Grand Games Spotlight: Sniper on the Game, Shotgun on the Ads

TL;DR: Grand Games launched Block Out at roughly $300K a day, iPhone only. The playbook is a contradiction: being strict on which games to make (nearly a year without greenlighting one), going nuts on the ads (1,000+ creatives for a single launch).

It's the same logic both times, a great game in a proven category is knowable, so be patient there, but the winning ad isn't, so flood the space until the winners surface. You can't predict a hit. You build the machine that finds it.

----------------------------------------------------------------

That is Block Out, a sort puzzler from Grand Games, six months out from launch and scaling on iPhone alone. The figures come from the teardowns the UA community runs on launches like this, so treat them as estimates. The exact decimals do not matter. The shape does. A game most people outside the industry have never heard of is scaling like a hit, on one platform, faster than almost anything around it.

That does not happen by accident. And Grand Games has done it before.

Two more numbers explain the first three. Before this launch, Grand went close to a year without greenlighting a single new game. Then it hit the market with more than a thousand ads. The whole playbook lives in that contradiction.

---

Keep exploring with Sett:
* How the UA manager became a scientist, and why the creative is now the targeting
* The two tracks every creative team runs: exploit the winners, explore for the next.
* What UA work becomes when your team is a set of AI agents you direct.

---

The studio behind the numbers

Grand Games is two years old.

Two years ago it was one founder designing puzzle levels with his back against the wall. When Balderton wrote the fastest Series A check in Turkey's history, the whole company was 20 people in a 70-square-meter office. Today: a $500 million annual run rate, six games shipped, still only about 75 people. That is north of $6 million of run rate per employee.

The sharpest signal in the $103 million Grand has raised is who got in. Nobody new. The same investors led the $70 million round at a higher valuation and took every share offered. The story comes from Bekir Batuhan Çelebi, the co-founder and CEO, in a long interview with Deconstructor of Fun.

That trajectory is the context for the launch. Block Out is one output of a machine built to produce this on purpose. There is no luck in it, and the machine is worth understanding before the ad account is.

First, the game is actually great

There is a temptation to skip this part, and it is the part that makes the rest work.

Block Out is not a thin reskin pushed out to game the algorithm. By every read it is one of the most polished games in its category, a sort puzzler built cleaner and tighter than the field. That matters, because everything that follows is an accelerant, and an accelerant on a mediocre game just burns money faster. Massive exploration does not rescue a bad product. It exposes it.

Grand knows this, and the discipline starts before a single line of code. A new game only begins if it clears one filter: real, impactful differentiation in art, level design, or core product. As Deconstructor of Fun put it, making an existing game slightly better isn't a strategy. The idea has to come from a specific pain point, someone who plays a game they like and can name exactly what feels off, and it has to survive a check against Sensor Tower market data before anyone prototypes.

Mark Pincus has a name for what comes out the other side, and a new book about it. He calls it Proven, Better, New. Copy what is proven, the exact experience people already love. Make it better, until ten out of ten people say yes. Then add one genuinely new thing.

The heresy is in the first step. Pincus founded Zynga, and by one count, eight of his ten major launches became hits. His advice is to get comfortable copying what works so you can spend your energy where it counts. "Only a master craftsman," he says, "can spot excellence and be comfortable copying it."

That is the foundation. Proven and Better, executed to the top of the category. It is the price of entry for everything else, and most of the market never pays it.

Where the New went

Run Block Out through that framework and the third word is almost missing.

In the product, it is Proven and Better with the New turned nearly to zero. Pincus has a name for that too. It is how he describes Supercell's Hay Day, which he calls FarmVille for mobile. Proven and better, no new at all.

A studio that refuses to gamble on novelty in the product still has to put the New somewhere. Grand moved it into the creative.

Pincus's own rule for handling anything new is that all new fails, so you have to try many, many variants of each new idea. He wrote it for product teams. It describes Grand's ad accounts just as well.

The number that explains the launch

Look at the creative output, before and after.

The last Grand launch ran something like 20 creatives. Block Out went out with over 1,000, plus 260 playables and 888 videos.

That is more than fifty times the creative volume, brought to a single launch by the same studio on the same kind of game. What the volume buys is the firepower to roll the dice enough times that the winners surface on their own, instead of getting guessed at in a room.

Sniper; shotgun

Here is the contradiction that turns out to be the whole point.

At the level of which games to make, Grand is the most patient studio in the market. Five autonomous studios of roughly fifteen people, each one owning its game end to end. Five shots that all have to count, which is why between one launch and the next the company went close to a year without greenlighting a single new game. Batuhan describes his own lane as neither Dream Games, all perfectionism and cadence, nor Rollic, all volume and machine-paced output. "We're kinda fast with kinda good quality," he says.

