What the FLoC? Io-SO confused?

Our web analytics is changing into more of a black box. So what’s changing? What are the practical implications of these changes? What can we do immediately and ongoing to mitigate the fallout? 

A little context before we dive in... (or skip to "ios Impact on Facebook" to get straight to the juicy stuff)

Yes, the game is changing *again* in digital and for many of us, this ‘ain’t our first rodeo. 

Remember back in circa 2014 when Google announced that sharing the specific keyword data for organic searches was a breach of privacy? Overnight digital marketers watched their Google Analytics accounts become riddled with (not set) keywords and cries ensued. 

Those of us who were implementing white hat SEO best practices, survived. We knew that the *real* game was in being contextually relevant and adding value in the form of high quality content. Equally, it flushed out the black hatters who were using this specific data to find the few opportunistic keywords to keyword stuff new articles they wrote in an attempt to rank and drive traffic - fast. 

The paid search strategies became more “trial and error” with a total lack of visibility over which organic keywords were driving us conversions. The result? More money for Google. As we started spending more on trial and error, we painted a new narrative to our clients that “PPC takes time” (which is something we once said about white hat SEO when the black hat SEO tactics stopped working). 

Wherever there is attention (read: eyeballs) marketers will find that attention and flock there, often ruining whatever platform was once used purely for entertainment or utility.

It’s true, I’m sorry. Marketers ruin everything.

Our tech giants of course, know this and toe the oh-so-careful line of generating the highest possible commercial value they can, while teetering on the edge of providing a useful or entertaining experience to keep consumer engagement (and therefore, *clicks* on their in-demand ad space) as high as possible. 

It would do us well to remember that the ultimate objective of our tech giants is a commercial one. As such, when an update is made, despite whether it on first glance appears noble in its nature to better secure consumer privacy, make no mistake, it has been pressure tested to model commercial impact. 

After all, our platforms have a lot of shareholders to answer to. 

So with that context, let’s unpack what’s coming (and what’s already started)! 

Ios14 impact on Facebook: 

What is it? 

Apple’s new ios14 rollout will prompt device users when apps request to track their activity and offer them the chance to “opt in or out”. 

This is part of Apple’s new initiative; the “AppTrackingTransparency framework” 

This rollout will prevent Facebook from tracking activity from app users. This will prevent advertisers from tracking and analysing mobile behaviour and offering personalised ad targeting. 

If you use an iPhone, you can head to Security > Privacy > Tracking and if your ios has updated to 14 (and you haven’t snoozed your “please update” notifications for a zillion years like I typically do) you’ll have the ability to toggle this according to your preferences.

Tell me in the comments whether you opted in or out ??

Why should you care?

According to Statista, In 2020, iOS held 52.84% of Australian phone market share. 

Without a crystal ball, it’s impossible to know what percentage of users will choose to “opt out” of tracking but for those who do, their activity will be entirely lost in a black box of data. 

It’s worth noting that while Apple has made this important update “on behalf of its users” in the name of “protecting user privacy”, Apple themselves have not limited their own data collection from device usage. This has suitably pissed off Facebook who have taken them to court on the issue. 

Okay, but why specifically? What are the practical implications?

Remember, not everyone you are targeting will opt out however for those who do, you can expect to see the following: 

• You will still be able to track impressions, clicks and conversions for 8 key events you define with Facebook leveraging their app pixel conversion tracking feature
• Conversion reporting on these 8 events will be delayed up to 3 days 
• For this reason, Facebook is recommending ad sets run for at least 3 days 
• Due to the limitation of tracking, remarketing and personalisation of ad audiences when executed via “web analytics” (i.e. use of the pixel) will not be possible for IOS opted out users.

Ios14 opted out users can still be reached via: 

• Contextual targeting 

i.e. if they “like” a page on Facebook/Instagram or the use of Facebook’s other contextual insight groups

• First party data targeting 

Uploading first party data (email addresses, phone numbers) for use in ad targeting 

There’s more, but I see the above to be the most notable. You can read Facebook’s full announcement here. 

What now? What should I expect to see and how should I change my strategy? 

You will see your Facebook “ROAS” and conversions decline

• This will be for two reasons:

(1) Facebook has been forced to change their default attribution window

(which frankly, I’m happy about since it allows marketers to view data as a clear last-click model rather than Facebook’s often greedy multi-touch attribution model which can at times, take credit for more than it deserves)

According to Facebook:

The following windows will be supported under the new attribution setting:

1-day click

7-day click (default after Apple prompt enforcement)

1-day click and 1-day view

7-day click and 1-day view (initial default)

Overall, we’ll lose our lookback windows which will make it difficult to understand how Facebook has “played a role” in view through and post-click conversions when users ultimately convert via another source (such as converting via email, direct or Google but where Facebook played a role along the way through). In any case, my recommendations listed down below remain. 

(2) Your web analytics based audience sets will be diluted.

According to Facebook:

As more people opt out of tracking on iOS 14 devices, the size of your app connections, app activity Custom Audiences, and website Custom Audiences may decrease.

