A weekly business review that survives: which metrics belong in the pack, how the numbers get pulled before the meeting, and what the room should decide.
A weekly business review is a fixed one-hour decision meeting run from a metric pack that was assembled before anyone sat down, with every decision tracked until it closes. For the DTC or marketplace operator who loses a weekend pulling Shopify, Stripe and ad numbers into a deck, the fix is mechanical: settle the pack once, build it before Monday, spend the hour on variances, chase the actions afterwards.
What a weekly business review actually is
A weekly business review (WBR) is a decision forum that answers three questions about the week that just ended. Colin Bryar and Bill Carr, who describe Amazon’s operating cadence on their Working Backwards site, put the first two as “What did our customers experience last week?” and “How did the business perform last week?”, with the third asking whether the business is on track to hit its goals. The same page notes the WBR “occurs every week, year-round.”
That draws the boundary with the other weekly meeting. A status review is a broadcast of updates that could have been a Slack post, which is why the async playbook argues that standup, weekly sync and status review should be replaced with reports. The WBR survives because the room decides something: move budget, pause a SKU, fix a fulfillment lag, change a target. A review that decided nothing has drifted into the meeting the playbook says to delete.
Pick controllable inputs before lagging outputs
The pack has two kinds of numbers. Bryar and Carr, who spent a combined 27 years as senior Amazon executives and wrote the book Working Backwards, split metrics into controllable inputs and outputs. Their guide to input metrics states: “Some companies refer to these as leading indicators, but we prefer a more precise term: controllable input metrics.” Outputs are the lagging results. The same page says “Output metrics like orders, revenue, and profit are essential but generally can’t be directly or sustainably manipulated over the long term.”
For a DTC operator the inputs are ad spend by channel, blended CAC, creative refresh rate, email send volume, in-stock rate and ship time: things the room can change this week. The outputs are ROAS, net revenue, contribution margin, repeat purchase rate and return rate: things the room reads and explains.
Outputs still open the pack. The input metrics guide says “We started every WBR at Amazon by reviewing our progress against our financial plan for the quarter.” The first page is revenue and margin against plan; everything after it is the levers that explain the gap.
Which metrics belong in the pack, and at what window
The Amazon-style format plots each metric in one chart with two horizons. Per the same input metrics guide, “For each metric, the chart plots the six-week trend and the twelve-month trend side by side,” and the line beneath the six-week trend “is the same metric for last year for the same weeks the previous year.” Bryar and Carr’s WBR App page calls this the 6-12 chart and notes that because the format does not change from metric to metric, “you can review hundreds of them in a one-hour meeting.”
A lean DTC team should not try for hundreds. Start with a dozen or so metrics on the same six-week and twelve-month view with a prior-year line, and add one only when it would change a decision. A first pack:
Metric | Input or output | Window | Owner |
|---|---|---|---|
Net revenue vs plan | Output | 6 weeks, 12 months, prior year | Founder or GM |
Contribution margin | Output | 6 weeks, 12 months, prior year | Finance |
Orders and average order value | Output | 6 weeks, 12 months, prior year | Ecommerce lead |
Repeat purchase rate | Output | 12 months, prior year | Retention lead |
Paid spend by channel | Input | 6 weeks | Growth lead |
Blended CAC | Input | 6 weeks, prior year | Growth lead |
New creatives launched | Input | 6 weeks | Creative lead |
Email and SMS sends and revenue | Input | 6 weeks | Retention lead |
In-stock rate on top 20 SKUs | Input | 6 weeks | Ops lead |
Order to ship time | Input | 6 weeks | Ops lead |
Support first response time | Input | 6 weeks | Support lead |
Return rate | Output | 6 weeks, 12 months | Ops lead |
See the jobs teams hand to Viktor →
Assemble the pack before anyone sits down
This is the unglamorous part, and the reason weekly reviews quietly stop. The numbers above live in Shopify, Stripe, Meta Ads, Google Ads, the email platform and the CRM. Each tool defines a week differently, reports gross or net differently, and exports on its own schedule. Someone spends Sunday reconciling them, and after a month of Sundays the pack arrives late, then thin, then not at all. The meeting runs on memory and opinions.
