Executive Wrapup: Apple privacy changes, Facebook Ads & Google 3rd party pixels
Not many of us enjoy digging into knotty areas like how websites and apps track users, or how advertisers use that data, but there is a rising tide of change driven by consumers and privacy laws so we all need to pay attention! I was trying to get my head around it to reflect on the implications for retail and direct brands. Spoiler – I can’t give you hard numbers because there are so many variables but I hope this gives you food for thought and helps you plan for the changes.
If you want to know my conclusions without having to wade through the background, jump to the end of this article. Otherwise, strap in for the whistle-stop tour! I’ve tried to make it understandable for those who may not be hands-on with these channels but, I’ve linked my source articles that go into more depth. It’s important to state at this point that things are evolving day to day.
This has been coming for a while
Back in January 2020, Google announced its plans to remove support for third part cookies by 2022. In March 2021 they confirmed they would not build alternative identifiers to track individuals. On Monday 26th April, Apple are releasing iOS 14.5 which forces all apps to request explicit consent to access the unique identifier (IDFA) for each user. Without this consent, they cannot connect data like conversions, add to carts, website visitors/tracking (which affects interest targeting) to individual users. This is likely to impact reporting, optimization and audiences.
These changes are driven by increasing consumer expectations and the rapidly developing legal environment for privacy and data protection around the globe. For now, the IOS 14.5 impact for any advertiser will depend on whether their customers and prospects opt out but it is likely to slightly reduce performance and will certainly impact measurement. Currently, about 70% of IOS users share their IDFA with app publishers, after this change it’s estimated that this number will drop to anywhere from 10% to 39%. The Google pixel change will affect the whole market but we have a little more time to plan for that. In either case, as a retailer or direct brand, you are in a stronger position than many advertisers for two reasons:
- You gather so much data through 1st party pixels on your website and through processing customer transactions. You should still be able to use this for modelling lookalike audiences.
- The majority of retail digital spend is through Google or Facebook, both of which will find ways to retain their ability to target and optimise via 1st party data that they control within their platforms.
By contrast, those using programmatic advertising will face the biggest impact. For each of these three platforms, let’s consider how the changes affect both performance and measurement…
What is the impact from the IOS change? (likely to be felt by end June 2021)
On Facebook Advertising
From a performance perspective it is hard to predict the impact, which will vary by brand according to your user base. Depending on the proportion who opt out, the impact may be quite modest because it will only affect those who opt out. Performance will be affected because the algorithm has less data to work with but Facebook have so many other data points within the platform, they may be able to bridge most of this gap. However, they will not be able to accurately recommend products based on the individual and their stage in the purchase path. They will find a way to serve relevant ads, but perhaps not quite so relevant as they were.
Our latest understanding is that users who said ‘No’ to the IDFA opt-in will be excluded from custom audiences. However, they will be included in ‘lookalike’ audiences. This will mean you cannot target some of your customers directly in your retention campaigns, but would probably reach them through lookalike audiences. The key difference here is that you would probably bid a lot less for lookalikes and therefore risk not winning the auction for those customers.
So, for performance, it is a case of wait and see, although there are specific measures you can take to mitigate the impact. If you are interested in these, please send us an email to the address at the end of this article.
Measurement will also be affected because all users who don’t consent to apps using their IDFA cannot be measured. Facebook will model the missing data but only when using 1 day post click as the attribution window. Any longer attribution window or post-view will exclude these users. In addition, they have reduced the available attribution windows from 28 days to just 7 days post click or post view. For most retailers this is less of a concern because they will often use 7 days as a conversion window anyway.
On Google Advertising
Google’s ad platform is much less affected because it is mostly web-based or through their own apps like Youtube where people are explicitly logged in so it is 1st party data. The majority of retail spending goes through search and shopping which will not be affected. If you use Google’s Display network to a large extent, you can read their guidance on it here.
On Programmatic Display Advertising
From an Ad Targeting perspective, many targeting types including device-level retargeting will no longer work for users that have opted out of sharing their IDFA. Platforms like Google and Facebook have a myriad of other deterministic variables they can use to identify devices (email, phone number) but other programmatic platforms that don’t have such deterministic ID graph information are likely to see a reduction in targetable audiences.
When it comes to measurement, most measurement platforms were built around the IDFA identifier. Apple have helpfully provided their SKAdNetwork API which allows conversion data to be captured at campaign level but your ability to measure will be compromised.
Until the dust settles and the industry better understands the extent of the impact and the efficacy of the alternatives, it seems likely that some spend will migrate from these less accountable campaigns to the ‘walled garden’ options like Google, Facebook, Amazon, who have richer 1st party data for targeting and will be less compromised on measurement.
