Improving traffic quality in a more private world

The latest change to affect marketers is the implementation of Apple’s App Tracking Transparency update (iOS 14). This update has forced apps to give users an option to either opt in or opt out of tracking. The underlying impact of this is that social media platforms will not have permission to track users who opt out – whether through Apple’s Identifier for Advertisers (IDFA) or any other identifier. Early data for this has suggested that the vast majority of users are choosing to opt out, with reports stating that 96% of Apple users have opted out – causing a headache for paid social advertising.

What’s actually changed?

Facebook, in particular, has been a vocal critic of ATT/iOS 14, arguing that it will harm the user experience by showing less tailored ads. Despite this, it has been forced to adapt how it operates along with every other social media platform.

For users who opt in to tracking there will be no change to how things currently work. For users who choose to opt out, data will become more limited, with conversion events being restricted, aggregated and delayed.

To account for this Facebook has enforced reporting limitations, from reduced user breakdowns, which no longer split data by age, gender or device for conversion-led campaigns, to delaying and restricting reports.

Further to this, attribution windows have also been heavily impacted by the changes. Since iOS 14 implementation, the default attribution window within Facebook has been changed to a seven-day click window and there has also been the removal of the 28-day click window. There are also heavy restrictions that have been put on view-through conversions, with one-day views being the only option available for conversion-led campaigns.

Outside of reporting implications there has also been limitations placed on targeting with custom audiences as a result of audience sizes being reduced. Similarly, there is also now less visibility of users’ app behavior across third-party sites, leading to fewer signals being reported back to Facebook and consequently affecting its ability to target users as effectively.

In an attempt to combat these changes, Facebook has introduced aggregated events by using statistical modeling for users who have opted out. However, while this is useful for larger businesses that can send enough signals from users who have opted in, this will potentially not be as effective for smaller businesses.

Although there is clearly a lot of change, it’s important to recap that much of what is available within Facebook and other social media platforms will continue to be:

• First-party data – platforms own their own first-party data

• On-platform behavior – still able to learn about the interests and behaviors of users who have opted out

• Ad eligibility – platforms are still able to serve ads to users, even if they’ve opted out of targeted or personalized ads

Impacts (so far)

Since the ATT rollout, the impact has varied across platforms, audiences and advertisers, with the complete picture being clouded by the lifting of lockdown restrictions coinciding with the roll-out in many locations. However, there are certain areas in which we can identify definite changes.

One of the most evident changes is the UI changes brought about by the change in attribution settings. As a result, account managers will now often not be able to tally data across multiple campaigns as attribution windows will no longer be in sync across the account. In order to combat this, campaigns will now need to be filtered or changed to the same attribution window to get a wider view of account performance.

Another problem that has been encountered is the lack of breakdown now available within the platform. For instance, if a campaign is reporting 50 conversions, this would previously have been split by age, gender and device, however this is no longer the case. While this was expected, it has nonetheless hindered what can be reported on, and subsequently reduced the learnings that can be made from campaign performance.

Reporting lags have been one of the less obvious impacts of the roll-out, with Facebook enforcing a three-day lag to certain metrics including revenue and website purchases. This in turn has impacted real-time/recent reporting. In particular, metrics such as ROAS will be heavily affected as Facebook will still account for media spend, while revenue will not necessarily feed through until the three-day period is over. As a result, it is advisable to caveat any reporting within this three-day lag period with this information in mind.

Hitting previously-set targets has become more difficult. There has been a move to increase media spend for businesses that can afford it in order to maintain consistent target sales volumes, despite generally higher CPMs and CPAs. The knock-on effect of this is that CPMs are being pushed up, often impacting smaller advertisers who are being priced out of their usual auctions.

What should you do next?

With all these changes and impacts to keep in mind, there is a raft of actions that should be considered to help optimize your paid social media account for a post-iOS 14 world.

The most pressing area of concern within any Facebook ad account is to ensure that all the tasks highlighted in the resource center tab have been completed. Within this section there will be prompts to:

• Adjust to new attribution windows – compare windows and download historic data

• Claim your domain

• Configure eight preferred web conversion events per domain

• Set up aggregated events measurement

• Implement server-side tagging – conversions API

Outside of this, we have also found certain tactics that have worked well, one of which is to consolidate account structures to give maximum data density. By implementing this change it is possible to maximize the signals that Facebook still receives. For example, consolidating remarketing audiences into one bucket instead of having granularized audiences may now be more effective, particularly if the audience sizes are small.

Alongside remarketing audiences being heavily impacted, lookalike audiences have also been affected with the audience source being reduced. With this in mind, increasing lookalike source size will allow Facebook to utilize more signals when it creates the lookalike audience.

Another tactic to maximize signals being reported back to Facebook is to ensure the attribution windows are not set to one day. Where possible, it is beneficial to increase this to the maximum option of seven-days click and one-day view. This is particularly important for conversion-based campaigns, which need a minimum of 30 actions to exit the learning phase.

With reduced third-party visibility, leveraging on-social signals is another method that should now be leaned on more frequently. In particular, creating engagement campaigns based on a user’s interaction with top-of-funnel content can help build remarketing lists without the need to track users outside of the platform.

Alternatively, another on-social signal to utilize is the creation of lead generation campaigns within the platform. By encouraging users to leave their details within the Facebook interface, there is no need to depend on conversion signals being sent from a third-party website.

One variable that hasn’t been affected by iOS 14 is the impact of creatives within paid social advertising. Having strong creative is arguably more important than tweaking targeting to cater for the newly-imposed limitations, and getting this right will have a huge say in determining whether a campaign will be successful. With this in mind, it is crucial to step up testing of creatives among different demographics to find which are performing effectively. Taking a more hands-on approach here is also recommended as letting Facebook take the reins will often mean some creative variants are not rotated enough to gather useful learnings.

With reporting becoming less user-friendly within the Facebook UI, it can be advisable to create dashboards using third-party software to help aggregate data that is most important to you. Having this data in an easy-to-read format can help you spot inefficiencies within the account that are now much harder to notice within the Facebook UI. Outside of this, it’s also important to recognize that there is likely to be a gulf in performance pre- and post-iOS 14 as we naturally see fewer conversions reported. Ensuring that this is understood and communicated clearly to the relevant stakeholders will help allay fears of diminishing performance/returns.

It’s important to avoid making knee-jerk decisions and take a more holistic marketing outlook. Paid social should not be seen as a silver bullet for businesses’ promotional activities and should be considered as one lever in a wider marketing mix. With this in mind, it’s important to not make changes based on a campaign’s initial performance, particularly as some metrics are now delayed by three days. On a longer time frame, it’s important to consider that a user journey may vary depending on the business sector. It should be emphasized that the initial groundwork carried out by paid social activity may not bear fruit for some time. It will also likely not be accounted for from an attribution perspective, however this does not mean the campaign did not have value.

Summary

The direction of travel for the paid social landscape is obvious, with greater privacy and transparency being a cornerstone of any activity moving forward. While this is causing a degree of panic for digital marketers, it’s important to recognize that through testing different methodologies, new best practices will be found, and that although there will be short-term pain, ultimately advertising works and paid social will continue to be hugely successful.

James Taylor, paid social specialist at Impression.