Two rules keep the pack alive. First, freeze the definitions: one week boundary, one revenue definition, one attribution window, written next to the chart. Second, nobody assembles the pack by hand on the day. The pull runs on a schedule and lands before the pre-read.
The mechanics of that pull are covered in how one Slack message replaced four hours of weekly reporting, with the exact prompt and the before and after. In short: Viktor, an AI employee who lives in Slack and Microsoft Teams, runs the Monday pull as a recurring Task. A recurring Task, as the setup guide puts it, “is a standing instruction plus a schedule.” Describe the pack once, tell him it runs every Monday at 7am into #wbr, and he pulls the numbers from the connected tools and posts the pack in Slack with what moved and why.
The integrations page lists Shopify, Stripe, Meta Ads, Google Ads and HubSpot among the 3,200+ tools he reaches, and states “Most connectors are one-click OAuth - pick a tool, authorize, done. Some use API keys.” The consequential actions wait for a person: the security page says “Money moves, code pushes, and customer emails wait for your explicit approval in Slack.” The pack assembles itself; a money move the room decides on still waits for a person to approve it.
Hand the Monday pack to Viktor →
Who attends and who owns each metric
The meeting owner is the founder, GM or head of operations. They keep time, decide what gets discussed and what gets parked, and close with the decision list.
The function leaders attend because they own metrics. Growth owns spend, CAC and creative throughput. Ops owns in-stock rate, ship time and returns. Retention owns repeat rate and email revenue. Finance owns margin. Each metric has exactly one named owner, who explains any variance before the meeting in a line under the chart: what moved, why, what they are doing about it.
The note taker is a separate person from the owners, and their only output is the action log: decision, owner, due date.
Connect Shopify, Stripe and your ad accounts →
A 60-minute WBR agenda that stays a decision forum
Fixed day, fixed hour, fixed length. Monday or Tuesday morning, after the pack has landed, works for a DTC brand because the week’s ad budget is still adjustable.
Minutes | Block | What happens |
|---|---|---|
0 to 10 | Silent pre-read | Everyone reads the pack and the owners’ variance notes. No talking. |
10 to 20 | Outputs vs plan | Revenue, margin, orders against plan and prior year. The meeting owner names the gap. |
20 to 40 | Variances only | Only metrics that moved outside their normal range get discussed, two to three minutes each. Charts that look like last week are skipped. |
40 to 50 | Customer experience | Ship time, support response, returns, plus two or three actual customer complaints from the week. |
50 to 58 | Decisions | The owner states each decision aloud: what changes, who does it, by when. |
58 to 60 | Confirm the log | The note taker reads back the action list. |
The silent pre-read makes the rest work: it removes the presenting and leaves the hour for the charts that need a decision.
What happens to the decisions afterwards
A decision that is not tracked was never made. The action log needs four fields: decision, owner, due date, and the metric it should move. It lives in the same channel every week and opens the next review: last week’s items, closed or not.
Misses need an escalation path that is not a blame session. Amazon’s mechanism is the Correction of Error. The Working Backwards COE template describes it as “a scalable, repeatable process” to “Identify and fix the root cause(s) of outages, and events (typically one-time or edge case events) that negatively impact the customer experience and/or business/financial results.” AWS, in its own post on the COE, stresses that “the focus is on corrective actions, not just documenting failures.” For a DTC operator the triggers are a stockout on a hero SKU, a carrier failure, an ad account suspension, a checkout outage.
The root cause tool the template suggests inside a COE is the Five Whys. The same Working Backwards template explains that “Each answer to the question ‘why did this occur’ then becomes the starting point for the next why.” Run it on the miss, not the person.
Because Viktor runs recurring work on schedule, a second Task can post the open action list with owners and due dates into the channel before each review, so the meeting opens on what was promised.
