Why a ritual and not a report
Most companies have plenty of dashboards and no rhythm. Numbers get looked at when something feels wrong, which means they get looked at through panic. The WBR is the opposite bet: look at the same numbers in the same order every single week, so that the week something actually changes, it cannot hide. This session is about the ritual itself - the agenda, the questions, and the failure modes.
The WBR as a ritual 8 min live
A WBR is 45 minutes with a fixed running order: scorecard first (the core four - GMV, orders, conversion rate, AOV - each with week-over-week and year-over-year), funnel second (the four gates from a3), channels third (where the traffic and money came from), then actions. The order never changes. That is not bureaucracy - it is the whole trick.
LiveWhy fixed order beats ad-hoc4 min▶
When the format changes every week, two things move at once: the numbers and the frame. Nobody can tell whether GMV "looks different" because the business changed or because the chart did. Hold the frame perfectly still and your team develops something valuable: pattern memory. After eight identical weeks, everyone in the room knows what slide three normally looks like - so the week it looks wrong, five people notice before the presenter finishes the sentence.
- Fixed order = free anomaly detection. The human eye is superb at spotting change against a constant background and terrible at spotting it against a moving one.
- Fixed order = no burying. When the deck is rebuilt weekly, a bad number can quietly migrate to slide nine. A fixed agenda makes omission visible.
- Time boxes protect the actions. Scorecard and funnel are review, not discussion - 10 minutes each. The 25-minute block at the end is where debate belongs, because that is the part that ends in an owner and a date.
Self-studyWho is in the room, and what each person brings3 min read▶
The ritual has roles, and they are not interchangeable:
- The analyst brings the deck - same template, numbers refreshed, anomalies pre-flagged. They present at pace and are never asked to compute live.
- Gate and channel owners attend for their block. When question 7 names their gate, the answer comes from the person who can actually change it, not from the presenter.
- The leader brings the question bank and enforces the time boxes. Your scarcest contribution is not insight - it is refusing to let the room skip the order.
Cadence matters too: early in the week, so the re-check dates from last Monday come due in the same meeting that set them. A WBR that drifts around the calendar loses its own follow-ups.
The question bank 10 min live
You do not need to compute anything in a WBR - you need to route. Any drop in any metric can be walked to its owner with a handful of questions, asked in roughly this order. Print this table; it is the leader's half of the meeting.
| # | The question | What it isolates |
|---|---|---|
| 1 | Which lever moved - traffic, conversion, or order value? | Splits GMV into its three drivers (a1). Everything else follows from this answer. |
| 2 | Is it one channel or all channels? | One channel = an acquisition story. All channels = a site or market story. |
| 3 | One device or all devices? | Mobile-only drops usually mean an experience or release problem, not a demand problem. |
| 4 | What did the same week last year look like? | Seasonality vs real change - the a5 question. |
| 5 | Did tracking or a definition change? | Data problem vs business problem. Ask before anyone panics - a6 makes this systematic. |
| 6 | New or returning customers? | Acquisition lever vs loyalty lever - two different teams, two different fixes. |
| 7 | Which funnel gate moved? | Names the owning team, exactly as drilled in a3. |
| 8 | Is this outside normal weekly wiggle? | Every metric breathes. Chasing noise burns the team; ask how unusual this really is. |
| 9 | What action, what owner, when do we re-check? | The closing question. A drop discussed without these three answers was entertainment, not review. |
Self-studyReading WoW and YoY together3 min read▶
The scorecard shows every number twice - against last week and against the same week last year - because the pair says more than either alone:
| WoW | YoY | Most likely story |
|---|---|---|
| Down | Flat or up | Weekly wiggle or calendar effect - watch, don't react yet |
| Flat | Down | Slow erosion - the most dangerous pattern, invisible week to week |
| Down | Down | A real drop - run the question bank now |
| Up | Up | Genuine growth - still ask which lever, so you can do it again |
The second row is why YoY earns its place on every slide: a metric can drift down 1% a week forever without any single week looking alarming. Only the year-ago comparison catches it.
Anti-patterns to ban 10 min live
WBRs rarely die dramatically. They rot - slowly, politely, one bad habit at a time. Four rot patterns account for most of it, and all four are visible from the leader's chair.
