Stop Archiving Steam Achievements Without a Player-Safe Plan - 2026

Stop archiving a live Steam achievement the night you discover sets. October 15, 2026 put Achievement Sets in the Steam client beta, and the Steamworks achievements guide now includes archive as the alternative to delete. Archive is a real tool. A same-night click on a shipping holiday row is not a plan.

The refusal is BETA_NIGHT_ARCHIVE. If the only reason you are in the achievements editor is the October 15 client announcement, you do not have a player-safe archive yet. Close the draft. The sets what-changed map already explains API names, the base set, and DLC restrictions. This page owns the archive decision.

Direct answer: Keep seasonal achievements earning until a written plan names the API, who still sees it, the build that stops awarding it, and the publish time. Then archive. Until that page exists, the decision is HOLD.

Why this matters now

Valve described seasonal and holiday sets as a reason the feature exists. The Verge quoted the line about temporary achievements for season-based games and time-limited holidays. That sentence will be read, the same week, as permission to retire last December’s event before Friday’s build.

The guide is more careful than that reading. Archive keeps an earned achievement visible to the player who unlocked it, and it can still be featured on a profile. Players who never unlocked it stop seeing it, and it stops counting toward their completion. An archived set is harsher. Nobody can unlock anything in it. Only players who already unlocked something in that set still see it, and they still see every achievement in the set, including the ones they missed.

Those are two different player outcomes. The editor lets you mark either an achievement or a whole set as archived, then save a draft, then publish from the same Publish tab you use for every other Steamworks change. The sets feature is still labeled beta. The new achievements page is in the client beta. None of that slows the Publish button down.

A studio that archives on beta night ships a live schema change to a client channel that not every player is on, from a build that still fires the old unlock, with store copy that still lists the holiday row. That is the failure this page is for.

Who this is for

Time: 25 minutes to choose KEEP_EARNING, PLAN_ARCHIVE, or HOLD. Writing the plan itself is a separate sitting. Do not finish both in the same hour you first opened the new editor.

Prerequisites: the app’s achievement list and a partner login. You do not need a new Steamworks SDK. Sets did not change SetAchievement. Archive does not require one either, which is why it is easy to publish too early.

The only three decisions

BETA_NIGHT_ARCHIVE is the click this table exists to block. It looks like PLAN_ARCHIVE and it has the evidence of KEEP_EARNING. The holiday achievement still has a global unlock rate. The game still grants it. The patch notes are not written. Someone saw “archive instead of delete” in the docs and treated the recommendation as a deadline.

The guide recommends archive instead of delete when players have already earned the achievement. That recommendation protects earners from a wipe. It is not a recommendation to archive tonight. Delete removes the achievement for everyone, including people who unlocked it. Archive is the safer retire after you have decided to retire. It is not the safer way to experiment.

One achievement is not one set

Read this before you pick a menu item. The achievements guide separates the two operations.

Archiving one achievement

• Players who unlocked it still see it, and they can still feature it on a profile.
• Players who never unlocked it do not see it.
• It does not count against completion for the players who never unlocked it.
• The game is unchanged. You have to stop awarding that API name yourself.
• The store-page achievements panel omits archived achievements, the same way it omits spoiler achievements.

Archiving a whole set

• No achievement in that set can be unlocked.
• The set stays visible only to players who previously unlocked at least one achievement in it.
• Those players can still see all of the set’s achievements, not only the one they earned.
• Players who never unlocked anything in the set stop seeing the set.

If your holiday event is five achievements inside the base set, archiving one of them hides that one row from people who missed it. The other four stay. If you instead created a “Winter” set and you archive the set, a player who earned the first night’s achievement still sees the four they did not earn, and they can never earn them. A player who earned nothing loses the set from the page entirely.

That second outcome is the one teams miss when they say “we’ll just archive the season.” You did not hide the season from the people who participated. You froze it in a half-finished state for them, and you erased it for everyone else. Write that sentence into the plan before you choose Edit set → Archived.

A set with no achievements does not appear. An archived set with achievements appears only to prior earners. Those are not the same lever as deleting the set. Deleting a set requires you to move or delete every achievement in it first. Archive is the path that leaves earned rows in place. Use it when that is the outcome you want, not when you want the rows gone.

What “player-safe” has to contain

A player-safe plan is a page in the repo or a ticket, not a memory. PLAN_ARCHIVE is blocked until every line below is filled. If one line is unknown, the decision falls back to HOLD.

