River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Product Launch Plan Template

Seven documents and four sheets that tie every messaging claim to the requirement it rests on, then order the calendar by what you can take back.

Free download  ·  No account needed

Claim to Requirement Map

Eleven claims, nine requirements, one scope cut

Ashwell Systems  ·  Rental Fleet 4.0  ·  status re-pulled at T-4, after the week 10 cut

ClaimReqStatusVerdictSurfaces
Inspect a unit with no signal at allR1CompleteClear3
Mark the damage on the photograph itselfR2CompleteClear3
The renter signs on the deviceR3CompleteClear3
Damage costs priced before the renter drives offR4Cut at T-11Blocked4
An end to disputed damage invoicesR4 R9Cut, partialBlocked3
The report reaches the renter before they leaveR6Email onlyNarrow3
Conflicts resolve themselves on reconnectR5CompleteClear2
Every inspection in the fleet in one viewR7CompleteClear3
Scan the unit and the record loadsR8CompleteClear2
Charges post straight to the ledgerR9Read onlyBlocked2
Cuts inspection time in halfnoneNo requirementUnsourced4
11 claims across 7 surfaces6 complete, 2 partial, 1 cut6 clear, 1 narrow, 3 blocked, 1 unsourced32

Four bad claims are thirteen sentences

The three blocked claims and the unsourced one appear thirteen times across six of the seven surfaces. Counting claims makes this a morning. Counting instances makes it a two-day job with six owners, which is the truth.

The gap that produced all of it

Messaging last edited on the Wednesday of week 8. Scope cut on the Monday of week 10. Fifteen days in which finished-looking copy was wrong and no review cycle would have caught it, because the review happened in week 9.

One surface was clean: the in-app notice, four claims, all clear. It was written in week 10 by an engineer looking at the running build.

Ashwell Systems cut a requirement from Rental Fleet 4.0 eleven days before ship. The messaging document had last been edited fifteen days before that, and nobody read it again, because re-reading finished copy is not on anybody's list. A launch checklist does not catch this. The checklist ran, every box got ticked, and four claims across six surfaces were still asserting a capability that had been removed from the release.

So this pack traces rather than reviews. Every claim carries the identifier of the requirement it rests on, and a requirement that changes status turns the claims standing on it into a list of files and slide numbers. Eleven claims, seven surfaces, thirty-two instances: six clear, one narrow, three blocked, one unsourced. Thirteen sentences had to be cut or replaced, not four, and thirteen is the number that decides whether the fix is an afternoon or a fortnight.

Then the calendar gets ordered by what can be taken back. The launch email sat at nine on day one, when a seven day phased release had reached one percent of the iOS base, 184 devices, against 12,600 app users on the list. Apple notes an update in phased release can still be downloaded manually by anyone, so the email would have made the phase decorative. It moved to day five. Installs into the marketing workspace beside the rest of the template library.

Readiness counted in open blockers, a calendar ordered by reversibility, and the note the reps actually read

The Readiness Tracker, the Launch Calendar with reversibility classes, and the Enablement Note.

Readiness Tracker

Thirty-four items. Every open one names what breaks and who notices first.

ItemDescriptionBlockerWhat breaks if it stays openWho notices first
B12Support macro for the question “where is the cost estimate”YesSupport tells customers pricing is coming and names no releaseSupport, within one shift of the email
B27Ledger write-back removed from the sales deckYesA rep demos a capability that does not existThe prospect, live on the call
B32Webinar abstract still carries the unsourced time savingNoAn unmeasured claim publishes; the surface is not live yetRegistrants, from day 6
B33Blog post header imageNoThe post looks unfinishedReaders
B30Named owner recorded for every Play rollout increaseClosedThe rollout sits at 10 percent indefinitelyNobody: it simply never ramps
B23Launch email send moved to day 5 and the schedule lockedClosedThe phased release is defeated by manual downloadsMobile lead, watching the ramp
At the go/no-goCountReads as
Items closed30 of 3488 percent complete
Open blockers2the number that set the date
Open non-blockers2tasks, not blockers

Eighty-eight percent was the number quoted in the status meeting. Two was the number that decided when the irreversible rows could be dated. Both blockers closed on day 3, which is why the email went on day 5.

An item is a blocker only if it can answer both questions. If the honest answer to “what breaks” is that something looks untidy, it is a task. A list where everything is a blocker is a list where nothing is, and the go/no-go becomes a vote.

Who-notices-first is the field that closes items. Two of the thirty-four answered “nobody, ever” and were deleted rather than done.

Launch Calendar

Nine activity classes ordered by recoverability, not by readiness. 18,400 iOS installs, 22,700 Android.

DayActivityOwnerReversibilityHow it is taken back
1In-app notice on, both platformsProductMinutesFlag off
1Pricing page paragraph publishedWebMinutesPublish an edit
1Phased release opens at 1 percent  184 devicesMobile leadPausablePause up to 30 days; manual downloads continue
1Play staged rollout at 10 percent  2,270 devicesMobile leadHaltableHalt; users who have the build keep it
2Blog post liveContentMinutesEdit or unpublish; URL may persist in search
2Sales deck reissued as 4.0-c, announced out loudEnablementFile onlyNew file; laptop copies cannot be recalled
3Play rollout raised to 50 percent  11,350 devicesMobile leadHaltableHalt the release
3App Store listing text and screenshots submittedMobile leadAfter reviewResubmit metadata and wait
5Launch email  41,300 contacts, 12,600 app usersLifecycleIrreversibleA correction, never an undo
6Wire press releaseCommunicationsIrreversibleSyndicated within the hour, stays syndicated

Three things moved during the sequencing pass. The email went from day 1 to day 5, where the phase reaches 20 percent, or 3,680 iOS devices, rather than 184. The wire release went behind the in-app notice, so the first people to hear about a feature can go and look at it. The store metadata went from day 1 to day 3, because a review queue is not a schedule you own.

