A Guide to Facebook Advertising Optimization with iOS14

Considerations for Optimization

Over the last few weeks, we’ve had multiple asks from colleagues regarding how we are optimizing your ads with the iOS14 data loss. If you’ve found yourself with less data, data showing up in a weird way or not at all, or it just looking at bit off, you are not alone.

This article aims to help us all understand options for optimization and what others are doing.

To begin, there are many parts to iOS14, but we know a few things to be true:

• We lost most attribution data, making the 7-day click attribution window our main source of data from Facebook.

• Ad reporting data will be delayed from those that have opted out (from iOS14 operating systems), giving us a three-day delay on data displaying. This is because there’s a process in which Apple shares attribution data with Facebook for those that have opted out via what Apple calls, “PCM.” The cadence and timing of this is not 100% certain just yet.

• Optimization and the entire process from “I launched, now what” to “I need to scale” is changing rapidly every single day. What we outline below will continue to evolve.

Process of Optimization Currently

As it stands currently, the process of optimization is varies greatly by advertiser. Some are just looking at money spent on ads and revenue, while others are trying to find a magical balance or ratio between certain numbers, giving it an acronym and giving people FOMO for not understanding it.

So, let’s just be transparent about it. Here’s what we’re doing:

First off, we updated all the attribution settings to 7-day click/1-day view so we could see it in reports properly. Our colleague, Brian first shared this on setting up a FB Dashboard, and we did ours this way also from the outset.

After setting it up we still encountered challenges in some cases with setup. So, if you’re having trouble getting your dashboard to reflect attribution window changes, you’re not alone. Here’s a conversation we had this week with a client on just that. (A must read, honestly)

Second, we decided which metrics we want to be optimizing off of:

• 7-Day Click Purchases

• Cost Per Unique ATC

• Unique Checkouts Initiated

• Unique Click Through Rate

• Cost Per Acquisition

The way we are approaching this is similar to what we did in the early days of scaling accounts like Pura Vida & Blenders Eyewear – while we didn’t have ROAS data in those days, we did know the client’s AOV and desired CPA goal. We look at this change in the same way. If purchases are coming in on a 7-day click and those purchases are converting within our desired CPA goal/range (we advise starting at something like 75% of the AOV), we then know we are on target, as there will also be ghost conversions (purchases no longer attributed as they are outside of the 7-day of from a user that has opted out) coming through. Helping guide this of course is our secondary metrics in Cost Per Unique ATC, Unique Checkouts Initiated & Unique Click Through Rate.

Delayed Attribution Dashboards

One way of understanding the data we have currently is studying the relationship between 7-Day Click (7DC) and the rest of the attribution data you once had. Once you understand those differences, it can make things look a bit better. For example, if you knew that for every conversion on 7DC you had another 20% lift on average ROAS within a 28 day click window, you’d want to reflect that across a dashboard to understand that relationship and optimize it better.

So, we made our own dashboard and then showed it to a few friends who improved on it. The best one we’ve worked with thus far is from our close friend Keith Martin at Lunar Solar Group who shared this Delayed Attribution Dashboard for you to download and use. (This link is just a template - please download the template as an Excel document for use).

What Are Other Advertisers Doing?

As a final step to understand what others are doing for optimization, we spoke to a number of friends in the industry and got their takes as well. Hopefully this guides you a bit more in your decision making.

Alex Afterman, of 11:11 Digital, says:

My take right now is to go off 7-day for most of my clients. I've also been tracking MoM ROAS changes in most of my accounts for a long time and, in the higher spending ones WoW for 4 weeks or even in some cases daily for 4 weeks, so I have a decent feel for how different it will be. I also wonder whether list building will become a bigger part of my strategy going forward. I've always kind of pooh-poohed using FB ads to list build for E-comm, but, if actual sales metrics get hazier or even less effective, maybe adding in list building starts to make more sense. You can accurately track the revenue if you are segmenting correctly in your CRM, and you own the data.  

Keith Martin, one of the heads at Lunar Solar Group, says:

We are doing a lot of “relative comparisons analysis.” So let’s say we look at our 5 campaigns and 1 is doing great, 1 is doing poor, and 3 are doing average (relative to one another). Before iOS14 we may have ramped up budget on the high performing campaign by 20% and perhaps we would have shut off the lower performing campaign. Post iOS14, we are going to have the same mindset but since we don’t have the full data picture, we are going to play things a bit more conservatively. In this example, perhaps instead of raising the budget on the high performing campaign by 20% we only raise it by 10% and instead of shutting off the lower performing campaign completely we just lower the budget their by 20%. It is all going to go back to eco-system ROAS though. If the blended ROAS across your entire business is looking good, you aren’t going to want to mess with things too much. Just because something looks like it’s performing poorly relative to the other campaigns, we still may not know whether or not that campaigns being effective for your business or not.

Tim Aton from Zach Stuck’s Homestead Team says:

Our day-to-day buying principles have shifted to flipping between a 2-to-3 day look-back and a 7-day look-back. In the short term 2-3 day look-back, we’re looking at CTR and CPC. If CTR is roughly above 1% but ROAS isn’t great, we let it ride/scale. Likewise, if CPC are under $1 (usually related with good CTR) we’ll also let that ride or scale. For ROAS, we’re looking at a 7-day look-back. This ensures we’re looking past the new 3-day delayed attribution threshold while using all possible data in the 7-day span for making buying decisions. Practically, we’re flipping between these two views constantly while moving through campaigns, ad sets, and ads to try to get a wholistic picture of the trends for an account.

Jess Bachmann from Fireteam says:

We are looking at 7-day click currently and trying not to make big decisions if tactics or strategies are working or not. On the ad set or ad level, we are giving things a wider tolerance. If something looks bad, it needs to look REALLY bad for us to turn it off at this time.

Side note, if you’ve never seen this thread from Fireteam on diagnosing issues, please read it.

It’s our hope this gives you some ideas on where to begin optimizing your ads from. If you have questions, always let us know via the contact form on our site or on Twitter.