1. API name exactly as the build passes it. Not the English display name.
2. Scope: one achievement, or the whole set. Circle one. Do not write “the holiday stuff.”
3. Set id and set name the row lives in today. Most shipping games still have everything in the base game set, which cannot be renamed or deleted.
4. Global unlock rate from the achievements table. Zero means almost nobody earned it, and delete may still be the honest tool if you are sure the row never shipped to players. Above zero means earners exist, and delete is the wipe the editor is warning you about.
5. Who still sees it after publish. For one archived achievement, earners. For an archived set, anyone who unlocked any row in that set, and they see every row.
6. Who stops seeing it. Non-earners of that achievement, or players with zero unlocks in that set.
7. Completion side effect. List the sets whose “still available” list shrinks. A player completes a set by unlocking every achievement that is still available. Archived achievements they have not unlocked are not required. If this archive is the last locked row for a large group of players, you are about to mark that set complete.
8. Showcase side effect. A game qualifies for the Completionist Showcase when the player has completed any of its sets. If step 7 newly completes a set, say so in the plan. Do not discover it from a player screenshot.
9. Build id that stops awarding the API name, already on the default branch players run. Archiving does not patch the game.
10. Publish owner and time. One person. One window. After the build in step 9, not before.
11. Copy list: store panel, About text, FAQ, and the fest or seasonal trailer that still names the achievement. The store panel will drop the archived row on its own. Your sentences will not.

Missing step 9 is the usual hole. The editor feels like it retired the achievement because the draft preview changed. Players on the current depot still run the old grant. For a single archived achievement, the guide puts the stop in your code. It states a client refusal for an achievement in an archived set, and for a DLC set the player cannot play. Do not treat those client refusals as a guarantee that a lone archived row inside the base set will no-op SetAchievement. Ship the build that stops calling it.

Order that will not page you:

1. Land the build that no longer awards the API name. Confirm it on a default-branch playtest.
2. Update the store and FAQ lines that name the achievement.
3. Publish the Steamworks draft that sets archived.

Order that will page you:

1. Publish archived.
2. Leave the awarding build live.
3. Watch a fanfare for an achievement the panel no longer shows, or a refusal the UI did not expect, depending on whether you archived a row or a set.

The Unity unlock help page is for the opposite bug, a grant that never arrives. Read it anyway if your “stop awarding” change is a comment around SetAchievement that still calls StoreStats on a stale flag. A half-removed grant is still a grant.

Completion is a latch, not a preview

This is the reason HOLD exists during beta week.

A player who has finished every still-available achievement in a set has completed that set. Once they have completed it, the guide says it stays completed even if you add achievements to that set later. Archived achievements they never unlocked are simply not required.

So archive can complete a set for people who were one holiday row away. Adding the holiday row back later, or un-archiving it, is not documented as a reversal of that latch. The documented latch is about adding achievements after completion. I will not invent a promise that un-archive walks the badge backward. Plan as if the completed state sticks.

That makes beta-night archive a one-way experiment on a live population. You can un-check Archived and publish again. You should not tell earners, or your producer, that every completion badge returns to yesterday’s math. If the seasonal row was the last gap for thousands of players, you manufactured completionists. The Completionist Showcase then has a factual basis under the current rules, and a trust problem in the comments.

Put a number in the plan before you accept that. You do not need a perfect forecast. You need the global unlock rate on the rows you would archive and the unlock rate on the set they sit in. If 40 percent of players have every base achievement except the holiday one, archiving the holiday one completes the base set for that 40 percent. Write “accept this” or do not publish.

New sets do not fix a mistaken archive after the fact. The guide’s advice to add a new set for new content is about keeping a completed badge from sitting on top of fresh locked rows. It does not rewind a set you already marked complete by hiding work.

Delete is not the cautious button

Teams reach for delete when archive sounds permanent. Delete is the wipe.

The guide’s FAQ is short. Archived achievements stay visible to players who unlocked them and no longer count against completion for players who did not. Deleted achievements are removed for everyone. If the achievements table shows a global unlock rate above zero, the editor recommends archive rather than delete.

Use delete only when you are sure the row never reached players. A test achievement in a developer-only set that only your accounts unlocked is a different conversation from a holiday achievement with a public unlock rate. Developer-only sets are hidden from players and only developers can unlock them. Archiving or deleting inside a set that players never saw is not BETA_NIGHT_ARCHIVE. Publishing a change to a set players already see is.

Do not “clean up” by deleting the earned row and recreating it under a new API name. Earners lose the old one. The new one starts locked. You have taken a profile trophy away and asked them to grind it again. That is worse than a late archive.

A winter set, worked as a decision

Five achievements, API names WINTER_1 through WINTER_5, display names you can ignore for the plan. They shipped in December 2025 inside the base set. Global unlock rates, the kind the editor table shows, might look like this. The rates here are an illustration of how to read a plan, not a measurement from a shipped game.

