Rando.gg

We tested 547 browser games for ads. 377 of them served ads.

Free browser games are free because somebody is being paid. Usually that is fine. Sometimes it means a thirty-second unskippable video before a game about stacking blocks. We wanted to know how common that actually was, so we measured it rather than guessing.

The method

We loaded each of the 547 games in our catalogue in a real Chrome instance and recorded every network request it made. Any request to a known ad network — Google IMA, DoubleClick, AdSense, Prebid, Amazon, Criteo and about fifteen others — counted as a hit. No judgement about how intrusive the ad was; just whether the machinery was there.

377 of 547games contacted at least one ad network

What that looks like in practice

The most common stack by a distance was Google IMA plus DoubleClick plus Ad Manager, which is the combination behind pre-roll video. Header bidding — Prebid, Criteo, Rubicon, AppNexus — showed up on the heavier sites, which usually meant the game was wrapped in a page selling several ad slots at once.

One game contacted eleven distinct ad networks and issued 464 requests before we stopped counting. The game itself was a browser port of a block-pushing puzzle.

The host predicts the answer

The single best predictor of whether a game served ads was not the game. It was the host it was served from. Distribution networks — the services that syndicate a game to hundreds of portals — bundle their ad SDK with the game, so every title from that host behaves identically. Games self-hosted by a portal on its own CDN were far more likely to be clean.

This has a practical consequence if you are building a catalogue: filter on the iframe host, not on the site you found the game on. A portal will happily embed syndicated titles alongside its own, and only its own are clean.

Why we bothered

Rando.gg is a shuffle button. If one press in three lands on a video ad, the button is not fun any more — the whole product depends on the gap between pressing and playing staying short. So the measurement is not academic for us; it decides what stays in the catalogue.

Not feeling this one?

Play something randomOne tap, a game you haven’t played.

More notes