Weekly Business Review: What It Is & How To Run One [Templates]

A weekly business review (WBR) is a fixed-format meeting where leaders walk the same set of metrics every week, see each one against last week, plan and the same week last year, and hear the owner explain anything that moved. Amazon made the format famous: its leadership has held one every week since the early 2000s, and it puts more weight on the controllable inputs that drive results than on the financial results themselves.

Review	Audience and question	Comparisons
WBR (weekly)	Leadership and metric owners: what moved last week, and who is acting on it?	Week on week, vs plan, vs same week last year
MBR (monthly)	Leadership and function heads: are we on plan for the month and the year?	Month on month, vs plan, vs last year, year to date
QBR (quarterly)	Leadership: did the quarter's bets work, and what changes next quarter?	Quarter on quarter, vs plan, vs last year, full-year outlook
Board pack (quarterly)	The board: is the company on track, and what does the board need to decide?	Vs budget, vs last year, trailing quarters, cash runway

What is a weekly business review?

Amazon's WBR is the best-documented version. Commoncog, which has interviewed former Amazon executive Colin Bryar about it at length, reports that Amazon's leadership has met for it every week since the early 2000s, even when the CEO or CFO is away. Bryar and Bill Carr describe it in their book Working Backwards, and Commoncog's podcast episode with Bryar covers its history and mechanics.

The deck is a weekly set of input metrics, output metrics, exception reports and customer anecdotes presented to company leadership (Holistics). Three ideas make it more than a status meeting.

Controllable input metrics vs output metrics

Output metrics are the results you report: revenue, orders, profit, retention. Nobody can move them directly this week. Controllable input metrics are the levers that lead to those results, the things a team can change by Friday. Spend the meeting on the inputs.

Jeff Bezos put the priority plainly in Amazon's 2009 letter to shareholders: "Senior leaders that are new to Amazon are often surprised by how little time we spend discussing actual financial results or debating projected financial outputs." The same letter counts the company's goals for 2010: 452 of them, "with owners, deliverables, and targeted completion dates." Across those goals, revenue appears eight times, free cash flow four times, and net income, gross profit and operating profit not at all.

Finding the right inputs takes iteration. In the classic example from Working Backwards (Holistics summarizes it), Amazon tracked new product detail pages, teams added pages in bulk, sales didn't follow, and the metric became detail page views. An input that stops predicting its output gets changed or dropped.

The same charts, in the same order, every week

Working Backwards describes charts that put the trailing six weeks and the trailing twelve months side by side on the same x-axis, so you see the recent detail and the long trend in one picture (a reader's highlight of the passage). Because the format never changes, a reader's eye goes straight to what is different.

Who attends and how the meeting runs

Outside Amazon, I'd set it up like this (our recommendation, not Amazon's rulebook):

• Chair: the CEO or general manager.
• Deck owner: finance or business operations, who make sure the numbers are final before the meeting.
• Metric owners: one person per area (acquisition, sales, product, operations, support), who review their section beforehand and come ready to say what they're doing about a variance.
• Data team: on call for definitions, not presenting.

Keep it to an hour:

1. Last week's actions: done, slipped or dropped (5 minutes).
2. Scoreboard of output metrics (5 minutes).
3. Input metrics by area, with owners speaking only to exceptions (35 minutes).
4. Customer anecdotes and operational issues (5 minutes).
5. New actions, each with an owner and a date (10 minutes).

The meeting spots what changed and assigns it. Investigations happen afterwards and come back next week as an action update.

Weekly business review template

Copy this into a doc or a deck. For each metric I'd show four numbers next to the six-week and twelve-month chart: this week, week on week (WoW), vs plan, and vs the same week last year (LY).