BETA_NIGHT_ARCHIVE on this table: someone archives all five because the sets beta mentioned holidays. The live build still grants WINTER_1 on login. Non-earners lose the rows and may complete the base set if those were their last gaps. Earners keep a private trophy the store panel no longer advertises. The FAQ still says the winter event is live. That is four contradictions. None of them required a new SDK.

KEEP_EARNING: the 2026 winter event is still scheduled. Leave the five API names granting. If you want a cleaner page, that is a new set for next winter’s new achievements, created when you are ready to publish a set with a name and a restriction, not an archive of the old rows. The what-changed page covers creating that set. Do it on a draft you do not publish the same night.

PLAN_ARCHIVE, honest version: the winter event is not coming back. The plan says scope is five individual achievements, not a set archive, because they live in the base set and you do not want a leftover set. The build on default removes the December login grant and any UI that still pops WINTER_1. FAQ and the store blurb drop the event. Publish archive the next morning, with the build id written on the ticket. You accept that non-earners no longer need these five for base completion, and you wrote the expected completion jump next to the 22 percent figure so it is not a surprise.

Wrong scope: you move the five into a new Winter set on Monday and archive that set on Monday. Players who earned only WINTER_1 still see WINTER_2 through WINTER_5 locked, forever, because an archived set shows all of its achievements to anyone who earned one. You also moved earned achievements between sets. The editor warns when a move can hide an earned achievement or break completion. A same-day move plus archive stacks both warnings. HOLD.

The illustration is the point. Your table needs your unlock rates and your build id. A plan that cannot fill the rate column is not ready.

What to say in public

Creators will want a devlog line the day the client beta hits their friends list. Use a line that matches the decision you actually made.

Do not post “we deleted the holiday achievements” if you archived them. Earners still have them. Do not post “everyone keeps the same completion percentage.” Non-earners may gain a completed set because those rows no longer count. Do not post “update your SDK.” Nothing in this retire path is an SDK bump.

If a trailer or a fest FAQ still counts those icons, fix the count in the same window as the publish. The fest hour-one parity checklist is about demo claims versus the overlay. A retired seasonal row creates the same class of lie on the full game. The store panel will hide archived and spoiler achievements for you. The sentences you wrote will not.

Governance for a partner read

Archive is a configuration publish. It is not a depot, and it still changes what players see on the achievements page and what counts as finishing a set. Treat the Publish tab like a price change.

Privacy does not move. You are not exporting a new list of earners by archiving. You are changing visibility and completion. If someone asks for “the list of players who earned WINTER_1” as part of the plan, that list is a Steamworks stats question with its own access rules. The global unlock rate on the achievements table is the figure this plan requires. Do not scrape profiles to decorate a ticket.

Cost is an evening of engineering if the grant is a single call site, or a week if the seasonal popup is copied in three modes. The Publish click is the cheap part. That is why it happens too early. Schedule the click after the expensive part, not before.

The Steamworks resource list is the wider partner-doc shelf. It will not choose KEEP_EARNING for you.

Common ways the beta night goes wrong

Archiving because the docs said not to delete. The docs are protecting earners from a wipe. They still expect you to stop the grant in the game and to mean the retire.

Archiving the set when you meant one row. Participants keep a graveyard of locked holiday achievements they can no longer earn. Non-participants lose the set. Those are both support tickets, and they contradict each other.

Publishing the draft because the preview looked tidy. Preview is your account. Publish is the population. The visibility column even has a state for “will be visible upon publish.” Archive has the same draft-until-publish gap. Saving is not shipping. Publishing is.

Using archive to dodge the 100-achievement cap. Archived rows still count. An unvetted app does not get slots back.

Moving earned rows into a developer-only set to “hide” them. The editor warns that a move into a hidden set hides them from players who earned them. That is a wipe wearing a different button. HOLD.

Assuming un-archive is a full undo. Clearing the flag is a second publish. The completion latch for players who already completed the set is documented as sticky when you add achievements later. Do not sell a round trip you cannot cite.

Writing the devlog before the build. Players read “retired” and still get the popup, or they get a client refusal from an archived set while the UI plays the success path. Match the post to the build id.

Cleaning a fest demo and the full game in one publish. The demo and the full app are different achievement lists if they are different app IDs. A plan that names one API name on the wrong app archives nothing useful and can retire the wrong product. The plan’s first line is the app id.

Monday, if you already clicked

If the draft is saved and not published, discard or leave it unpublished. HOLD is still available. An unpublished archive changes nothing for players.

