Go-to-Market Playbook: Five Fields That Make a Play Runnable

Ask a revenue team for the go-to-market playbook and what usually arrives is a folder: a positioning deck, three case studies, an objection sheet, a pricing one-pager and a link to the CRM guide. Every document in it is real work. None of it tells a person what to do on Monday with the forty accounts in front of them, which is the thing a playbook is supposed to settle.

A playbook is not a library. It is a set of plays, and a play is a specific, repeatable motion with a trigger that starts it, criteria that decide who is in it, a message, an owner and a condition under which it ends. The document is the container. The play is the unit, and a playbook with no plays in it is an archive with an ambitious name.

What a play contains

Five fields, and a play is unusable when any one of them is missing.

1. Step 1Trigger

   The observable event that puts an account into the play. A funding round, a new executive in a named seat, a hiring pattern, a product signal. Observable from outside, and dated.

2. Step 2Entry criteria

   What else has to be true for the trigger to matter. Size band, geography, existing relationship status, and the disqualifiers that keep customers and live deals out.

3. Step 3The message

   The premise, written out. One sentence that is true of every account the trigger admits, in words the buyer would use about their own situation.

4. Step 4Owner and volume

   One name, and how many accounts a week the play can carry with current capacity. A play with three owners has none.

5. Step 5Exit and expiry

   What ends an account's participation, and the date the play itself gets reviewed. Triggers decay, so a play without an expiry is a play that quietly stops being true.

The trigger field is what separates a play from a segment. A segment is a group of companies. A play is a group of companies plus a reason to write today, and the reason is the part that decays. A hiring signal from four months ago is history, and a play built on it is contacting people about a situation that has already resolved.

Why the folder version fails

The asset library and the play library look similar from a distance because both are collections of documents. They fail differently, and the failure of the library is the one that goes unnoticed for longer.

The asset kitWhat usually arrives

• Organised by document type
• Grows on request, never retires
• Nothing in it has an owner or a date
• Cannot be run, only read
• No condition under which it is wrong
• Reviewed when someone reorganises the drive

The play libraryWhat a playbook has to be

• Organised by trigger
• Capped at the number a team can actually run
• Each play has one owner and a weekly volume
• Runs on a schedule
• Each play has a stopping figure
• Reviewed on its own expiry date

The kit is never wrong, which sounds like a strength. Nothing in it makes a claim specific enough to be contradicted by a quarter, so nothing in it ever gets removed, and the collection grows until search across it degrades and people rebuild what they need from memory. A play library has the opposite property: each play asserts that a particular trigger predicts a particular problem, and the quarter can settle that.

How many plays a team should run

Fewer than it wants to. The constraint is not creativity, it is that every play consumes the same scarce resources: list-building attention, sending capacity, and the reply handling that follows.

Three plays running at a volume where their results are readable beat nine plays running below that threshold. Below a certain volume a play does not produce a smaller version of its result, it produces an unreadable one, and the team then makes a judgement call and files it as evidence. The number of accounts a play needs before its reply counts mean anything is a function of your own rates, and the arithmetic for it is the same funnel arithmetic that sizes a campaign, covered with the counts shown in the ideal customer profile guide.

A practical ceiling for most teams is one play per owner, plus one experiment. When a fourth idea arrives, it goes in a queue behind an expiry date rather than into the rotation, and the honest version of adding a play is naming which existing play it replaces.

Where the plays come from

Three sources, in descending order of reliability.

Closed-won pattern. Take the last twenty accounts that bought and ask what was happening at each of them in the ninety days before the first conversation. Where a situation repeats, that is a trigger with evidence behind it. This source is the most reliable and the most often skipped, because it requires reading deals rather than inventing plays.

Closed-lost pattern, inverted. Losses tell you the entry criteria, which is the field most plays get wrong. An account that reached a meeting and died on a constraint that was visible from outside should have been excluded by the play rather than discovered by a person.

Observable market events. Funding, leadership changes, hiring, expansion, regulatory dates, competitor moves. These are the easiest triggers to detect and the easiest to over-read, because a detectable event is not the same as a predictive one. What various signal types genuinely predict, and which decay fastest, is set out in the B2B intent data guide.

