Why one American publisher bought three Istanbul studios

Between May 2018 and October 2020, Zynga bought three games studios. All three were in Istanbul.

Zynga was not the only buyer looking. Rovio and SciPlay bought Turkish hyper-casual studios too, in İzmir and Ankara rather than Istanbul.

That is well over two billion dollars, spent in one city, inside thirty months, by a company headquartered in San Francisco.

Gram Games, 2018

Gram Games was the first, acquired in May 2018 for a total purchase consideration of $299 million, a figure that was widely reported at the time as a $250 million deal because that was the cash portion.

Gram had built 1010! and Merge Dragons!, and the merge mechanic in that second game turned out to matter a great deal to what mobile puzzle games looked like over the following five years. Buying it was Zynga's first real bet on Istanbul, and at the time it looked like a reasonably sized acquisition of a studio with a good game.

In hindsight it was a test.

Peak, 2020

Two years later Zynga announced it was buying Peak for $1.8 billion, roughly half in cash and half in stock. Peak had two puzzle games, Toon Blast and Toy Blast, more than 12 million average daily players between them, and about a hundred employees. The deal took Zynga's own daily active users up by roughly 60%.

It remains the largest exit in Turkish startup history, and there is a whole post about it.

Rollic, 2020

Four months after the Peak announcement, Zynga closed on Rollic, paying about $180 million in cash for 80% of the company, with the remaining 20% to follow over three years based on performance.

Rollic had been founded in December 2018, by Burak Vardal, Deniz Basaran and Mehmet Can Yavuz. So it was about twenty-two months old.

In that time it had passed 365 million downloads and had put eight games at number one or number two on the US App Store. In one quarter of 2020, the two most downloaded games in the entire US App Store were both Rollic titles.

Vardal's comment on joining was characteristically low key: "We are memorably excited to be joining Zynga, a perfect home for Rollic's culture and core values."

What was actually being bought

Three studios, three different kinds of game, one buyer. Ask what they had in common, because the answer is not the games.

Gram Games made merge puzzles. Peak made match-and-collapse puzzles with enormous production quality and years of live operations behind them. Rollic made hyper-casual games with lifespans measured in weeks.

What all three had was the ability to spend money on user acquisition and get more money back. That is the whole skill. In hyper-casual it is nearly the entire business: the games are cheap to build, dozens get tested, and the winners are the ones where the cost per install comes in under what the player will eventually be worth. In puzzle it is slower and the numbers are bigger, but the discipline is the same.

Istanbul produced an unusual concentration of teams who were very good at this, partly because the first generation trained the second. People who learned it at Peak or Gram went off and did it again somewhere else.

The bit that is different now

Every one of these three companies paid for its own user acquisition out of its own revenue, in the years before Türkiye had a serious support programme for the sector. Decision No. 5447 only came into force in April 2020, after Gram Games had already been sold and weeks before the Peak announcement.

The capability that Zynga paid over two billion dollars for was efficient ad spending. Under Decision No. 10962, that spending is reimbursed at 50%, and at 70% when the campaign targets one of 20 target countries including the US, where all of these studios were buying their most valuable players.

It is a strange thing to sit with. The single skill that made Istanbul worth two billion dollars to an American publisher is now the skill the Turkish state pays you back for using.

If you are running that playbook, work out what it is worth.