The iOS14.5 Curveball

Did the iOS14.5 curveball crush your ads?

If it didn’t, congratulations!

If it did, hang in there.

In this post, I’m going to share a few insights we are gathering based on what we see in our accounts.

From 100 to 0

If you’re (unlucky) like we were, your ads were probably crushed.

Within days, we went from KILLING IT to KILLING OURSELVES (figuratively).

In our minds we already killed Apple and Facebook a hundred times.

We were well-prepared (or we thought we were).

  • We verified all the domains ahead of time.
  • We prioritized the relevant events on AEM.
  • We championed Hyros and turned over 50% of our clients to use Hyros.
  • We pushed real-time calls booked among our high-ticket coaching clients into Slack.

But nada.

Either Facebook or Apple decided to spring a nasty surprise on us.

We stopped getting calls even on our most reliable account (which has been producing $60-100K per month from $15K per month in spend for the past 18 months without fail).

We expected the worse and hoped for the best…but only the worst came.

So if you experienced the same thing, I know how you feel.

Turning things around

Let me be clear: there isn’t one thing we did that worked for every account.

Just like there wasn’t a magic pill 9 years ago when I started advertising on Facebook, there still isn’t one today.

So instead of giving you a generic list of recommendations, I’ll give you a breakdown of the key things we did for a few of the accounts.

Hopefully these transparent examples give you an idea of what to do with yours.

Accounts 1 & 2: High-ticket coaching programs

  • We brought back some of our old winners in 2020.
  • We rewrote some of the ads (similar angle, different copy).
  • We stopped using dynamic creatives. Instead, we tested each element separately (hi 2014!) For example, we had 4 headlines for a single ad, we tested them individually and found one which produced 4 calls at the same spend the other ads took without producing a single call. Interestingly, Facebook spent the least amount of money on that ad initially.
  • We asked our client to double down on organic traffic (email + FB group 2-step promos). This strategy was highly effective and it helped both accounts boost immediate calls and sales.
  • We found unusual winners among lookalike audiences using Facebook saved post and page interactions.
  • We also found unusual winners among higher % lookalike audiences (not common in these accounts).

Accounts 3 & 4: Online courses

  • We turned on some old ads which had delayed attribution (i.e. according to Hyros, they were ROI-positive but we had turned them off during our 3-day and 7-day optimization windows). We saw an immediate spike in sales for several days.
  • We found a winner in combining two lookalike 1% audiences. Our standard approach is to test all lookalike audiences individually, so this is the first time we’re testing it and we’re now expanding this test to other audiences to see if it works, or if we’re just a lucky pony. (Try it at your own risk.)
  • We lowered the ad spend on several campaigns/offers, which according to Depesh, is one of the most underrated Facebook ad optimization tactic.
  • We amped up our email strategy. Apart from doing ads, we also run the entire email marketing strategy for one of these accounts. We managed to produce an additional $10K sales in this account through our emails, which cushioned the negative impact of the ads. We ended the month at 2.8x ROAS ($80K+ sales).

What to do if you’re stuck

First, stay connected with other advertisers.

One of the most useful things I’ve learned is that one tactic can sometimes change your life (literally).

For example, the idea of testing Facebook custom audiences who saved your post or page came from a tweet. He also regularly shares tactics like this.

(Shameless plug: if you’re on Twitter, follow me too. I can’t guarantee your ads will do well but I can make you laugh and laughter is the best medicine when you feel depressed.)

I also regularly get ideas from chatting with my agency friends.

Second, make a list of what you plan to test, prioritize the list based on your conviction, and review the list after each test is complete.

For example, I used to think that a lookalike audience of paying customers is the best list you can find.

Later, I found that the lookalike audience produced the cheapest calls, but none of them converted. By that time, we invested $4,000 in it. We were able to review our list of audiences and re-prioritize them. Eventually we found winners among those who visited the testimonials and podcast sections of the website, rather than the paying customers.

This is just one example.

Third, don’t give up.

Keep going.

Not everyday is going to be rainbows, but storms don’t last forever.

Photo by Brett Jordan on Unsplash

Subscribe below to get future posts from Jason How