The distinction underneath all three is between a filter and a signal, which is the highest-leverage idea in prospecting and is covered directly in B2B prospecting.

Writing the play so somebody else can run it

The test is handover. Give the play to a person who was not in the room and ask them to produce this week's list and the first message. If they come back with questions, the play is a description rather than an instruction, and every question they ask names a field that is not finished.

Two failure modes account for most of the questions. The trigger is stated in a form nobody can check, so membership becomes a judgement call that drifts week by week. And the message is stated as a theme rather than a sentence, so each person writing from it produces a different premise and the play stops being one play.

Is this a play or a description

• ~Depends: Someone outside the team can produce this week's list from the trigger and criteria
• ~Depends: The trigger is observable from outside the account and carries a date
• ~Depends: The disqualifiers name customers, live deals and anyone another play is working
• ~Depends: The premise exists as a written sentence, not a theme
• ~Depends: One name owns it and a weekly volume is stated
• ~Depends: The play has an expiry date and a stopping figure

What a quarter of plays looks like

Plays interact through shared capacity, which is the part a play-by-play view hides. Three things have to be true across the set rather than within any one of them.

No account sits in two plays at once. Two plays running through overlapping lists is how one person receives two different pitches in a week, and it is invisible from inside either play. The suppression rule is that an account belongs to exactly one live play, with the tie broken by whichever trigger is more recent.

The plays do not all peak together. Triggers arrive at different rates. A funding-round play produces a trickle and a hiring-signal play can produce a flood the week a segment starts scaling, and if both are owned by the same person at the same volume, the flood week is the week the trickle stops being worked.

One slot is reserved for the experiment. A rotation with no room for a new play produces a team that stops generating them, and the closed-won pattern that supplies the good triggers keeps moving. The experiment slot exists to be wrong cheaply, which means it carries the same five fields and a shorter expiry.

The failure that these three prevent is the common one where a playbook is technically current, every play is defensible on its own, and the quarter still produces an unreadable result because the plays were competing for the same capacity and the same accounts.

The retirement rule

The habit that keeps a playbook usable is retirement, and it needs to be mechanical rather than discretionary. Every play carries a review date. On that date the play either produces its numbers and continues, or it stops. A play that is kept because removing it would feel like giving up is a play that will still be there next year, consuming capacity that the closed-won pattern has since pointed somewhere else.

Retirement also protects the thing a playbook is for, which is that the same motion runs the same way twice so its result means something. A library that only ever grows loses that property, because the plays at the bottom of it are being run occasionally by whoever remembers them, in slightly different forms, producing results nobody can compare.

Where we differ from standard practice

The single-channel version of all of this, written out as an operating model rather than as a framework, is the outbound sales playbook. One thing in it is ours rather than the market's and it changes how plays are constructed.

We run one message per campaign. No bumps, no thread replies, no scheduled second attempt at somebody who did not answer. Re-approaching an account that stayed quiet happens as a new campaign on a new premise, normally because something changed at that account, rather than as a reminder underneath the original message. In play terms, that means a play ends at the send, and the next play is a genuinely different trigger rather than position two of the same one. The test we apply is whether the new message would stand up if the reader had never received the first.

The cost of that policy is fewer touches per account, which pushes the work into the trigger and the premise. Where the sourcing is what is missing rather than the plan, we run that half, with the meeting criteria agreed in writing before anything sends. The full argument for the policy, including what it costs us, is in why we stopped using follow-ups.

The short version

A go-to-market playbook is a set of plays, and a play is a trigger, entry criteria, a written premise, one owner with a weekly volume, and an exit with an expiry date. A collection of documents organised by type is an asset kit, which is a useful thing that answers a different question.

Run fewer plays than you want to, because a play below its readable volume produces a judgement call rather than a result. Take triggers from your closed-won pattern first, entry criteria from your losses, and treat detectable market events as the easiest source to over-read.

Test every play by handover before it enters the rotation, and retire on a date rather than on a feeling. The decisions the plays inherit, including who is in the target set at all, belong on the one page that a go-to-market strategy produces.