and you have to figure out how to use it, and you're not allowed to combine it with any other source to make it better.

You would need to build your own DSP, or change the industry.

Let's imagine that you managed to figure out how to use all the first-party data from your app and you've got a solution for only using that first-party data (that doesn't contain any PII) without combining it with any other source.

Now you have to get ads to show to the user.

You either have to get your own advertisers and essentially build your own DSP to have the inventory of ad creatives to return to users based on your first-party data modelling, or work with existing networks and make them accept your unique data for them to understand and serve ads with, and to pay you a higher eCPM in the process.

There's no way building your own ad network makes sense, and existing ad networks can't consume unique data from every publisher and make sense of it to serve relevant ads.

There are a lot of problems with first-party data

It's no wonder why everyone talks about first-party data. It is the only real solution to show better ads to users when they opt out of tracking and to increase eCPMs.

BUT - that doesn't mean it's a practical or commercially viable solution to solve the problem of diminishing eCPMs and in-app advertising revenue.

The solution to first-party data usage to increase eCPMs and ARPDAU.

I believe that there is only one solution, and it's something we've spent the last 9 months figuring out and building at Nefta.

1. You have to have a standardised form of data that can be easily added to any game, so a model can be built from any game and make good decisions about which advert to display and how much to pay.
2. You have to collect data without any PII in any capacity (this is important today, but critical to future-proof),
3. You have to keep all data collected from any single app in its own sovereign tenancy, separate from data from any other app, and not combine or mix any data in any capacity.
4. You would have to run your models on each app and have a way of understanding what these insights mean, to return a relevant advert.
5. You actually have to build all of the infrastructure of a high-scale ad network.

So that's what we've built at Nefta.

It means that when users opt out, you can still use Nefta to show a relevant ad and earn a better eCPM than showing a non-personalized ad.

This is possible because we don't track, link data to a user, collect or process any PII. We don't know anything about the user or the advert. We don't mix any data, ever. We process your first-party data (the standardised tags you add to your app), and we test adverts against the patterns of usage behaviour that we spot.

Our new privacy-first and first-party data approach means we can compliantly show a more relevant ad and pay a higher eCPM for users who opt out of compliance frameworks (such as ATT or TCF).

Summary

It's only getting harder to effectively monetise users, and the world is only getting more restrictive regarding data compliance and legislation.

It is obvious that the "silver bullet" is first-party data, but it is also obvious that it's a monumental task to actually do anything with it.

However, there are a few options available to you:

1. Build your own DSP that can process your first-party and manage your own advertisers.
2. Spend hundreds of millions of dollars to acquire competitors and own their data, then feed this non-standardised data to a new team that can process it, and then try to use this data via LiveOps to increase IAP.
3. Integrate the Nefta Ad SDK and standardised events, and earn higher eCPMs on users who opt out, with very minimal effort on your part.
4. Do nothing and accept your revenue has decreased.

If you're interested in exploring Nefta and our first-party ad network, feel free to contact me via LinkedIn, or alternatively, you can book a meeting with me directly using my link here: Book a chat with Andrew