Discover client recommended agencies

Marketers still scrambling to recover ‘lost learnings’ on Facebook after Apple privacy push

Marketers still scrambling to recover ‘lost learnings’ on Facebook after Apple privacy push

Meta-owned Facebook’s ad business was among those “massively compromised” by Apple’s privacy update in March. As part of The Drum’s Data Deep Dive, we ask marketers how they’re regaining clarity on their campaigns on Facebook and its sister apps.

Amid a wider deprecation of third-party cookies – the web’s go-to tracking tool on most sites – Apple’s contribution was to update the permissions process for its advertising identifier (IDFA), with boss Tim Cook saying the company was “putting the power with the user”. Users on the latest iOS were opted out of in-app tracking and instead served an opt-in message from each app. As Cook put it: ”Many of them are deciding no.”

The ‘Identifier for Advertisers’ is an anonymized identifier that tracks, personalizes and measures behavior on Apple devices. Facebook was among many apps using it to track off-site behavior in the Apple ecosystem, but when Apple turned off that data funnel, most users disappeared from graphs overnight. Recently, CNBC claimed at least 62% of iPhone users remain opted out, although figures vary with some showing almost half of global users allowing tracking.

Meanwhile, Apple’s burgeoning SKAD Network alternative hasn’t impressed. Snap Inc chief executive Evan Spiegel said the tool ”did not scale as we had expected, making it more difficult for our advertising partners to measure and manage their ad campaigns”.

Here’s how Apple’s App Tracking Transparency (ATT) measures impacted marketers this year.

Huge impact

Conor Johnston, business development executive at Impact.com says there has been a ”huge impact on clients”.

Data gathered by Facebook’s pixel wasn’t matching clients’ data, sparking distrust of Facebook’s reporting. As a result, many clients switched to cold targeting based on basic identifiers such as age, location and interests, says Johnston.

Back in September, Graham Mudd, vice-president of product marketing at Facebook, admitted: “The impact on your advertising investment has been greater than you expected. The cost of achieving your business outcome may have increased and it has also gotten harder to measure your campaigns on our platform.”

He claimed that Facebook was under-reporting in some elements, likely by 15%. “We believe that real-world conversions, like sales and app installs, are higher than what is being reported for many advertisers.”

But Facebook couldn’t prove it, so it promised to better its conversion modeling, close reporting gaps and enhance the ability to measure in-app conversions.

Johnston says: “The Facebook pixel was massively compromised and caused advertisers to lose a lot of confidence in it. And the prices of ads are rising each month”.

Can CAPI fill the gap?

Pollyanna Ward, head of paid social at Social Chain, says the impact has been felt across all of social, but that her agency was prepared to handle the loss of third-party data through its publisher and talent network. For now, contextual placement deployed using first-person data is the go-to. But that’s not to say it’s all smooth sailing. On Facebook, campaign reporting is no longer in real-time and data is aggregated at a campaign level, which isn’t overly useful for optimizing individual ads.

Ward adds: ”It is no longer possible to break down conversion results by demographics, which might have been useful for advertisers who want to get a granular understanding of who’s converting on ads.” Events reporting still has its limitations too.

Ward says Social Chain is recommending its clients that are tracking website events install the Conversion API (CAPI) [fairly new] to report on their data, with it matching up with Facebook’s data to create a more holistic view of the events that have happened. ”It helps to fill in the gaps using the brand’s first-party data and is a valuable tool in helping Facebook campaigns optimize towards better targeting and results.”

In October, Facebook’s Mudd boasted of “continued improvements to ad performance and measurement“ and said CAPI reduced the cost-per-action on average by 8% for advertisers that “followed best practices while sharing data via both the Facebook pixel and the Conversions API”.

It’s not all bad news. Ward says this issue has forced agencies to work closer with clients and lessen dependence on Facebook’s measurement to share learnings around first-party data. Even Facebook has been sharing more around some of the more traditional measurement approaches, including marketing mix modeling, which includes using historic data to predict and optimize campaigns. “It’s still very much looking at measuring in the short term and I would like to see Facebook looking to offer long-term solutions too,” says Ward.

Facebook advertising’s effectiveness recovery

Sophie Strong is head of display and social at Wavemaker. She says there has been a race to test new targeting solutions and predictive attribution modeling. But she has had to change the way she brings clients to the market in the meantime, going for fewer, bigger, better ads.

”We need to be less targeted and optimize further up the funnel to allow the platforms to be more effective. This is a big shift to how we used to use the platforms for niche audience targeting. It really is forcing the need to think about how first-party data can be incorporated.”

Again, CAPI is doing the heavy lifting on that front. Strong thinks it can recover a “lot of lost learnings” and, in theory, should be able to match audiences accurately.

But for now, there remain difficulties in proving Facebook’s effectiveness. Strong suggests ”a macro diagnostic, like an econometrics, to effectively guide budget across channels, but then use the in-platform or ad-server/analytics measurement to then guide investment overall”.

In the long-term, Facebook is going nowhere. Strong believes Meta can get better and “piece together the gaps”. She concludes: “Once the new technology is in place, I suspect we will see performance resume as we have seen in the past – as the reality is, it has a large base of users and a strong algorithm that has worked for us in the past. It just needs the right data to feed into the machine so it can work its magic.“

By continuing to use The Drum, I accept the use of cookies as per The Drum's privacy policy