A game just blew up. How do I get it on every PC today?

A title breaks out on a Thursday. By Friday afternoon three of your regulars have asked whether you have it. You do not. It goes on the list. You get to it the following Wednesday and by then those three have already played it at the place two streets over and discovered that the place two streets over is fine.

That is the entire mechanic. It has nothing to do with your judgement about which games are worth the disk space.

Taste is not the moat

Every owner in your city is reading the same charts. The same streamers, the same clips, the same friend groups pushing the same title into the same Discord servers. Nobody has a secret. If you think your edge is knowing which game is about to blow up, you are competing on a piece of information that is public and free.

The gap that actually decides who gets the traffic is turnaround. Between a title breaking out and a customer sitting down in front of it, some number of days passes. In most cafes it is however long it takes the owner to get round to it, which is to say the following week. It does not need to be. It needs to be one.

Which games belong in your permanent lineup is a separate and slower question. I have written about that elsewhere. This post is about the sprint, not the library.

Three clocks are running from the moment you decide. The catalog entry. The seats. Telling people it is there. Owners are good at the first, unhappy about the second and skip the third almost every time.

Three things I am not going to pretend are solved

Before the clock, the honest part. Nothing below is a workaround. It is what the job actually costs.

Pushing a menu entry is not installing the game. These are two different jobs and only one of them is instant. The catalog entry, the artwork, the description and the launch tile go out to every connected PC in seconds. The binary does not. Forty gigabytes still has to land on forty machines and that is a deployment problem you own, whether you solve it with an image, a script, a sync tool or a remote session on each station. Doing that from the counter instead of walking to every seat is what remote management is for. The download is still a download.

Seats are the real ceiling. Disk space is cheap. Accounts are not. On a breakout title the whole room wants the same game at the same hour on the same night, which is the one scenario where a handful of shared logins runs out. Plan the account count before you plan anything else.

A delisted title comes back thin. If a game has been pulled from the Steam store, you can still bring it into the catalog. The storefront data it would normally carry is gone. Genre, rating and description arrive empty and you type them in yourself. It is not common. It is deeply irritating at eleven at night when you assumed the entry would fill itself.

The catalog entry

The Games page in HandyCafe is a browse surface, not a list you maintain by hand. It runs on Steam charts and Steam storefront data with artwork from SteamGridDB. It needs a Cloud subscription to work.

Three rails: Most played right now, Trending games and New releases. There is also a country selector with a Global option. The country selector is the part owners set once and never look at again. Use both settings. Your own country tells you what your customers want this weekend. Global tells you what has not reached your market yet.

Add a title from the catalog and it arrives with cover art, poster, genres, rating, release year and a description in the client language. Then it appears in the client menu on every connected PC immediately. No restart, no walking the floor, no telling the evening staff to re-scan anything.

That step is genuinely fast. It is also the step that gives owners a false sense of being finished, because a tile now exists and nothing else has been done.

The seats

Now the expensive clock.

If you are running shared launcher logins, put them in a pool rather than a document behind the till. A Game Account Pool holds Steam, Epic, Battle.net and GOG accounts and lends one out for the length of a session. The same account is never checked out to two machines at once, which on a breakout Friday is the only thing standing between you and a customer getting thrown out of a match by a stranger signing in on their credentials. There is a CSV import that takes username,password,label in bulk and reports how many landed, how many were duplicates and how many rows failed. Buying six accounts at once takes two minutes.

Two settings matter more on a new title than on a settled one, because new titles crash more. A suspension window from 0 to 1800 seconds holds a customer's account through a crash instead of releasing it to the next person in the room. A heartbeat timeout from 15 to 600 seconds pulls accounts back off a frozen PC that is never coming back. Get both wrong and your pool looks empty while half of it is stuck.

You can charge for the pool if you want to, per session, per minute or per account switch, with no charge as an option and a custom rate on any single account. On a breakout title I would not charge in the first week. You are buying attendance, not margin.

Here is the opinion. Buy fewer accounts than your gut says, then read the Events tab. It logs "denied no capacity" every time a request could not be filled. If that line shows up eleven times on Friday between eight and ten, you have a purchase decision backed by a number instead of a feeling. You make it on Monday. The account pool mechanics are worth understanding properly before a busy night rather than during one.

If what you own is a fixed number of seats rather than shared logins, Software Licenses is the other tool. It counts concurrent use of a title and blocks overuse without handing any credentials to anybody.

The step everyone skips

A game nobody sees earns exactly what a game you never added earns. Owners who got through the first two clocks in an afternoon leave this one for whenever.

The client menu is server-managed and it has structure you should use deliberately. Categories hold apps. Every app has a kind, a launch URI, icon and poster images and a sort order. You can link it to a store platform such as Steam or Epic with an app id and store URL. Every app carries a visibility toggle. You can build the entry quietly on Thursday and switch it on when the download finishes.

Two placements do the real work:

• The hero carousel. Featured apps rotate as banners across the top of the client desktop. This is the position with the highest chance of being seen by somebody who was not looking for the game. Give the new title a slot for one week.
• The quick launch bar. Pinned favourites at the top for one-click access. Put it there for the first weekend and take it off after. Otherwise your quick launch turns into a second menu and stops meaning anything.

Then push an announcement to the client desktop saying it is live. Announcements track impression count and click count, which means you get to find out whether your customers ignored it. That is uncomfortable and useful.

Every one of these changes goes to connected clients instantly. You can sit at the counter at seven on a Friday, promote the game to the hero slot and watch the room's desktops change.

And tell people who are not currently in the building. The desktop only reaches customers who already walked in. The same weekend logic that fills a new gaming cafe applies to a new title: the announcement in your group chat is doing more work than the tile.

A week later, check whether you were right

Open the Statistics page and the Applications tab. It ranks apps by cumulative usage time and shows a trend view across a date range.

This is the step that makes the next breakout cheaper. If the title has hours on it, you know your three-clock routine works and you repeat it. If it has almost none, you learn something specific rather than general. Either nobody saw it, in which case your placement was wrong. Or they saw it and did not want it, in which case the chart lied to you and your local market is not the global one. Those two failures look identical from behind the counter and completely different on the trend line.

What I would do

1. Thursday night, add the catalog entry with the app hidden. Start the download on every station.
2. Friday morning, load the accounts by CSV and set the suspension window and heartbeat.
3. Friday afternoon, switch visibility on, drop the title into the hero carousel and pin it to quick launch.
4. Friday evening, push the announcement and post the same thing wherever your regulars talk.
5. Monday, read the "denied no capacity" count and decide whether to buy more accounts.
6. The following week, open the Applications tab and find out if any of it mattered.

Six steps across five days. The cafe down the road starts the week after next.