Steam Next Fest preparation checklist: the multilingual version

You have a demo, a date, and a slot in Steam Next Fest. For most games this week is the single largest burst of attention they get before launch, and Valve's rules have generally allowed a game one appearance, before release — so the preparation window matters more than the week itself. Whatever is not ready when the event opens is not going to be part of it.

The part that is easy to underestimate is where the attention comes from. Next Fest is a global event on a global storefront. The people browsing it are not filtered to your language, and the first thing most of them see is a capsule image and a one-line description rendered in whatever language their Steam client is set to. Anything of yours that does not exist in their language is simply not part of their decision — they do not struggle through it, they scroll past.

This is a checklist for that reality: which languages to commit to, what to prepare in each of them, how to run the week itself, and what to read out of the results afterwards. Dates, registration deadlines, and eligibility rules have changed between editions, so confirm the current ones in the Steamworks documentation and the event announcement before you plan anything around them.

What the week actually does to your store page

During the event, registered demos are gathered into one browsable place for a limited period, and the traffic that produces is unlike your normal store traffic in two ways. It is front-loaded — the opening days carry the heaviest browsing — and it is overwhelmingly cold. Most visitors have never heard of your game, are sampling many demos quickly, and will decide whether to look further in a few seconds.

That changes the job of every surface you own. A player who follows your devlog will forgive a rough English description because they already care. A player who arrived from an event grid three seconds ago will not, and they are the majority of the audience for this particular week.

It also changes the cost-benefit of translation, which is the reason a multilingual checklist is worth making at all. The work of translating a short description or a demo's UI is the same whether you do it in a quiet month or before a fest — but the number of people who see the result is concentrated into a few days. If you were ever going to translate the front of your game, this is the moment when the same effort is worth the most.

Decide the language list by what can be finished, not what you want

Translation is a lead-time item, not a last-week item. Text has to be exported, translated, reviewed, put back into the build, and then read again inside the game where it can overflow a button or hit a missing glyph. Each of those steps has a waiting period attached to it, and none of them compress well when you are also fixing demo bugs.

So build the language list backwards from the registration deadline rather than forwards from ambition. A language belongs on the list only if three things can all be true before the event opens: the store page is translated, the demo's own text is translated, and the supported-languages declaration on the page honestly matches what the build does. A page that promises a language the demo does not have produces exactly the kind of public disappointment you do not want during your one week of visibility.

Fewer languages done properly beats more languages done roughly, and this is more true during a fest than at any other time, because the feedback loop is immediate and public. Broken text gets found within hours — in discussion posts, in Discord, in a stream where someone reads your untranslated menu out loud to an audience.

• Count the demo's actual text first — a demo is usually a small fraction of the full game's word count, which is what makes it affordable to translate at all
• Put the store page's short description and About section into the same batch, since they are small and they gate everything downstream
• Reserve time after the text is in the build for the fix pass, not just for the translation itself
• If a language will not be finished, drop it from this fest rather than shipping it half-done — the demo can gain languages afterwards and keep earning

Example shape of a deadline-driven plan (scale the intervals to your text volume)

D-8w  demo text frozen; word count taken from the export, not from memory
D-7w  translation starts: demo text + short description + About section
D-4w  translated text back in the build; first full read-through in-game
D-3w  fix pass: overflow, missing glyphs, leftover untranslated strings
D-2w  store page updated per language; supported-languages checkboxes set
D-1w  broadcast assets, pre-written replies, social posts scheduled

The store page gates everything else

If you can only finish one thing in a given language, finish the store page. A player who cannot read the page never downloads the demo, so page translation sits upstream of every other number you are trying to move that week. It is also the smallest piece of text you own, which makes it the highest ratio of visibility to effort anywhere in this checklist.

Go through the page element by element rather than translating it as one blob. The short description does disproportionate work because it follows the game around browsing surfaces and hover cards. The About section is where a browsing player decides whether the game is for them. Tags and genre affect who sees you at all. And any text baked into your capsule or header art does not switch languages with the rest of the page — if your capsule contains a tagline in one language, every visitor sees it in that language regardless of their settings.

The supported-languages declaration deserves its own moment of honesty. Interface, subtitles, and audio are separate claims, and a demo that has a translated menu but untranslated dialogue is a partial state you should represent accurately rather than optimistically. Players check that list before they download, and they remember when it was wrong.

Broadcast, trailer, and everything with text burned into it

Next Fest gives developers a chance to broadcast to their own store page during the event, and a broadcast is by nature a single-language performance. That is fine — you cannot stream in six languages at once. What matters is that the page around the stream is not also single-language, because a viewer who cannot read the page has nowhere to go after the stream ends.

Video is the surface where language quietly hardcodes itself. Text overlays in a trailer, on-screen tutorial prompts in captured footage, UI visible in every screenshot: all of it is fixed at export time. Steam supports per-language store assets, and it is worth checking in Steamworks exactly which assets can be swapped per language before you decide whether to cut a second trailer or simply edit the first one to lean on visuals instead of words.

Screenshots are the cheapest fix in this category and the most commonly skipped. Capturing a parallel set with the game running in each of your main languages costs an afternoon and removes the single strongest signal that a game was not really made with that audience in mind.

• Trailer text overlays — either localize them or design the cut so it works with very few words
• Screenshots — capture a set per language; a visitor browsing in Simplified Chinese should not see only English UI
• Capsule and header art — burned-in taglines never translate, so consider keeping them text-free
• Broadcast overlays and the schedule graphic — at minimum, put times in a format that is unambiguous outside your own time zone

Running the week itself

Once the event opens, your job shifts from building to watching. Feedback arrives in languages you may not read, and it arrives fast. Machine translation is entirely adequate for reading that feedback — the risk in machine translation is on the output side, not the input side — so do not let a language barrier stop you from at least knowing what is being said.

Replies are the part worth preparing in advance. Writing a warm, correct message in a language you do not speak, at speed, while your demo is live, is where mistakes happen. Draft the handful of messages you know you will need before the week starts: thanks for playing, we have seen this bug and are fixing it, this language is not supported yet but we are considering it, here is where to report problems. Having those ready turns a stressful improvisation into a copy-paste.

Be conservative about pushing builds mid-event. A text-only fix for an untranslated string or an overflowing button is low risk and worth doing. A structural change to the demo during the busiest browsing days is not, because a build that breaks on day two costs you the rest of the week.

Finally, capture the feedback in a form you can sort later. Tagging each report with the language it came from turns a chaotic week of messages into an actual dataset: which languages produced the most reports, which produced the most text-fitting problems, and which produced enthusiastic players asking when the full game arrives.

Reading the data afterwards

The most useful comparison after the event is not the total, it is the shape. Look at where wishlist activity came from during the week versus where your page traffic came from. A region that sent you plenty of visitors but comparatively few wishlists is telling you something about how your page reads to those visitors — often that it does not read at all, because it is not in their language.

Be careful about confounds before you act on any of it. A single streamer with a large audience in one language can dominate a week and make that language look like a market when it was really one person's audience. A language you already supported will naturally outperform one you did not, which is a fact about your page rather than about the market. And absolute counts are always tangled up with how many Steam users a region has in the first place; the Steam hardware and software survey is a reasonable public baseline to sanity-check your own proportions against.

The output of this reading should be a short, concrete decision: which languages the full game will ship in, and which one you will add next to the demo. That last part matters because the demo does not disappear when the event ends. It keeps running, keeps being found, and keeps converting — and a language you add to it a month later is still working for you every week after that.