QBR Deck Outline
Nine slides. 30-minute meeting. Deck sent 48 hours early as a standalone pre-read. Slides 1-6 make the quarter legible (pre-read), slides 7-9 make the next quarter decidable (the meeting). The room...
From this piece
The template
QBR Deck Outline
Nine slides. 30-minute meeting. Deck sent 48 hours early as a standalone pre-read. Slides 1-6 make the quarter legible (pre-read), slides 7-9 make the next quarter decidable (the meeting). The room's energy belongs to slides 7-9.
Ground Rules
- Same five scoreboard numbers every quarter. No metric shopping. A team that changes its scoreboard has no scoreboard.
- Pre-read protocol: deck out 48 hours before the meeting, full speaker notes included so it reads standalone. If attendees arrive unread, the first 10 minutes become silent reading. They will only let that happen to them once.
- Speaker notes carry the evidence and sources. The slide carries only the claim. If a slide needs a paragraph, the paragraph goes in the notes.
- The meeting ends on slide 9 with answers, not on slide 1 with applause.
Time Budget (30 minutes)
| Slides | Minutes | Job |
|---|---|---|
| 1-6 (legibility) | 5 | Confirm the pre-read landed, take questions on misses only |
| 7 (bets) | 12 | Debate and weigh the portfolio |
| 8 (anti-bets) | 8 | Settle what is out, on the record |
| 9 (asks) | 5 | Get the decisions |
Slide 1: Scoreboard vs Plan
Time: part of the 5-minute pre-read confirmation.
| Metric | Target (set Q start) | Actual | Gap |
|---|---|---|---|
| Direction metric: [e.g. weekly active accounts using X] | [t] | [a] | [+/-] |
| Eval trend: [score on most important AI surface] | [t] | [a] | [+/-] |
| Cost per outcome: [unit economics line] | [t] | [a] | [+/-] |
| Adoption: [uptake of the quarter's main bet] | [t] | [a] | [+/-] |
| Customer-signal theme: [top theme, mention count] | [expected] | [actual] | [shift] |
Guidance: the SAME five metrics every quarter, in the same order. Targets are the ones you committed at quarter start, not revised ones. Speaker notes: one line per metric on the driver of the gap. No spin; slide 2 will expose it anyway.
Slide 2: What We Said vs What Happened
The honesty slide. The room calibrates its trust on this one.
| We said (verbatim, last quarter's deck) | What happened |
|---|---|
| "[Exact quote from last QBR slide 7]" | [Reality, one line, unspun] |
| "[Quote]" | [Reality, including the miss] |
| "[Quote]" | [Reality] |
Guidance: quotes are verbatim from last quarter's deck. Include every commitment, hit or missed. Misses stated plainly: "We said the migration would finish in Q2. It is 60% done." Not "migration progressing." Spin this slide once and slides 3-9 get discounted for a year. Speaker notes: for each miss, one line on cause (bad estimate, changed priority, external event) without excuse language.
Slide 3: Decision Review
The quarter's big calls, scored two ways. Decision quality and outcome quality are separate columns (Annie Duke: don't judge the bet by the result).
| Call | Decision quality (given evidence at the time) | Outcome |
|---|---|---|
| [Sunset the legacy reporting module] | Good: 3 quarters declining usage, tripwire defined upfront | Good: no churn, support load down as projected |
| [Sequenced onboarding ahead of pricing test] | Good: funnel evidence pointed there | Miss: pricing was the bigger lever |
| [Call] | [verdict + why, one line] | [what happened, one line] |
Guidance: 3-5 calls. All four quadrants are legal verdicts: good/good, good/bad, bad/good, bad/bad. A column of all-good verdicts means you are resulting, not reviewing. If you keep a decision log, this slide assembles itself from the quarter's rows. Speaker notes: the evidence each call was made on, and the lesson line from the log.
Slide 4: What We Killed
| Killed / declined | Why (one line) |
|---|---|
| [Initiative or request] | [reason] |
| [Initiative or request] | [reason] |
Guidance: the quarter's subtractions. Include senior-sponsored items you declined; those are the highest-signal rows. A quarter with zero kills is a yellow flag, not a clean record. Speaker notes: what the kill freed up (people, weeks, support load) and where it went.
