-
Winning Adset and how to scale?
Posted by on July 2, 2020 at 4:39 pmSo I’ve split tested different adsets and ads on Facebook for a week. Kept trimming off the fat and at the moment I’m down to one adset which seems to be a real winner.
My question is now that I want to scale my ad spend is it simply a case of increasing my budget at the campaign level and leave it at one adset to the single winning audience? I heard here and there that some people increase budget and duplicate adset to the same audience but cannot understand the logic behind it.
Kinda new to Facebook ads so pardon my ignorance on this.
-
2 Replies
-
1. If something is working don’t touch it.
2. Duplicate the campaign with just that ad set at a higher budget.
3. Increase the new campaign budget at 10% day over day as needed. Hope that helps. -
There are different schools of thought here, but the central truth is that you shouldn’t touch something that’s working. You can increase budget every day to a limit. If the increase to spend does not go over a 20% increase to daily budget, the algorithm will not reset. You can scale pretty quickly just notching it up day over day like this. The worst thing you can do is fuck with your campaign, reset, the algorithm, then start over with inefficiencies and a lot of learnings lost (you’l never get that first algorithm’s performance back until you get new winning creative).
The people who say to clone campaigns or ad sets say this because it’s a way you can “game” the algorithm, in theory. If your budget is small enough and the audience big enough, it is true that your first few hundred dollars spent could make or break that instance of an ad set because it will find pockets of people within the audience and inform the campaigns CPMs & performance from there.
In my campaigns, I am testing a new audience every month. Winners are added into the “main campaign.” I want to create the biggest possible audience in that main campaign because my budgets are pretty large, so I go for a discounted CPM and count on the machine learning to optimize quickly because I generate enough data. I have, in some circumstances, cloned adsets and campaigns, but even anecdotally speaking I have not found this to be a winning strategy. Eventually, performance will be performance when you are spending enough dollars, and it’s more valuable to have the algorithm optimize once a month.
Establish a baseline this month with a SINGLE scaling strategy. Measure where that gets you, and then run it for another month if it is profitable. If it is profitable, run it again the 3rd month and then also run a different scaling strategy. If it is profitable, you are winning & should establish a baseline and begin recording your playbook.
Log in to reply.