The question

We know where we’re stuck. What do we actually do about it, and how will we know it worked?

Naming the problem is the easy half. The hard half is doing a few specific things in order, then being honest about what changed afterwards, rather than declaring victory because everyone was busy.

Where it sits on the loop

A fix starts at your first constraint. There are five common ones, each tied to one step on the loop: leads not turning into conversations (Imported to Working and Engaged), replies not becoming deals (Replied to Pipeline), deals sitting still in Pipeline, deals not closing (Pipeline to Won), and won customers with no proof of value (Won to Evidenced).

It belongs in the weekly review: one fix running, its next step and who owns it, and the date you’ll look at the result.

How to think about it

A fix is a short, ordered plan with its measurement decided before you start.

Diagnose before you push. The first step is usually looking: which imports were poor fit, which deals lost and to whom. Working harder on the wrong leads makes the number worse.
Change the cause, then act. Revise the ICP, refresh the battlecard, agree what a reply needs before it becomes a deal. Then put the change to work.
Write the baseline down. The rate you’re trying to move, and the date you read it. Without it, there’s nothing to compare against later.
Choose the window in advance. Long enough for the change to show: a couple of weeks for lead conversion, six weeks or more for deals to close. Don’t peek and call it early.
Say what changed after, not what you caused. Other things move at the same time: new hires, the season, a big deal. “The rate went from 18% to 26% after the fix” is honest. “The fix lifted it 8 points” usually isn’t earned.

Common mistakes

Measuring before the work is done. Half the steps happen, someone checks the number, and the fix gets judged on work that never landed.
Running the same fix twice at once. Two people start it from different meetings and nobody knows which steps count.
Crediting the fix with everything. The rate moved, so the fix worked. Next quarter the same fix “fails” and nobody trusts the method any more.

In Summit53

A Fix playbook is a named set of steps for one constraint. A leader starts a run, the team works through the steps, and once every step is done Summit53 watches the matching conversion rate for a set number of days and shows what changed after. Each run is recorded in the Loop, so it counts in Did our advice work?

Where to find them

Fix playbooks open in a panel on the Revenue Engine. You reach one from:

Fix this: <playbook> in the Stage Inspector’s first-constraint badge, and on the first-constraint card in Diagnostics’ Action Brief. See Diagnostics.
Fix this: <playbook> on a firing Blueprint check. See Blueprint & Knowledge.
Run the Stalled pipeline fix in the Field guides on the Flow and Risk Signals pages.
A Fix playbook row in the Loop’s Intervention ledger, which opens that run.
A link: /revenue-engine?fix=<playbook> opens the playbook’s open run, or a preview if none is open; /revenue-engine?fix_run=<id> opens one run, including a finished one.

The five playbooks

Leaky top of funnel: Imported leads aren’t turning into conversations. Check who you’re importing before you work harder. Suggested when the constraint is Import → Working or Working → Engaged, or when recent imports are poor fit or leads were scored on an older ICP. Steps: Check the fit mix of recent imports; Revise the ICP; Rescore leads scored on the old ICP (can be skipped); Put strong- and good-fit leads into a sequence. Measured on Import → Working for 14 days.
Replies not becoming pipeline: Leads reply but few become deals.Suggested when the constraint is Reply → Pipeline, or a lead source gets replies but little qualified pipeline. Steps: Review replied leads that haven’t become deals; Agree what a reply needs before it becomes a deal: pain, budget owner, timeline; Shift prospecting toward the sources that convert. Measured on Reply → Pipeline for 30 days.
Stalled pipeline: Deals are sitting still. Suggested when deals have gone quiet, slipped their close dates or stopped moving, or when the constraint is Pipeline and nothing else fits. Steps: Open Deal Drag and find the stalled deals; Re-engage the deals that have gone dark; Close out or re-date overdue deals. Measured on the win rate for 45 days.
Win rate slipping: Fewer late-stage deals are closing. Suggested when the constraint is the win rate, or when a competitor on open deals has no current battlecard or late-stage deals are weakly qualified. Steps: Review recent losses by competitor; Publish or refresh battlecards for your top 2 competitors (can be skipped); Score frameworks on late-stage deals. Measured on the win rate for 45 days.
Value not evidenced: Won customers have no recorded proof of value.Suggested when the constraint is Onboarding Velocity (won to evidenced). Steps: Create value contracts for recent wins; Log the first impact for each new contract. Measured on that rate for 30 days.