Slide 5: Customer Signal Themes
| Theme (named as a pattern, not a category) | Frequency | Verbatim quote |
|---|---|---|
| [e.g. "Bulk export friction blocking weekly reporting"] | [n calls / tickets] | "[Exact customer words, with source]" |
| [Theme] | [n] | "[Quote]" |
| [Theme] | [n] | "[Quote]" |
Guidance: 3-5 themes that recurred across calls and tickets this quarter. Theme names are specific patterns ("export friction blocking weekly reporting"), never categories ("product feedback"). Quotes are verbatim; paraphrase turns evidence into opinion. Speaker notes: which next-quarter bet (slide 7) each theme feeds, if any. A theme feeding no bet needs a stated reason.
Slide 6: Margin and Cost per Outcome Trend
One chart: cost per outcome across the quarter's months (or weekly), with the trend line.
Below the chart, two lines:
- Biggest driver up: [e.g. model cost per generated report, support minutes per account]
- Biggest driver down: [e.g. caching, prompt compression, deflection rate]
Guidance: for AI-touched products this slide is not optional; the cost line is product work now. If margin is finance's number and you cannot produce this slide, that gap is itself a finding worth saying out loud. Speaker notes: the math behind cost per outcome, sources, and what would bend the curve next quarter.
Slide 7: Next Quarter Bets
The meeting's main event. 12 minutes.
| Bet | Outcome it targets | Confidence | Door | Displaces |
|---|---|---|---|---|
| [Bet name, one line] | [the number it should move, by how much] | [%] | [one-way / two-way] | [what stops or shrinks to fund it] |
| [Bet] | [outcome] | [%] | [door] | [displaced] |
| [Bet] | [outcome] | [%] | [door] | [displaced] |
Guidance: 3-5 bets, never more. Confidence is a number; "high" cannot be debated, 70% can. Every bet names what it displaces, because a bet that displaces nothing is a wish. One-way doors should have a premortem already run (note it in speaker notes). Speaker notes: the evidence behind each confidence number and the kill criteria per bet.
Slide 8: Anti-Bets
What we are explicitly NOT doing next quarter. Usually the most discussed slide in the room. 8 minutes.
| Not doing | Why (one line) | Revisit if |
|---|---|---|
| [The thing people keep asking about] | [reason] | [tripwire that would reopen it] |
| [Zombie initiative] | [reason] | [tripwire] |
| [Plausible adjacent bet] | [reason] | [tripwire] |
Guidance: name the ideas that re-litigate every planning cycle and kill them here, in writing, in front of leadership. The "revisit if" column converts a no into a tripwire, which makes the no easier to accept. Speaker notes: who has sponsored each item before, so you walk in ready for the pushback.
Slide 9: Asks
The meeting ends here, with answers. 5 minutes.
| Ask | Decision needed | By | Default if undecided |
|---|---|---|---|
| [Ask] | [the yes/no or A/B] | [date] | [what you do absent an answer] |
| [Ask] | [decision] | [date] | [default] |
Guidance: every ask is a decision, not an FYI. The default column converts silence into an answer. 1-3 asks; more than three means you have not prioritized your own needs. Do not let the meeting end before each row has a yes, a no, or an owner-and-date. Speaker notes: the one-paragraph case for each ask, ready to deliver verbally.
Building Checklist
- Build slide 2 first. It is the hardest and it determines whether you have the honesty the rest depends on.
- Build slide 7 second. The bets table is where prep time pays.
- Pull slide 3 from your decision log, slide 5 from your call-triage themes, slide 1 and 6 from the same numbers your weekly exec update tracks. If those systems exist, the QBR is assembly, not authorship.
- Write speaker notes for every slide so the deck reads standalone.
- Send 48 hours early. Put the pre-read rule in the calendar invite text.
- In the meeting: 5 minutes on slides 1-6, 25 minutes on slides 7-9. Leave with every slide-9 row answered.
Nine slides. The first six earn the trust, the last three spend it on decisions. If the room applauds and decides nothing, the deck failed, however good the quarter was.