The reasoning behind the shrug is testing logic. Ship a game too rough and the market hands you a false signal of what it could have been. A sniper in a market full of studios firing shotguns at the wall.

Then that same sniper walks up to the launch and empties a thousand rounds into the creative.

Both are the correct move, and they are correct for the same reason. You can be disciplined about the foundation because a great game in a proven category is knowable. You cannot be disciplined about the winning ad, because nobody knows which creative works until it runs. Grand spends its patience where judgment pays, on the game, and spends volume where judgment fails, on the creative. Restraint and firepower each point at the half of the problem where they work.

This is the Chinese model, ported west

If the volume looks familiar, it is the launch model the best Chinese studios have run for years, and the one the West kept filing under hyper-casual. Cheap installs, disposable games, someone else's trick.

Block Out is the counterargument. Point that same creative-first launch at a genuinely great IAP game and it does not cheapen the launch. It accelerates it. A premium puzzler scaling to these numbers on a thousand-creative launch is real validity for a model the West spent years underrating.

What massive exploration actually buys

A thousand creatives is not one ad printed a thousand times.

The range is the point. Concept after concept, colorful, genuinely diverse, altered gameplay, deep variants.

Each of those is a bet on a different kind of player, a different reason to install. Run a thousand of them and you are not hoping to find the winning angle. You are covering enough of the space that the winners cannot hide. That is what testing at this scale gives you. Speed to the answer.

And speed to the answer is what lets you do the next part.

Then they floored it

For the better part of half a year, the entire launch was small on purpose. One campaign, one network, one country, one platform. AppLovin, in the US, on iOS. Nothing else.

That looks like caution. It was patience with a purpose. They held the footprint tight while the creative engine did its work, and the moment the search had surfaced its winners, they moved fast. Tier-one countries in late April. A global launch by mid-May. Somewhere near 50 countries today. Months of patience, then weeks of acceleration.

You can only accelerate like that if you already know what works. The thousand-creative exploration is what earned them the right to floor it. Distribution did not lead the launch. The creative did, and distribution followed once the answers were in.

The ads they added later run on the same logic. Block Out was sold as a game with no ads, and for a while it had none. Then the interstitials arrived, because interstitials unlock the blended-ROAS math that lets you buy users at this scale. Their hit Magic Sort draws somewhere between 32 and 36 percent of its revenue from ads, a widely cited figure, almost all of it interstitials after levels.

Every move waits for the last one to pay off, and then it does not wait at all.

Why it is a model and not a fluke

One hit is a hit. Two off the same playbook is a system.

The ad account is only half the system. The studio around it is the rest. Five squads of fifteen, small on purpose, because Batuhan found that above twenty people the sense of individual ownership evaporates and product management turns into people management. Every Head of Studio ships without asking a founder. Shared code and AI-assisted development let a feature built for one game move into the next in three days. That is the engine that lets a team that small run a game at real scale, greenlight with a sniper's patience, and still launch with a thousand creatives when the moment comes.

Grand can run all four steps again, and it raised the money to keep running them. The skill on display is a repeatable way of finding out which great games deserve the firepower, then bringing enough firepower that the winners cannot stay hidden.

You cannot predict a hit

Grand did not predict Block Out would work. Nobody can predict a hit.

What they did was build a machine that finds one. Start from a genuinely great game that cleared a real differentiation filter. Put the New where it is cheapest to be wrong, across a thousand creatives. Hold the launch small until that testing answers, then accelerate hard.

Proven and Better are the cost of entry. Everyone can copy and polish a proven mechanic. The winning happens in the search, and that is where most studios fall short, because you cannot staff a thousand creatives and 260 playables by hand.

That is the firepower we built Sett to add. Agents and a system that explore the angles, build the playables and the videos to test them, and feed every result back into the next round. Not to invent the hit. To find it the way Grand found Block Out, at a scale a team could never reach by hand.

You still cannot name the winner before you have seen it. But put enough exploration behind a great game and you can be almost sure it is in there, and that you will be the one who finds it first. Grand spent two years and $103 million building its machine. Be the sniper on your game. Sett is the shotgun.

---

Sources: Block Out launch figures (revenue, downloads, creative counts, geo timeline, ad-revenue share) are estimates and details drawn from a Two and a Half Gamers teardown of Block Out, June 2026, treated as well-informed industry estimates rather than audited figures. Studio history, funding ($103M total, $30M Series A, $70M Series B), structure (five studios of ~15, ~75 people), the greenlight filter, Sensor Tower, and all Batuhan Çelebi quotes are from his interview on Deconstructor of Fun, "Grand Games from $0 to $500M in Two Years," by Michail Katkoff. The "Proven, Better, New" framework and all Mark Pincus quotes are from Pincus's book Life at the Speed of Play (2026) and surrounding interviews.