Once launch week is over, a daily game's visitors come from three places: people typing its address or opening a saved link, people searching its name on Google, and other daily games recommending it. In October 2026, 85% of DailyGeo's Google clicks were searches for its own name, cross-promotion brought 55.3% of Geopolitix's visits, and social networks and AI assistants each stayed under 2% for all seven games. Players sent by other games came back the next day at 8.8%, about as often as players from Google (8.4%).
Disclosure: these seven games are made by relaydle's founder, and the cross-promotion measured here is mostly the relaydle rail, which we run. The numbers are our own analytics; the queries are listed in the method note. For what happens in launch week itself, see how to promote a daily game.
Where did the visits come from in October 2026?
Share of visits, by the source of each visit's first page:
| Game | Direct or none | Bing and other search | Cross-promotion | Social | AI assistants | |
|---|---|---|---|---|---|---|
| DailyGeo | 57.7% | 34.7% | 3.5% | 3.5% | 0.1% | 0.3% |
| Who's More Famous? | 48.6% | 44.7% | 2.0% | 4.0% | 0.3% | 0.0% |
| DailyMetro | 69.2% | 22.9% | 2.6% | 2.4% | 1.7% | 0.2% |
| Devine le prix | 56.7% | 32.8% | 2.5% | 6.3% | 1.0% | 0.1% |
| Le Bon Rang | 49.3% | 12.7% | 4.1% | 30.7% | 0.4% | 0.6% |
| Geopolitix | 34.3% | 6.3% | 2.1% | 55.3% | 0.1% | 1.5% |
| DailyFoot | 49.1% | 1.4% | 1.7% | 47.3% | 0.5% | 0.0% |
Social groups Reddit, X, Facebook, Instagram, TikTok and YouTube together; DailyMetro's 1.7% is mostly Reddit. Rows do not add up to exactly 100% because of a small "other" group. "Direct or none" includes in-app browsers that strip the referrer and navigation inside the game's own site.
Is Google traffic to a daily game really search traffic?
Mostly not in the sense of being found. It is people who already know the name. In Search Console for October:
| Game | Google clicks | Share that searched the game's name |
|---|---|---|
| DailyGeo | 34,859 | 85% |
| Who's More Famous? | 16,725 | 89% |
| DailyMetro | 11,456 | about half or less |
Name searches include typos ("dayligeo", "daily geo", "whos more famous"). The two shares are lower bounds: Google hides rare queries, and those were counted as non-name. DailyMetro's hidden share is too large to say more than "about half or less"; its visible non-name clicks are generic searches such as "jeu metro".
For DailyGeo and Who's More Famous?, Google mostly works as the address bar of people who already know the game; this data does not say where they first heard of it. A new game with no name recognition should not expect search to find it players on its own.
Do social networks and AI assistants send players?
Very few, as referrals. Apart from Reddit on DailyMetro (1.3% of its visits, mostly from the Reddit app), every social network stayed under 1% of visits on every game. A viewer who types or searches a game's name after seeing it on social media lands in the direct or Google column, so this measures referral clicks only, not the reach of a post.
AI assistants sent between 0.0% and 1.5% of visits, mostly from ChatGPT (it tags its links with utm_source=chatgpt.com). Visitors from them were also among the least likely to return: 5.2% the next day (562 first-time visitors), against 8.4% for Google.
Do players from cross-promotion come back?
About as often as players from search. Next-day return of first-time visitors, all seven games pooled:
| Source of the first visit | First-time visitors | Back the next day |
|---|---|---|
| Bing and other search | 6,367 | 10.0% |
| Direct or none | 82,212 | 9.7% |
| Cross-promotion | 16,043 | 8.8% |
| 68,164 | 8.4% | |
| 566 | 5.7% | |
| AI assistants | 562 | 5.2% |
| 351 | 2.8% |
The 0.4-point gap between cross-promotion and Google is within noise, and the pooled figures mix games with different base rates: most cross-promotion arrivals went to Geopolitix, DailyFoot and Le Bon Rang, most Google arrivals to DailyGeo. Game by game there is no fixed winner: cross-promotion returned better than Google on four games and worse on two (DailyGeo 8.2% against 9.9%, Devine le prix 7.6% against 8.5%). Two of the four wins are partly Google's weakness: on DailyMetro and Le Bon Rang, Google visitors returned at only 4.2% and 2.9%, partly because of searches that were never looking for a game. These are floors measured with a cookieless tool; the retention post explains why and gives each game's overall rate.
The useful reading for a creator: players who clicked from another daily game's result screen returned about as often as visitors from Google. They are not low-quality clicks.
Which source should a small daily game work on?
Here, the split follows size. The three biggest games got 2.4% to 4.0% of their visits from cross-promotion, alongside 22.9% to 44.7% from Google. The three smallest got 30.7% to 55.3% from cross-promotion and 1.4% to 12.7% from Google. For DailyFoot, with 183 visitors a day, other games brought 47.3% of visits while Google brought 1.4%.
A new daily game has no name to search yet, and social referrals barely registered for any of these games. One channel that does not depend on being known is other daily games recommending yours to players who just finished theirs. That is what the relaydle rail does, for free: submit your game, then apply to the rail. The rail's own data shows how many clicks that produces and how cards get chosen.