There’s been a lot of talk about how iOS14 and Apple’s App Tracking Transparency updates will change Paid Social. So far, the conversation has been led by Facebook, particularly how Apple’s changes will affect retargeting audiences and performance measurement.
This is a huge topic which marketers all over the world are still grappling with – while getting on with managing and reporting on their day-to-day ad campaigns.
But there’s one impact that not many people are talking about that could be adversely affecting your reporting, campaign management and results. Particularly if you spend your days knee-deep in Ads Manager.
That’s delayed reporting.
What is delayed reporting?
You probably spotted this as a bullet point in Facebook’s How the Apple iOS 14 release may affect your ads and reporting:
Delayed reporting: Real-time reporting will not be supported and data may be delayed up to three days. For iOS 14 app install campaigns, conversion events will be reported based on the time that they are reported to Facebook by the SKAdNetwork API and not the time they occur. Web conversion events will be reported based on the time the conversions occur and not the time of the associated ad impressions.
Delayed reporting has always been a consideration for Facebook, especially if you used Facebook Pixel to attribute conversions within a certain period after a user clicked or viewed an ad.
Although the default Facebook Pixel attribution method was criticised for being overly generous, it gave a fuller picture of attribution than just ‘last click’, especially if you have customers who take longer to make a buying decision than usual or high value products. This helped campaigns to optimise towards purchases by giving more conversion data.
Before – if users take their time converting, but still convert, generally you’d see associated transactions and revenue in Ads Manager. This would be credited on the date of impression so you could tell how your campaigns and creatives were performing day-to-day.
After – some of this is still in place – but there’s been a few major changes:
- Facebook’s removed the old default 28-day click view, 1 day view window.
- Now the new default is 7-day click – although 7-day click and 1-day view can be enabled
- Facebook now attributes conversions on the day of the sale, not the ad view/click (this is the opposite to pre-iOS14)
- Real-time reporting will not be supported and web conversion event data may be delayed up to three days.
But what about what the conversion modelling Facebook’s implemented around missing data for opted out users?
Estimated results: For iOS 14 app install campaigns, the SKAdNetwork API will report results to Facebook, aggregated at the campaign level. Statistical modelling may be used to account for results at the ad set and ad levels, unless the campaign is composed of a single ad set and ad. For web conversion events, statistical modelling may be used to account for conversions from iOS 14 users.
According to this, Facebook’s still going to estimate what conversions you might be missing out on, right? Yes… but this may be up to 3 days later and the delay may include opted-in users as well as those who’ve opted out of tracking via iOS14. This is to protect the privacy of individual users.
What impact does delayed reporting have?
This could make your recent results look a lot worse than they actually are. Why? This delay just affects the app install or web conversion data (e.g. purchases, revenue, ROAS) not the on-platform data (e.g. media spend, CPC, CPM)
Facebook knows how much you’ve spent, who you’ve reached, and if they’ve clicked, so it’ll show you the full figures for on-platform data but the impact on your revenue will be noticeable after several days when all the results are in.
This means that when reporting week-on-week, you are likely missing a few days of website conversion data, but not spend, so ROAS looks poor.
However, it also means that data from 7-14 days ago is going to be higher than when you captured it last week, making the 0-7 day data appear even worse when comparing the two within the Facebook Ads interface – as you’ll be comparing a full week of revenue data to 4-5 days of revenue data instead.
Example:
Across several ecommerce clients, we saw on average a 20% improvement in ROAS for the same week when comparing data gathered with a week’s time lag, to data gathered immediately after the reporting period ended.
What should I do about it?
The situation is still evolving, but for now we suggest:
- Implementing a min. 3-day lag to your weekly or monthly reporting schedules to allow delayed reporting information to be processed
- Avoid making knee-jerk decisions about very recent performance and changing campaigns frequently – as when revenue is processed the picture may be different and lead you to make alternative optimisations
- Leverage Google Analytics data – for example, if you still want to assess performance trends in real-time, or sense-check unexplained spikes and dips. For this, ensure you are tagging your campaigns correctly so you can compare campaign, ad set and ad-level data within Google Analytics
If you want an introduction to iOS14 please read our blog post How iOS14 Affects Marketers.
Learn more about the wider conversation within the ad tech and measurement ecosystem, or how to best manage your Paid Social strategy in this new world, check out our webinar on the topic or get in touch.