Section	What goes in it	Comparisons
1. Header	Week number and dates, when the data was last loaded, a one-line summary of the week	None
2. Last week's actions	Each action, its owner, and its status: done, in progress or slipped	None
3. Scoreboard	6 to 10 output metrics, such as revenue, orders or new ARR, gross margin, active customers, retention	This week, WoW, vs plan, vs LY
4. Input metrics by area	3 to 6 per area: acquisition, conversion, product usage, operations, customer experience	This week, WoW, vs plan, vs LY; 6-week and 12-month trend
5. Exceptions	Every metric outside its band (say, more than 5% off plan, or three weeks moving the wrong way), with the owner's one-line explanation	Vs plan, vs LY
6. Customer voice	Two or three anecdotes, top support themes, notable accounts won or lost	None
7. New actions	What, who, by when	None

A starting point for sections 3 and 4:

Business	Output metrics	Controllable input metrics
B2B SaaS	New ARR, net revenue retention, gross margin	Qualified pipeline created, trial-to-paid rate, weekly active accounts, time to first value, support first-response time
E-commerce	Revenue, orders, contribution margin	Sessions by channel, conversion rate, in-stock rate, delivery time, return rate
Marketplace	Gross merchandise value, take rate, repeat buyer rate	New listings viewed, search-to-purchase rate, supply in the top demand areas, cancellation rate

Monthly business review template (MBR)

The MBR is where the weekly noise averages out and the numbers meet finance: books closed, actuals against plan, an updated outlook for the year, and the management reporting pack. Run it a few working days after month end, once the close is done.

1. Summary: three bullets on what went well, what didn't and what we're changing.
2. KPI scorecard: the output metrics for the month and year to date, month on month, vs plan and vs last year.
3. P&L vs plan: revenue, gross margin, operating expenses by function, EBITDA or burn.
4. Function pages: one page each for sales, marketing, product, operations and customer success, with 3 to 5 input metrics, their trend, and the top issue.
5. Cohorts and retention: customer or revenue retention by start month.
6. Forecast: the full-year forecast against plan, and what changed since last month.
7. Risks and decisions: the two or three decisions you need from the room.

Quarterly business review template (QBR)

This is the internal QBR. For the customer QBR account managers run, swap the financial pages for the customer's usage, outcomes against their goals, support history and next quarter's plan.

1. The quarter on one page: results vs plan, three wins, three misses.
2. Goals scorecard: each quarterly goal or OKR, its owner, status and score.
3. Financials: quarter on quarter, vs plan, vs last year, and the full-year outlook.
4. Input metric trends: the trailing 13 weeks and the trailing five quarters for each area.
5. What we learned: the bets that worked and the ones that didn't, with the evidence.
6. Next quarter: priorities, a target for every metric, and an owner for each.
7. Resource asks: headcount and budget changes.
8. Appendix: metric definitions and data notes.

Board report template (board pack outline)

A board pack is the QBR rewritten for people outside the business: fewer metrics, more context, and explicit asks. Send it far enough ahead that the meeting is spent on discussion, not reading.

1. CEO letter: one page with the headline, what's on and off track, and what you need from the board.
2. KPI dashboard: 8 to 12 metrics with the trailing four to eight quarters, vs budget and vs last year.
3. Financials: P&L vs budget, cash, burn and runway.
4. Go-to-market: pipeline, bookings, retention.
5. Product and operations: what shipped against the roadmap, and the key operational metrics.
6. People: headcount vs plan, key hires, attrition.
7. Strategic topics: one or two questions for discussion, each with a short pre-read.
8. Decisions and approvals: budget, option grants, policies.
9. Appendix: metric definitions and detailed financials.

How to automate a weekly business review from your warehouse

Most of a manual WBR is mechanical: pull the numbers, paste the charts, write "revenue up 4%, led by EMEA" twenty times. Get the numbers right in the warehouse first, then decide how much of the writing a machine should do.