The ramp cell needs a name in it. A staged rollout percentage does not rise on its own, so day 3 at 50 percent was a hope until somebody's name went in the cell on day 2.

An irreversible row that depends on an open blocker stays undated. That is the correct output, and it beats a date everybody privately expects to slip.

Enablement Note

For 14 reps, 3 solution consultants and the support team. Read once, quickly, before a call.

Say this

The inspection report is emailed to the renter the moment the inspection closes. Say emailed. Do not say delivered, texted, or reaches them: the message path in the requirement did not ship.

What it does not do, and what to say instead

Asked aboutThe plain answerThen say
Pricing the damageNot in 4.0. The rate card estimate was cut.Damage is captured, marked and timestamped at handover, so the charge is built from evidence rather than memory
Posting to the ledgerRead only. Billing still actions the charge.The inspection is matched to the open rental agreement, ready for billing
How much faster it isNo published figure exists.The instrumentation went in with 4.0 and the comparison is being run across 50 inspections

Never name a future release. “Coming soon” from a rep becomes a commitment in the prospect's notes, and the next conversation starts from it.

The thing most likely to get someone into trouble

Three sentences were live on the deck and the pricing page as recently as day 1 and are now false: damage priced at handover, an end to disputed invoices, and charges posting straight to the ledger. All three read well. If a slide in front of you says any of them, you have an old file.

Deck 4.0-a and 4.0-b are withdrawn. The reissue was announced out loud in the sales channel, because a file drop does not reach a laptop.

What's in the pack

01

Claim to Requirement Map

One row per claim, not per document, with every surface it appears on and one of four verdicts against it.

02

Launch Calendar

Owner, reversibility class and how each activity is taken back, with the irreversible rows dated last.

03

Readiness Tracker

Thirty-four items where every open one names what breaks and who notices first, in that order.

04

Asset Checklist

Per surface: audience, size, claims carried, edits needed and reversibility, so the fix list has owners.

05

Messaging

Each claim with its requirement and verdict, keeping every struck-through original visible above its replacement.

06

Comms Plan

The sequence with device counts at each rollout step, and one sentence on why every moved activity moved. A price or package change riding along gets its own ceiling and comms plan rather than a line item here.

07

Enablement Note

The two questions a cut requirement generates, the words to use, and the sentences that are now false. Whether the shipped capability is actually adopted once it reaches a new account is a separate content problem this pack does not try to solve.

08

Reversibility Sequence

Five recoverability classes, and the staged rollout mechanics that decide where the announcement goes.

How to use it

  1. 1

    Open in River, or take it blank

    Open the pack in River and send your requirements and your copy, or download the Word documents and CSV sheets and build the map yourself.

  2. 2

    Send the requirements with their status

    Status is the field people leave off and the field the whole space runs on. Complete, partial or cut, with a date against each one.

  3. 3

    Compare two dates first

    When the messaging was last edited, and when the scope last changed. If the second is later, there is copy nobody has read since it stopped being true.

  4. 4

    Count instances, then sequence

    Verdicts multiplied by surfaces gives the real size of the fix. Then put the things you cannot take back behind the things you can.

Frequently asked questions

Is this template free?

Yes, with no account, no card and no email gate on the download. Edit with AI is the optional half: River reads the requirements and the copy, builds the map, and routes the blocked claims to surface owners with replacements drafted. Every pack sits in the template library.

What format are the downloaded files?

Seven Word documents and four CSV sheets, in a zip. The documents open in Word, Pages and Google Docs. The sheets open in Excel, Numbers and Sheets. Nothing needs converting, and the sheets are real spreadsheets rather than tables flattened into a document.

What if there is no PRD, only a ticket list?

A ticket list works, as long as each item has a current status. What the map needs is the identifier, one sentence of what it does, the status and the date that status was last true. If you want a real one, start from a PRD.

Is this different from an announcement kit?

Yes. The feature announcement kit writes four surfaces that agree with each other for one small change. This coordinates a whole launch: nine requirements, seven surfaces, six teams and a go/no-go, and it starts from what product wrote rather than from a brief. The blocked claims it finds also sit on the competitive battlecards, which is where a false claim survives longest.

Why order the calendar by reversibility?

Because the reversible activities are the ones that surface a problem, and running them first buys the chance to fix it. An in-app notice comes back with a flag. An email to 41,300 contacts comes back as a correction, which costs more credibility than the announcement bought.

Does it replace the go/no-go meeting?

It changes what the meeting looks at. Percent complete gets quoted and open blockers decide, so every open item carries what breaks and who notices first. An item that cannot answer both is a task, and moving it out is what makes the count mean something.

What does Edit with AI actually do?

It signs you up, installs this pack as a private workspace, and starts with the claim map. Send the requirements and the copy, and the finding comes back as a count: how many claims no longer have a requirement behind them, and how many surfaces they sit on.

Find the claims that lost their requirement

Take the Word documents and CSV sheets blank, or open this exact pack in River and send it your requirements and your launch copy.

Edit with AI