Every playbook ends with a measure step, such as Measure after 14 days. Some constraints have no playbook yet (Engaged → Replied, Expansion Rate and Renewal Rate), so they show no Fix this button.

The panel

The panel is headed Fix playbook and the playbook’s title, with a status: Not started (a preview), In progress, Measuring, Finished or Abandoned.

Why this was suggested appears when this is the playbook suggested for your first constraint: the stage, the reason, and any Likely causes that match the playbook.
What you’ll do (in a preview) or Steps (in a run). Each step shows its state: Current step, Up next, Done · <date>, Skipped · <date> or Done automatically · <date>. Every action step has Go → to the page where the work happens.
Some steps Checks automatically: they complete when Summit53 sees the work (a published ICP, a finished rescore, five good-fit leads enrolled, two battlecards published, three late-stage deals scored, a value contract created) done after the step started. The others need someone to press Done.
On the current step, people who can act see Done, Skip (where allowed) and Add a note (What you did, or why you skipped it, up to 500 characters). A finished step in an active run can be Reopened, which also resets the steps after it.
Outcome: in a preview, When every step is done, we watch the <rate> rate for N days and show what changed after. In a run, the rate at the start, then Measuring until <date>, then the rate after and a result: Improved, Didn’t improve or Not enough data yet.
The footer has Start this playbook (or Start a new run after a finished or abandoned one), Abandon during a run (Abandon this run? It won’t be measured.) and Close.

The rules

One open run per playbook. A run that’s measuring still counts as open. Starting another shows the open one: This playbook already has an open run, so here it is.
Leaders start and abandon runs. That’s a workspace owner or admin, or anyone with a leader role. Others see Ask a sales leader to start this.
The leader who started a run, or any leader, moves its steps. Everyone else can open the panel, follow the steps and use the Go → links: A sales leader runs this playbook. You can follow along.
Measuring starts when every step is done or skipped, and lasts the playbook’s window.
A run that never finishes is closed with no result after 90 days.

Run a Fix playbook

On the Revenue Engine, click Fix this: <playbook> on your first constraint.
Read Why this was suggested and What you’ll do.
If you’re a leader, click Start this playbook. The rate at the start is recorded.
Work the current step with Go →. Press Done on manual steps; automatic steps tick themselves. Add a note where it helps the next person.
When the last step is done, the status changes to Measuring with an end date.
After that date, open the run from the Loop to read the outcome.

When the data is thin

What it can’t see. If a conversion step is under its sample floor when you start, the starting rate reads Not enough data yet, and the result can end the same way: Not enough data yet to say what changed after this run.
How the result is judged. The rate has to rise by at least 5 points to count as Improved. A smaller move counts as improved only if the signal that prompted the fix has cleared, meaning the rate is now healthy against its benchmark; otherwise it’s Didn’t improve.
Correlational, said plainly. The panel says it: This is what changed after the run, not proof the run made it happen. Other things moved too.
Two fixes, one rate. Stalled pipeline and Win rate slipping both watch the win rate. If both run at once, their results can’t be told apart.
Abandoned runs have no result (Abandoned before measuring, so there’s no result.) and drop out of the Loop.
What it changes. Opening a playbook or a preview changes nothing. Starting, completing steps, adding notes and abandoning are recorded on the run.

Do it this week

Write down your first constraint and its rate today, with the date.
Write three steps in order: what you’ll look at, what you’ll change, and how you’ll put the change to work. Give each an owner.
Decide the measurement window now: about two weeks for lead steps, six for closing deals.
Put the end date in the calendar, and don’t check the result before it.
On that date, write one sentence: “After the fix, the rate went from X to Y.” Not “the fix moved it”.

Troubleshooting

There’s no Fix this button. No playbook matches your constraint yet, or nothing is constrained.
I can’t start a run. Starting needs a leader role or workspace owner or admin access.
An automatic step isn’t completing. It only counts work done after the step started. Do the work again (for example, publish the ICP) or check you’re in the right workspace.
I can’t start a new run. One is still open or measuring. Finish or abandon it, or wait for its result.
The result says Not enough data yet. Too few records went through that conversion step during the window to judge it.
I abandoned a run and it disappeared from the Loop. That’s expected: abandoned runs aren’t measured.

Diagnostics: find your first constraint
Signals & the Loop: where runs are counted
Blueprint & Knowledge: the bets several steps revise
Action plans & tasks

Fix playbooks: work one constraint to a measured result