LiveThe four rot patterns6 min▶
- Metric theater. Numbers reported with no owner and no possible action - a slide everyone watches and nobody uses. Test any metric with: "who changes what they do if this moves?" No answer, no slide.
- Anecdote-first meetings. "A customer emailed me about shipping" becomes the agenda, and the scorecard never gets read. Anecdotes are excellent hypotheses and terrible openers - park them for block three, where the data can meet them.
- Comparing to last week during a seasonal peak. Every week of a peak "beats last week" on the way up and "collapses" on the way down, and both readings are noise. In peaks, YoY beats WoW - a5 gives this its full treatment.
- Changing definitions mid-quarter. Redefine "conversion rate" in week 7 and every trend line on the deck silently breaks. Definitions change at quarter boundaries, in writing - that discipline is a6's whole subject.
The Amazon-style WBR. Amazon is widely described - in books by and about its executives, not in any single official document, so treat this as a well-attested paraphrase - as running weekly business reviews on a fixed deck of standing metrics, walked at speed, where anything in line with expectations gets seconds and anomalies get the meeting. The lesson survives the paraphrase: the deck is boring on purpose. A WBR where every slide is interesting is a WBR where nothing is being compared to anything.
A mini-WBR: the week GMV fell 6% ★ 10 min · everyone reads
Mango Lane, Monday morning. The headline: GMV came in around $621,000 against the usual $660,960 - down 6% week over week. The room tenses. Run the ritual, in order, and watch it turn a scary headline into a 15-minute diagnosis.
LiveWalking the three blocks8 min▶
Block 1 - scorecard (10 min). GMV -6% WoW. AOV flat at $68. Conversion rate flat at 2.7%. Orders down 6% - and sessions down 6% with them. Question 1 answers itself: the lever that moved is traffic. Conversion and order value are innocent; do not let anyone start redesigning checkout.
Block 2 - funnel (10 min). All four gates flat: PDP penetration 40%, add-to-cart 25%, checkout start 45%, completion 60%. Exactly snapshot 3 from a3 - the funnel kept its shape and shrank. The site is fine; fewer people walked in.
Block 3 - channels (say 15 of the 25 min). Organic, CRM, and social: flat. Paid: sessions down roughly 20% - which, at paid's share of traffic, accounts for the whole 6%. One channel, not all channels. Now the anti-panic question: same week last year? No, this is not seasonal. Did tracking change? The analyst confirms tagging is intact - so it is a real paid-traffic drop, likely budget, campaign, or platform-side.
Close - action, owner, re-check. Action: audit paid campaigns and spend pacing against last week, check for a platform policy or bidding change. Owner: the growth lead. Re-check: next Monday's WBR, same slide. Written down before the meeting ends.
Try it yourself - this week ◐ 20 min total
- Write your own three-slide WBR skeleton: slide 1 scorecard (your core four, each with WoW and YoY), slide 2 funnel (your four gates), slide 3 channels + an actions table with columns for action, owner, re-check date.
- Next to every number on the skeleton, write who owns it - the person who changes what they do when it moves. Any number with no name next to it is a candidate for deletion (metric theater).
- Bring the skeleton to your next team meeting and run the first ten minutes as block 1 - scorecard only, no discussion. Notice how strange, and then how useful, the discipline feels.
Sources this session draws on
The WBR is a practice, not a spec - so this session leans on well-attested industry patterns for the ritual and on platform documentation for the numbers inside it. This page draws on:
Three questions before you go 🎯 ◐ 90 seconds
1 · Why does the WBR keep the same numbers in the same order every single week?
A fixed frame gives the room pattern memory: after eight identical weeks, everyone knows what slide three should look like, so the week it looks wrong, it cannot hide. Moving frames camouflage moving numbers.
2 · GMV fell 6%. The funnel gates are all flat and AOV is flat. What is the first routing question?
GMV has three levers: traffic, conversion, order value. Flat gates clear conversion; flat AOV clears order value. The drop must live in traffic - and "one channel or all?" is how you walk it to an owner, as in the mini-WBR where paid alone had fallen.
3 · A metric has appeared on the deck every week for a quarter. It has no owner and has never produced an action. What is it?
The test for any WBR metric is "who changes what they do if this moves?" No answer means the slide is decoration. Reporting without owners is the first and most common way weekly reviews rot.