You will only be able to leverage data sampling for optimisation decisions.

When we don’t have a complete data set, our best effort is to leverage sample data and mirror insights across. From experience, we know that data sampling can be risky business (insights don’t always apply across) but it’s worth a try. Ultimately, despite a 3 day delay, you will be able to track conversions that occur from ios14 opted out users and therefore, you will be able to notice within a few days whether the insights can be applied across.

Keep in mind, this doesn’t apply to all ios users, just those who have opted out of tracking.

For tracking, setting KPI’s and attribution analysis, I recommend:

1. Review historical data sets and compare impact as audience dilution occurs. Separate out campaign structures for other devices (Android) and model mirror top performing strategies across to ios campaigns (this may not always be possible depending on the target set).
2. Set up GA4 (I’ll talk more about this in the FLoC update below)
3. Ensuring you have UTM tagging in place for Facebook ads

Overall, reconsider your strategy with Facebook’s guidelines in mind and restructure campaigns with tight ad messages and placements, built to achieve very specific KPI’s aligned to your 8 allowed events. 

Gone are the days where we used to A - Z split test. Keep campaign structures tight and think long term. 

Oh and of course, if you weren’t already listening to all the digital marketers who have been banging on about getting first party data strategies sorted, it’s a good time to start listening and start enriching.

This is my specialty (and I’ve found a few fun ways of enriching first party data sets from “nothing”) so be sure to reach out if you’d like to explore this. 

Alright now it's Google's turn. What the FLoC?

Google’s move to FLoC Privacy: 

What is it?  

Google has been open about joining many other platforms to ensure a Cookiepocolypse in our digital world. 

FLoC stands for Federated Learning of Cohorts. It’s Google’s proposed solution to ending cookie tracking and improving customer privacy, while serving advertisers needs to segment groups of people based on their demographics, behaviours and interests - without risking the unlawful or immoral sharing of PII (personally identifiable information). 

In a practical sense; FLoC is Google’s answer for safer advertising targeting and analytics tracking. 

Recent reports highlight that it apparently doesn’t adhere to GDPR standards (even just for testing!) and it’s copping quite a bit of criticism more broadly… so jury is out whether this will in fact be, the final solution. 

Why should you care?

Google have themselves even outright announced that the FLoC solution is “nearly as effective as cookie-based approaches” going on to say that “at least 95% of the conversions per dollar spent when compared to cookie-based advertising”.

So basically, Google are pulling their "(not set)" of the twenty-twenties decade on us and is basically saying “Yep, we know it won’t be as effective as what we were doing previously but we’re doing it anyway.” 

Of course, as I noted in my intro above, this effectively means that advertisers will cop the brunt of this, paying for less effective advertising, right into Google’s already deep pockets.  

Okay, but why specifically? What are the practical implications? 

Unlike what I could share with you above re: the ios impact, we don’t know much on the FLoC front yet.

We won’t *really* know how to address nuanced strategies until we start seeing the tools roll out but we can start to prepare for this new digital world in the following ways: 

1. Capture and enrich your first party data

Whether it’s server side analytics set up or actual first party customer data capture, the more data you own about your customers and prospects, the more targeted opportunities you'll have for segmentation.

If you're not yet in a position to get yourself a CDP, you don't necessarily have to. I've helped a number of businesses set up patchwork systems that allow them to start capturing the data they need for hyperpersonalisation in advertising and email - and without huge budgets or in-house data teams. It is possible and if there were ever a time to start considering this as your insurance policy, it's right now!

2. Start thinking about how you would alter your strategy targeting a contextual audience

Would you need to change your messaging on your landing page?

Airbnb shared their 2020 win for maintaining traffic despite slashing performance budgets by $800M and while most of us are not working with nearly as much brand equity to leverage, it would do us well to remember that the funnel isn’t linear.

I ask you this: do you know whether you could bolster revenue and convert more overall users if you were targeting new markets and converting “cool” eyeballs through a combination of broad contextual advertising with targeted landing page messaging? There’s an argument to say that those hyper-targeted ads were just spending money on people who were going to buy from you anyway.

In any case, start testing now and you’ll have a leg up on your competition when the rug is pulled by our good friends Google.

3. Start using GA4

GA4 is really just the fancy new name for Google’s “App + Web” property but they have added a bunch of new fancy features in this update. What it will do is give you aggregated new measures to look at along with more behaviour based events (such as scroll depth etc.) which will help train your analytics brain to consider how we should optimise and run attribution models with aggregated measurement. 

If you haven’t already installed this you can follow my tutorial here. 

So what now?

We're still waiting for the full impact of this rollout and as I learn more, I'll share more but what I do know, is that I'm set up over here with the popcorn watching, waiting and ready to pivot.

Ultimately, the tactical algorithm game may have changed but wider marketing mix modelling and attribution strategies will remain for those who are measuring marketing effectively.

So if you'd like to know whether your marketing mix is effective and how to adopt an attribution model that will truly help you measure the return on your marketing efforts OR if you need some immediate guidance and support to improve your first party data strategies or whip your analytics into shape - reach out for a chat!