If you already published:

1. Confirm whether you archived one achievement or a set. The player-visible difference above is the one you now have to explain.
2. Check whether the live build still awards the API name. If it does, ship the stop before you write a longer post.
3. Read the store panel and the FAQ. Remove lines that still offer the retired achievement as earnable.
4. Do not delete the row to “finish the job.” Delete takes it from earners. You are already in the archive outcome. Stay there unless the plan says otherwise.
5. Tell support the two populations: earners who keep it, non-earners who do not see it. If you archived a set, add the third fact. Anyone who earned one row still sees the whole set, locked rows included, and cannot unlock the rest.

That is recovery. It is not permission to archive the next season the same way.

What this page will not decide

It will not fill in a first-time sets lab. That form is the achievement sets worksheet. It will not list every free tool around the editor. That list is the twelve free set tools. It will not tell you to freeze a Next Fest demo’s icon count. That checklist already exists, and it uses “set” in the ordinary sense of “the list you promised,” not as a Steamworks set object.

It will not tell you that archive is bad. Archive is how you retire a mode without taking a trophy off the profiles that earned it. The opinion is narrower. Stop doing it without the plan, and stop doing it the night the client beta shows up.

Key takeaways

• Refuse BETA_NIGHT_ARCHIVE. October 15, 2026 made archive easy to find. It did not make a seasonal retire automatic.
• Choose KEEP_EARNING, PLAN_ARCHIVE, or HOLD. A Slack “let’s archive the holidays” is HOLD.
• Archive of one achievement keeps it for earners, hides it from everyone else, and drops it from their completion math. Your game must stop awarding it.
• Archive of a whole set blocks every unlock in that set. Only prior earners still see the set, and they see every achievement in it.
• The guide recommends archive instead of delete when people have already earned the row. Delete removes it for everyone. That advice is not a deadline.
• A completed set stays completed even if you add achievements later. Do not treat un-archive as a documented undo of a completion wave you just caused.
• Completing any set can qualify the game for the Completionist Showcase. Archiving the last gap can create that qualification.
• Archived rows still count toward the achievement cap. You do not free the 100-achievement unvetted limit by archiving.
• Publish only after the build that stops the grant is the default players run. Archiving does not change the game.
• Say “client beta” when you talk about the new achievements page. Do not claim every stable client already shows the new bars.
• One named owner presses Publish. The plan’s first line is the app id and the exact API name.

FAQ

Should I archive Steam achievements the day achievement sets launched?

No. The sets beta on October 15, 2026 is a reason to read the rules, not a reason to retire a live row. Archive after a player-safe plan, or leave the achievement earning.

Do players who earned an archived achievement keep it?

Yes. The Steamworks guide says an archived achievement stays visible to players who unlocked it, and they can still feature it on a profile. Players who never unlocked it do not see it, and it does not count toward their set completion.

What if I archive the whole set instead of one achievement?

No achievement in an archived set can be unlocked. The set remains visible only to players who already unlocked something in it, and those players still see all of the set’s achievements. Players with no unlock in that set stop seeing it.

Does Steam stop my game from granting an archived achievement?

Archiving does not change your game. You must stop awarding that API name in the build. The guide says Steam will not unlock an achievement in an archived set, or in a DLC set the player cannot play. Ship the build that stops the call either way, and do it before you publish the archive.

Will archiving change who has completed the game?

It can. A set is complete when the player has unlocked every achievement that is still available. Archived achievements they never unlocked are not required. If you archive someone’s last locked row, that set can flip to complete, and a completed set is what qualifies a game for the Completionist Showcase.

Can I undo an archive if the beta night was a mistake?

You can edit the achievement or set and publish again with archived cleared. Do not promise that every completion badge returns to the previous math. The guide says a completed set stays completed even if you add achievements later.

Is delete safer than archive while sets are in beta?

No. Delete removes the achievement for everyone, including players who earned it. The editor recommends archive instead of delete in that case. The safe action during beta week, without a plan, is to leave the row alone.

Do I need a Steamworks SDK update to archive an achievement?

No. Archive is a Steamworks configuration change. The same API names and unlock calls remain. The risk is the Publish tab, not a missing SDK.

What to do next

Open the achievements table and write down API names, set names, and global unlock rates for any seasonal row someone suggested retiring. Do not mark those rows archived in that sitting.

If the event is still on the calendar, the decision is KEEP_EARNING. If the mode is gone and you can name the build that already stopped the grant, you may write PLAN_ARCHIVE and schedule the publish. If you are here because the client beta is new, the decision is HOLD.

Share the three-row decision table with anyone who has Publish rights. The button is older than the sets beta. The player outcome is not.