Stuck Attribution, Black Box Algorithms, and Why UA Is Just an Amplifier
Most UA conversations stay on the surface. Budgets, channels, ROAS, the visible layer everyone debates on LinkedIn. This one goes underground.
Noam Gohary spent two decades on the side of mobile gaming marketing nobody wants to talk about: the data pipes, the attribution logic, the marketing tech that decides whether the dashboard you're staring at actually reflects reality. His path runs through Matomy, Aditor, Playtika, and most recently Scopely, where he was Director of Product & Marketing Tech. Across all four, he built the same kind of team: two or three deeply technical UA people who can talk to Google's solution engineers and the studio's developers in their own language. The kind of team that fills a gap most companies don't realize they have, until something quietly costs them millions.
This is a conversation about what's hiding under the floorboards. What's stuck. What's missed. And where AI takes the whole stack in 2026.
From Economics to a Second Laptop
Noam didn't grow up dreaming about mobile games. He came in through economics, financial consulting, and a stint as an analyst at Nielsen, where he discovered exactly what he didn't want to do.
"I want something with much quicker interaction, where I can do something and see the results."
A friend pulled him into Matomy at a moment when, in his words, half the country worked there. He started managing the long tail of streaming sites, about five hundred of them at once, which left him two options: automate or drown.
"I built two Excel files. I'd dump the data, click a button that ran a script behind it. It would optimize everything, tell me what to close, what to move. I got them to give me a second laptop so I could split it."
The pattern that built his career was already there. Find the tedious thing. Automate it. Then walk around asking questions everyone else thinks are weird.
"My curiosity came from my first positions, even in economics. It was about asking questions outside of what I do."
The Attribution That Never Grew Up
Here is what nobody in the industry wants to say out loud. The ad networks turned into something close to magic over the last decade. Attribution did not.
"Attribution is the only piece of our ecosystem that hasn't really changed since it was created. We went from rocks and hammers to spaceships. Attribution started well, somewhere in the middle, like a car. But it's still a car. It never evolved into a spaceship."
The whole industry runs on last-touch. The default settings inside MMPs become law because most teams don't understand the logic well enough to challenge them. Defaults stop being recommendations and start being rules.
"If a company this big, that all the big names are using, says this is the default setting, then they did their research. So I leave it at that."
That deference has a price. Not catastrophic, never catastrophic, just slow.
"I don't think things completely break. It's not that things break, it's that things are being missed."
Black Box Problem
The harder admission lives one floor down. Senior UA people, the ones with a decade in the trenches, no longer fully understand the algorithms they spend on every day.
His test is brutal. Ask a Google rep what actually moves the algorithm. Is it CTR? Early purchases? Creative rotation? Watch what happens.
"They don't know, because the algorithms got so complex, so much of a black box."
If UA can't explain it inside its own walls, it can't explain it to a GM or a CEO either. That gap is where bad strategic calls get made and nobody can trace why.
The Three-Headed Beast
Most studios think they have a structural problem. Noam thinks they have a translation problem.
"This three-headed beast has to mesh together well for anything to run perfectly."
The three heads are the studio, the game team, and UA. When they don't talk, the cost shows up as wasted experiments, Facebook data misread as a Facebook problem, and early funnels nobody on the game team cared to optimize. The early funnel is exactly where UA lives, and exactly where game teams stop paying attention.
"UA is really built on the first thirty days of the game, to exaggerate. Most game teams don't care as much about the first thirty days, because what they care about is the money makers, and the money makers are the whales."
The whales pay the bills. The first thirty days build the funnel that finds the next whales. That mismatch in what each side optimizes for is, in Noam's view, where most studios are quietly leaving real growth behind.
The New UA Manager Is a Diplomat
The Excel-deep, manual-optimization UA manager is fading. What's replacing him is closer to a politician.
"The UA manager becomes what I call public relations. They have to go and sell the product manager inside the game on why it's important to change something in the funnel."
The other half of the role is network conversations. The ones that surface useful intel before the rest of the market catches up.
"You need the relationships. You need to be a people person."
UA Is an Amplifier, Not the Engine
This is the line many UA people don't want to hear, and the one Noam doesn't soften.
"A shit game doesn't work, no matter how good the UA is. UA is an amplifier of the game. If the game is good, UA can make it explode."
His proof point is Monopoly Go.
"The game is insane. The cycles, everything works. UA was extremely correct, and the team there did some magic, knowing exactly where to put the dots."
UA doesn't make or break a company. It's the multiplier between a good game and a phenomenon. Get UA right on top of a great product and you get Monopoly Go. Get it right on top of a weak one and you still don't have a hit.
Where AI Takes This in 2026
Noam splits the question in two. What AI does to games. What AI does to UA.
On the game side, personalization stops being about segments and starts being about individual profiles. Lobbies, menus, characters, skins, the store layout, all tuned to the single player.
"Change the whole menu. We know what the lobby is, we know how powerful the menu is. Change the whole thing to match what I like to do."
On the UA side, smaller teams. Smarter flagging. More prediction.
"We'll see things that shrink UA teams and make them more efficient. I don't think UA teams will disappear. It always needs some human touch."
The bigger move, in his read, isn't OpenAI. It's Google.
"Google is really good at taking tech and making it actionable. Google built trust. That's their biggest weapon right now."
Wrap Up
The thread under this entire conversation is that the foundation of mobile games marketing is shakier than most teams admit, and the cost is never a crash. It's a slow leak. Attribution that never grew up. Algorithms that turned into black boxes. Studios, game teams, and UA still operating like three countries with three languages and no shared dictionary.
Noam's read on the corporate world is honest in a way most industry guests refuse to be. Big companies making real money have no incentive to rock the boat. The next leap probably comes from a startup willing to break the system, or from one big player whose leader is brave enough to take the risk.
"It's like going on an adventure that might hurt you a lot, and a lot of people are very afraid of that."
His take on AI is the opposite of doom. AI doesn't replace UA, it amplifies it, the same way UA itself amplifies the game. The teams that adapt won't disappear. The teams that don't will be left behind anyway.
And the discipline he carries from Aditor through Playtika and Scopely lands on one principle: don't build the cathedral, ship the MVP plus, and trust what people actually touch.
"If they don't touch it, if they don't tinker with it, understand that it's not good."
