Insights, strategy and trends.

I assume your red flags were raised by this topic, and rightfully so - but don't you worry - it's not a clickbait. I've been running marketing campaigns since "Drop it like it's hot" was destroying the Billboard 100 and I've got a bit of knowledge to share with you.

See, I wasn't the smartest kid during math class as a youngster, actually, pretty far from it. Even though, when starting out with marketing campaigns and optimizing for conversions a couple of years later, I started to understand something that has been - and still is - crucial for marketing success.

But what's really remarkable is that only 1 out of 20 people I talk to about marketing seem to understand this. And you know what? The marketers who've understood this "12 second math lesson" I'm about to share are far more successful than the ones who don't. Although, to be honest, the 51% I mentioned in the header of this article is just a number that sounds pretty good. It could be 25% lowered CPA. It could be 92% lower.

If you're that one in 20 people who actually know this, please don't be frustrated - this article is for those 19 others.

⏱ OK, fasten your seatbelts (wow...) or whatever you need to fasten, here it goes:
Always make decisions based on a statistical significant number of conversions. It might seem trivial but if this was an isolated event I wouldn't be writing this post. This is the part where you either leave because it's as obvious as putting on pants in the morning or this might be a huge, but small, but also huge wake-up call.

So, my question would now be; did you fully understand that sentence? Or let me re-phrase; if you didn't get the part with the fancy words "statistical significance": are you sometimes analyzing your marketing campaigns, ad sets or ads and find yourself pausing them if CPA is too high after like...  6 conversions? Well, think again - because you can't make any decisions based on such few numbers of conversions or data.

Why, you ask?

  1. The Facebook algorithms have probably not been able to go past the learning phase, which means Facebooks algorithms haven't even reached enough conversions to start highly efficient optimizations.
  2. Mathematically, it's impossible to make decisions based on such few conversions. Simple as that! It's difficult so say where the "breaking point" is, since it's more case by case. But I'd definitely say at least 100 conversions are to prefer before making hasty decisions. Of course, you can also automate this instead of manually keeping track of all your campaigns, all ad sets and every ad 24 hours a day, 7 days a week. (At Zalster, we've developed algorithms with AI to automate and optimize marketing campaigns, making them much more profitable and high-converting.)

Any questions about this? Shoot me an email to albin@zalster.com and I'll answer as soon as I can :)

Do you like to read heavy and useful articles and want to go deeper on the topic - making you look like the meme below? Read this article about statistical significance in the Harvard Business Review.