Playbooks › Why did the number move?

Why is the number different from last week? Explain the movement

Why is the number different from last week? Explain a forecast move in four parts: level, movement, what moved it and what you can't see, ready for the call.

Summit53 Team8 min readUpdated October 23, 2026

The question

Last week you said one number. This week it’s different. Why?

You can always produce a number. The question in the forecast call is why it moved, and the answer has to survive the follow-up. “A couple of deals slipped” rarely does.

Where it sits on the loop

Where Pipeline turns into Won: the weekly forecast or commit call. The number is built from parts that reach across the loop: how the pipeline is shaped, how well deals are being worked, how clean the data is, renewals, and how accurate past forecasts were. The parts are how you explain a movement.

How to think about it

Every movement answers four questions, in the order they get asked.

  • Level. What is it now: the number, the unit and the date you read it.
  • Movement. What it was, when, and the signed difference. Check both dates before the number: “last week” is sometimes whenever someone last saved a snapshot.
  • Attribution. Which parts moved it, each with its sign, and they have to add up to the total. If they don’t, the gap is part of the answer.
  • Coverage. How much of the move you’ve explained, and what you aren’t measuring.

Deals are evidence, not arithmetic. Lead with the part of the number that moved, then name the deals that are consistent with it. The answer fits in one breath: “We’re at X, down Y on last week. Execution accounts for most of it; these three deals are consistent with that. Forecast accuracy isn’t measured yet.”

Common mistakes

  • Answering with the level. “We’re at 50%” is the number. Nobody asked for the number.
  • Telling a deal story as the explanation. Deals might be true, but they’re corroboration. Say “consistent with”, not “caused”.
  • Comparing the wrong window. Two snapshots five weeks apart, presented as last week.

In Summit53

The four answers sit on the Revenue Engine home, and a skill puts them on one slide for the call.

1. Level: the forecast call

  1. Open the Revenue Engine and check the period in the dropdown next to REVENUE ENGINE.
  2. Read the headline (for example FY2026 Q4 · NZ$13.7M forecast) and the Forecast Rail’s basis line: projected landing · rep forecast calls, or the fallback closed-won + weighted pipeline.
  3. Click the FORECAST tile for Forecast Command: Most Likely, the Low to High range, the confidence badge and the trend label.

See Revenue Engine masthead and Forecast Command.

2. Movement and attribution: why confidence moved

  1. Scroll to Forecast gap diagnosis and, at its foot, WHY DID CONFIDENCE MOVE?
  2. Read the move in points, then the five drivers: Pipeline shape, Execution health, Data quality, Account & renewal and Forecast accuracy. Each measured driver shows its share; together they add up to the total move.
  3. A driver that reads building… isn’t measured yet. Say so in the call rather than treating it as zero.

See Signals & the Loop.

3. Corroboration: the deals behind it

  1. Compare Most Likely with WEIGHTED PIPELINE on the rail. A gap between what reps say and what the pipeline says is worth naming.
  2. When confidence is below 60%, follow View At-Risk Deals in Forecast Command to Risk Signals.
  3. Pick the deals that match the driver that moved, and describe them as consistent with it.

From Claude or Codex

With the Summit53 plugin, the forecast-delta-brief skill answers all four questions on one slide for the call, with a five-line summary in the chat. Ask, for example:

  • Why is the forecast number different from last week? Prep me for the commit call.
  • What moved the forecast this week?
  • Forecast confidence dropped — why?

It compares exactly seven days, shows the parts adding up to the headline (with any remainder), names up to three deals consistent with the movement, and lists what it can’t see. It saves nothing. See forecast-delta-brief.

When the data is thin

  • Which method. If fewer than 70% of quota-carrying sellers have a call from the last 14 days, the headline falls back to closed-won plus weighted pipeline, and the rail says so.
  • Not measured yet. Drivers that read building… are unmeasured, not zero. The skill lists them as not yet measured and doesn’t plot them.
  • No history yet. WHY DID CONFIDENCE MOVE? doesn’t appear until there’s an earlier reading to compare. The skill gives a short readiness note instead of inventing a movement.
  • Whose numbers. The masthead shows the widest view you’re allowed. The skill asks for the organisation-wide view and labels the slide if it only got a team view.
  • Consistent with, not caused by. Deals corroborate a driver’s direction; they don’t prove which deals moved the number.

Do it this week

  1. On Friday, save the number and its parts, with the date.
  2. On Monday, read the number and its parts again, and check the two dates are a week apart.
  3. Work out each part’s change and check they add up to the total. Write down any remainder.
  4. Write the one sentence: level, movement, the main driver, and what you’re not measuring.
  5. Add up to three deals, as “consistent with”.

Troubleshooting

  • The headline doesn’t match the ML forecast. Expected: the headline is your sellers’ call; the ML figure is a one-month input.
  • WHY DID CONFIDENCE MOVE? is missing. There’s no earlier reading to compare yet.
  • The skill gave a readiness note instead of a slide. It couldn’t find a reading from seven days ago, or the reading wasn’t from today. Run it again once there’s a week of history.
  • My numbers differ from my manager’s. The masthead shows each person their own scope: organisation, team or own book.