The other day I had a conversation with three co-founders, for which we ran their first UA campaign for a newly released title. The game is actually close to my own interests and I enjoy playing it on weekends. Maybe it is even a little too “nice”, but that is a monetization discussion for another post.

The initial campaign numbers are flowing in and they are good. Good early numbers are exciting. But they are not the same as scalable confidence. The cohort numbers are still young, mainly D1 to D7. So quickly the question from one of the founders came up: 'can we scale?'. And this is where my caution kicks in. We often see the so-called golden cohort, where the initial cohort shows the best results compared to those that are following. However in this case there is no big brand or community behind it. Early numbers like this can be a real signal.

The ROI numbers we are seeing are gross. The real picture is usually less visible. When we start including store fees, taxes, overhead, but also our fee for the UA operations and creatives (we also need to pay our bills), etc. So before scaling, we need to have a discussion with the team on the total costs beyond ad spend, and deduct that total amount from the total revenue coming week over week. That weekly contribution view gives us the fuller picture and, more importantly, confidence.

And with that confidence new decisions need to be made, and as a UA manager you are standing a crossroads:

Do we let cohorts mature?
Do we scale the current campaign?
Do we segment the campaign to understand where quality comes from?
Do we recreate the setup on another network?

I believe that at this point an experienced UA manager should ask himself the question; which levers do I have left to further improve. The confidence grows from knowing there are still clear levers to improve. The campaigns only just started with good early signals, but we can still improve the store conversion, creatives can become stronger, campaign setups can be tweaked, the network mix can expand. If the numbers are already strong before these things are optimized, that makes the opportunity more interesting.

So before scaling ask yourself these questions;

Are the cohorts mature enough to trust the direction?
Do we understand where the strong performance is coming from?
Does ROI still work after real costs and margins?
Can we repeat the result with more creatives, geos, campaigns, or networks?
Is the product or monetization expected to change over time?
Are there obvious optimizations left that can improve the model?
What is the next best step to reduce uncertainty?

Scaling is not binary, not a "yes or no". It's a sequence of decision making to build confidence step by step. The goal is not to scale because the numbers look good. The goal is to scale when you understand why they look good and how likely they are to hold.