1. You've been in mobile UA since 2011 and you've seen hundreds of game launches from both inside studios and as an external consultant. In your experience, what is the single most common mistake teams make when deciding whether to keep investing in a game post-launch?

2. The go/pivot/kill framework sounds clean in theory. What does the actual decision-making process look like in the room? Who's in the conversation, and who often gets left out?

For me CPI is only relevant when running campaigns that are not CPI campaigns. You do not know what traffic you are buying there, you are only paying for cheap, and the network might be buying users that are completely uninterested in your product. Your CPI might radically change when optimizing towards “tutorial completed”, for example. 

Day 1 retention is a nice early indicator of user interest and performance, but you do need the users to stay longer if you want to be able to have a viable business. 

The golden metric for me is ROAS, because even if your retention rate is not ideal, if the rest of the metrics combined are able to get a positive ROAS, then you should be good to go.
That said, profitable isn’t always the same as "successful". Certain studios might need a minimum scalability to be profitable on a studio level. So in reality the conversation is a bit more complex, it needs to start by defining the meaning of the word ‘success’ for each company, then depending on that the team will set up their needs for the rest of the game KPIs.   

There are many reasons why I always preach about creating long soft launch processes (at least three months). One of them is that long-term data might be surprisingly good, and that might completely change your overall results. So for me, early retention can sometimes be counterintuitive.

We hear that we need a minimum of 30% day 1 retention to have a viable hybrid casual game, but if your day 1 is lower and at the same time your long-term retention is exceptionally good, then you might still have a really good product! Niche but sticky.

We have to remember that we cannot look at metrics as a unique thing. If we do so, we commit big mistakes — one bad metric or one good metric alone does not make a good game. It is the combination of all of them that will make your LTV good enough to scale.
Especially if you are a studio that is just starting, do not get obsessed with big studios' goals. As said before, success metrics need to be set taking into account cashflow, studio size, and short-term targets.

5. "Pivot" is often used as a softer word for "we don't know what to do." When is a pivot a genuine strategic move versus a way of avoiding the kill decision?


I wouldn't call it "we don't know what to do." This industry runs on testing, and a pivot is usually the output of a specific diagnosis.

Say your team is testing a card-based mechanic. In soft launch, CPI comes in too high relative to what your monetization can support. The data might show the real issue isn't traffic quality or creative. The mechanic itself isn't generating enough engagement to monetize at that cost. That's a clear, specific reason to pivot. Maybe you switch monetization strategy, if your team has that expertise. Maybe you switch the mechanic itself, cards to a board (board games are usually cheaper than card games), for example. Either way, you're acting on a known cause.

There will be times where the team might be unable to find where the problem is, and then decides to make a radical move to try to keep the game alive. The team might have a hypothesis but can't back it up with data. Perhaps the results of the test prove that hypothesis right. Then, eureka! you learned something new!

6. You operate as an embedded UA team inside studios. When you're brought in post-launch, what do the first 48 hours of diagnosis look like? What are you actually looking at?

First thing we do is download and play with the game. We make all type of questions to the team: what is their goal, why this app, what are their hypothesis and their competitors.

Then we look at the competition,  and lastly we do an audit, we look at the MMP and network set up, the stores…

We look for errors and opportunities, and we try to learn everything from that specific market, since we need to work with very different games, sometimes from the vertical we play on our free time, but others are completely different.

7. What does a kill decision actually cost - beyond the obvious sunk costs? Are there hidden organizational costs that teams consistently underestimate?

Beyond production costs and opportunity cost, there's a cost teams consistently underestimate: the cost of the team itself.

Hiring a team to run a soft launch is great if it works. Those same people move into the scale phase with the game they already know. But if it fails, you're left with a team you now need to fire or relocate, which is its own cost — financial and organizational.

What happens more often is that companies don't hire at all. They just add the soft launch workload onto their existing team's plate, reorganizing priorities and slowing down delivery on their main, already-live games. That's a real cost too, it just doesn't show up on the project's own budget line. It shows up later, as a delayed update or missed content on the games that were actually making money.

This is exactly why studios bring us in during these early stages. Either they want a senior team with soft launch experience to carry the new title through its first months, without committing to a permanent hire before they know if the game works. Or they use us as interim support on their scaling games, so their in-house team can put more hours into the new title in development instead of splitting focus.