Why hand-made thumbnails don’t scale past 500 games.
A lobby of 200 games is a design task. A lobby of 2,000 is a logistics problem. The work that feels manageable at launch quietly turns into the thing nobody has time for — and the maths behind that shift is more unforgiving than it looks.
Every operator’s lobby starts the same way. A designer opens a folder, picks a house style, and starts cutting thumbnails: a character, a logo, a background, a frame. The first fifty look great. The first hundred still look great. Somewhere between the five-hundredth tile and the thousandth, the wheels come off — not because the designer got worse, but because three costs that were invisible at small scale all come due at once.
Cost one: the catalogue grows, the output multiplies.
A thumbnail is never one file. A modern lobby needs a portrait tile for the mobile grid, a landscape card for the desktop rail, a wide banner for the hero slot, and often a square for search and recommendations. That is four renders per game before anyone mentions a seasonal variant or a “NEW” badge. One thousand games is four thousand image files that must be named, versioned, colour-checked and kept in sync. Add a fifth size next year — because a new front end wants it — and you are not adding one job, you are re-cutting the entire back catalogue.
Hand production scales linearly with games times sizes. The catalogue grows every week. The team does not.
Cost two: drift.
The second cost is the one operators feel but rarely name. When every tile is made
by hand, every tile is a slightly different decision. The logo sits two per cent
lower on Tuesday’s batch than Monday’s. One designer likes the character
bleeding off the top edge; another centres it. The provider accent is
#E4002B in March and #E10028 in September because someone
colour-picked from a new press kit. None of these is wrong on its own. Together they
turn a lobby into a patchwork, and a patchwork reads as cheap to the one audience
that matters: the player deciding whether this casino looks trustworthy enough to
deposit into.
Drift is invisible per tile and obvious in aggregate. It is also nearly impossible to fix by hand, because fixing it means re-opening hundreds of source files and re-making the same decision identically every time — the exact thing humans are worst at.
Cost three: the launch delay.
Providers ship new titles constantly — often dozens a month across a full aggregator feed. Each one arrives as a press kit: a key visual, a logo, maybe a background plate. Before that game can go live in the lobby with art that matches the house style, someone has to cut its thumbnails. So the art queue becomes the thing standing between “the game is integrated” and “the game is earning.”
At a hundred games a month, that queue is a part-time annoyance. At the volume a serious multi-provider operator runs, it is a full pipeline with a backlog, and every day a title sits in it is a day of revenue the operator has already paid to license but cannot display.
Why “just hire another designer” doesn’t fix it.
The instinct is to throw people at the queue. But adding designers adds throughput and adds drift — more hands making more independent decisions — and it does nothing about the re-cut problem when a new size appears. You end up paying more to make the consistency problem worse. The task isn’t really a design task any more. It is a rendering task wearing a design task’s clothes.
What a template system actually changes.
The fix is to separate the two things that got tangled together: the decision (how this brand’s thumbnail should look) and the production (drawing it for every game, at every size). Make the decision once, as a template — anchors, safe areas, how the key visual fits, where the logo locks, which band and glow sit behind it — and the production becomes a function. Feed in a game’s three assets, and every size comes out the same way, every time.
That collapses all three costs at once:
- Multiplication stops hurting. A new size is a new layout in the template, applied to the whole catalogue in one pass — not a re-cut.
- Drift goes to zero. There is one decision, so there is nothing to drift from. The thousandth tile is laid out by the same numbers as the first.
- The launch queue disappears. A new title’s thumbnails are generated the moment its assets land, not whenever a designer gets to it.
This is the whole reason thumbs.store exists. We render thumbnails from a provider-aware template instead of cutting them by hand — four sizes from one set of numbers, the same frame logic across a 12,000-game catalogue, new games added as the feed grows. The designer’s judgement still sets the look. It just stops being the bottleneck for every file that look has to appear on.
The honest threshold.
None of this means hand-made art is wrong. For a boutique lobby of a few dozen hero titles, a designer and a folder is exactly right — the control is worth the effort, and the volume never bites. The threshold is roughly where catalogue size times size count outruns the hours a team can give it without the queue or the drift becoming the player’s problem. In our experience that lands somewhere around five hundred games. Past it, the question isn’t whether to template. It’s how long you can afford not to.