What is the impact of the Google 3rd party pixel change?
On Facebook Advertising
Your Facebook pixel can be configured either as a 3rd party cookie or a 1st party cookie. They explain in a little more detail here. As a consequence, campaigns should be able to continue operating as they do now. However, browser privacy measures and user-controlled ad-blocking are already compromising pixel performance, even before Google stops the use of 3rd party pixels.
For this reason, it is best practice to implement Facebook’s Conversion API which can run in parallel to the pixel. You are effectively connecting directly from your server to Facebook’s to record every transaction. With the right setup, Facebook will dedupe the two sources and ensure that events missing from the pixel are properly counted towards your campaigns. This is important so that Facebook can optimise based on a full picture of conversions and will give you a competitive advantage in the auction. (by the way, if you have retail stores, you should already be using this to upload store transactions)
On Google Advertising
On Google owned media such as Google.com and Youtube, you will continue to be able to use your first party data for targeting or use Google’s audiences as you do now. Measurement will also be unaffected because it is using 1st party data. You can read more about how Google is moving their Global Site Tag to 1st party here.
For the Google Network publisher partners (GDN – ads that they serve via third party websites), advertisers will need to rely on contextual targeting or privacy sandbox API’s such as FLoC or FLEDGE. These group users into anonymised cohorts based on their profiles and behaviour. Much of the industry thinks this is a bad idea and not so far from the current cookie-dependent world we currently operate in. Google explain their rationale here.
The EU has strong views on the legality of the FLoC approach which is forcing Google to delay tests in Europe. So much is likely to change before Google actually roll out the 3rd party cookie change that it is hard to make specific plans. It seems likely that GDN may suffer some of the same limitations as the broader programmatic world – more of this below.
On Programmatic Display Advertising
The loss of the ability to identify people through third party pixels or identifiers will have the following impacts as identified by the IAB:
- Frequency capping will no longer be available in its current form so your campaigns will be less able to limit spend at a user level
- Marketers will no longer be able to syndicate their first party data to a publisher, which transforms it into third party data. As a consequence, retargeting or other forms of cross publisher audience targeting will no longer be possible.
- First party publisher data will also not be able to be used by publishers for audience extension without cross publisher identifiers.
- Audience based dynamic creative optimisation will be greatly hindered without cross publisher identifiers.
- View through or multi touch attribution will no longer be possible.
The combined impact is significant but the whole industry is still working on ways to mitigate that impact and enable publishers to continue to monetise their ad inventory. IAB Europe (the Interactive Advertising Bureau – industry body) have written a detailed guide called ‘The Post Third Party Cookie Era’ that expands on the issues above and the solutions proposed so far.
So what is the prognosis for retail/direct brands, and what should I do next?
As mentioned near the start of this article, many retail and direct brands already have a head-start in the new privacy-centred era because they have plenty of opportunity to capture and use 1st party data. Customers also have compelling reasons for sharing their data – to enable or enhance service delivery. For example, Amazon and Asos give me better product recommendations because I’m logged in. Bravissimo offer a more personalised fitting service in-store because their customers allow their measurements to be stored. The brands who win will be those who show customers the connection between sharing their data and having a better customer experience.
ACTION 1: Improve 1st party data capture at every touchpoint. Show clear benefit to customers in them sharing the relevant information or logging in.
Clearly Facebook (amongst others) will have less data to work with moving forwards. As a result, you need to implement the Facebook Conversion API so that you can attribute your campaign performance to its full extent because Facebook can ‘see’ all of your orders. Google are also addressing their 3rd party ‘blind spots’ by ensuring their pixels are 1st party. Now is the time to check that your brand is using the latest 1st party implementation.
ACTION 2: Implement Facebook Conversion API and Google 1st party pixel approach.
ACTION 3: If you haven’t already done so, you must urgently verify your domain in Facebook.
It seems likely that a proportion of programmatic ad spend will migrate towards 1st party ‘walled gardens’ like Google and Facebook because they will be more accountable and optimisable when 3rd party cookies have gone. This increased competition for what is finite ad inventory will mean that brands will have to work smarter and be better optimised to maintain their position.
ACTION 4: Carry out a review or audit of your current performance activity – are you adopting the latest best practice across the board?
Thanks for reading this far. more2 exists to help retail and direct brands address these areas and achieve sustainable long term growth by understanding and harnessing their customer data. Please email email@example.com if you would like more information or have challenges to share.