WBR vs monthly business review vs QBR
Bryar and Carr’s page on monthly and quarterly business reviews describes the MBR or QBR as a narrative document of “6 pages or fewer” reviewed in a one-hour meeting with the CEO, held “within 5-10 business days” after the period ends, that checks progress against the annual operating plan. The same page says “about 80% of the metrics and discussion should focus on controllable and mostly customer-facing inputs.”
| Weekly business review | Monthly business review | Quarterly business review |
|---|---|---|---|
Cadence | Every week, year-round | Within 5 to 10 business days after month end | Quarterly where the page says that makes more sense; same document and one-hour meeting |
Metrics | Full pack of inputs and outputs on six-week and twelve-month charts | Key inputs and outputs vs plan and prior year, about 80% inputs | Same as monthly, at quarter level |
Audience | Metric owners, meeting owner, note taker | Unit leader, CEO, direct reports, relevant stakeholders | Unit leader, CEO, direct reports, relevant stakeholders |
Format | Charts with owner notes, silent pre-read | Narrative of six pages or fewer plus appendix tables | Narrative of six pages or fewer plus appendix tables |
Purpose | Decide this week’s actions on variances | Check progress against the annual plan, adjust it | Check progress against the annual plan, adjust it |
Why these reviews fail and how to keep yours
Three failure modes kill reviews.
Drift into status. Owners narrate their charts instead of explaining variances. The fix is the silent pre-read and the rule that unchanged charts are skipped.
Performative rigor. The pack keeps growing because every function wants representation. Any metric nobody has acted on in eight weeks leaves.
Assembly fatigue. The person building the pack burns out, the pack arrives late, the pre-read is skipped, and the meeting runs on opinion. The fix is mechanical: freeze the definitions and schedule the pull.
None of this needs a particular tool. A small team can run it from a Google Sheet and a Monday calendar block, as long as someone owns the pull and it lands on time. When that someone becomes the bottleneck, the pull should become a scheduled Task, so the pack is in the channel before the first person opens their laptop.
Pricing for Viktor is usage-based per workspace with no per-seat licenses, with up to $100 in free credits and no credit card to start. Pricing checked October 2026.
See usage-based pricing: up to $100 in free credits and no credit card →
FAQ
What is the difference between input and output metrics in a WBR?
Inputs are the levers the team controls this week: ad spend, creative volume, inventory position, ship time. Outputs are the lagging results: revenue, margin, ROAS, repeat rate. Bryar and Carr call the first group “controllable input metrics”. The pack opens with outputs against plan and spends its time on the inputs.
Who should own the weekly business review?
The founder, GM or head of operations owns the meeting: timekeeping, what gets discussed, the closing decision list. Each metric has its own named owner who writes the variance note before the meeting. A separate note taker keeps the action log.
How long should a weekly business review take?
Sixty minutes, same day and time every week: ten minutes of silent pre-read, ten on outputs against plan, twenty on variances, ten on customer experience, ten on decisions. If it runs long, the pack is too big or the pre-read is being skipped.
What tools or templates do you need to run a WBR?
A pack of charts in one format, an action log, and a scheduled pull that builds the pack before the meeting. The Working Backwards site publishes a free WBR App and a sample WBR report for an e-commerce company. A spreadsheet holds the pack; the pull can be a recurring Task that assembles Shopify, Stripe and ad data every Monday.
What is the difference between a WBR, an MBR and a QBR?
The WBR is a weekly, chart-driven decision meeting on last week’s inputs and outputs. The MBR and QBR are narrative reviews of progress against the annual plan: six pages or fewer, discussed with the CEO in a one-hour meeting within 5 to 10 business days of period end. The weekly meeting decides this week’s actions; the others adjust the plan.
What should happen when a metric misses two weeks in a row?
The owner’s variance note should already say why and what is being done. If the miss persists, open a Correction of Error: write up the impact, run the Five Whys to the root cause, and log corrective actions with owners and dates. The point is the fix, not the blame.