The data model a weekly business review needs

• A date spine with your fiscal calendar. One row per day with the week start (pick Monday or Sunday and never change it), fiscal week, month, quarter and a holiday flag. Compare with last year by going back 52 weeks (364 days), not one calendar year, or a shifted weekend will distort every comparison.
• Facts at daily grain. Orders, sessions, signups, pipeline, tickets, so they roll up to any week, month or quarter.
• Shared dimensions. Region, channel, product and segment spelled the same way in every table.
• The plan, in the warehouse. A targets table at week, metric and segment grain. If the plan lives in a spreadsheet, "vs plan" is typed in by hand and is the first number to go stale.
• Metric definitions in one place. A semantic layer, such as dbt metrics or LookML, keeps "active customer" identical across the WBR, MBR and board pack, so any board number traces back to the weekly deck.
• Owners and freshness. A table that maps each metric to an owner, and a load timestamp, so the report only runs once the week's data is complete.

With those in place, the scoreboard is one query. Here it is in BigQuery SQL for revenue, returning the trailing six closed weeks with all three comparisons:

-- Trailing 6 closed weeks: revenue, WoW, vs plan, vs same week last year
WITH weekly AS (
  SELECT
    DATE_TRUNC(order_date, WEEK(MONDAY)) AS week_start,
    SUM(revenue) AS revenue
  FROM analytics.fct_orders_daily
  WHERE order_date >= DATE_SUB(DATE_TRUNC(CURRENT_DATE(), WEEK(MONDAY)), INTERVAL 60 WEEK)
    AND order_date < DATE_TRUNC(CURRENT_DATE(), WEEK(MONDAY))  -- closed weeks only
  GROUP BY week_start
)
SELECT
  w.week_start,
  w.revenue,
  SAFE_DIVIDE(w.revenue, prev.revenue) - 1 AS wow,
  SAFE_DIVIDE(w.revenue, p.target) - 1 AS vs_plan,
  SAFE_DIVIDE(w.revenue, ly.revenue) - 1 AS vs_ly
FROM weekly AS w
LEFT JOIN weekly AS prev ON prev.week_start = DATE_SUB(w.week_start, INTERVAL 1 WEEK)
LEFT JOIN weekly AS ly ON ly.week_start = DATE_SUB(w.week_start, INTERVAL 52 WEEK)  -- same weekdays
LEFT JOIN analytics.plan_weekly AS p ON p.week_start = w.week_start AND p.metric = 'revenue'
ORDER BY w.week_start DESC
LIMIT 6;

Swap the fact table and the metric name for every other scoreboard row. Change WEEK(MONDAY) to MONTH or QUARTER (and the offsets to match) and it feeds the MBR and QBR.

Dashboard subscriptions vs AI-written reviews

Your BI tool already has subscriptions. As of October 2026, Power BI emails a snapshot of a report or dashboard on the schedule you set, and outside Premium capacity each subscriber needs a Pro or Premium Per User license (Microsoft Learn). Looker schedules dashboards to email, webhooks, Amazon S3 and SFTP, and to Slack once an admin enables the integration (Google Cloud docs). The alternative is a review an AI writes from the same data.

	Dashboard subscription	AI-written review
What arrives	The dashboard as an image, a PDF or a link	The numbers plus a written account of what changed and why, as a message, PDF or deck
Explains why	No. Someone still writes the commentary	Yes, if the tool can investigate rather than only describe
Run to run	Identical layout every time	Same queries, different wording
Cost	Usually inside the BI license you already pay for	Paid per run, and a deeper run costs more
How it fails	Nobody opens it	A confident, wrong explanation when metric definitions are loose

A subscription is enough when the readers know the charts, the metrics move slowly, and the meeting supplies the story: a good dashboard subscription plus a shared doc for the owners' comments. Don't pay per run for words nobody needs.

An AI-written review earns its cost when the commentary is the bottleneck (an analyst spends hours every week explaining the moves), when people act on the review between meetings, or when the readers don't open dashboards. Fix your metric definitions first, because an AI explains whatever the data says, including a broken join, and check anything bound for the board.

How GetDot's Automated Reports produces a weekly business review

This is how our own product does it, as of October 2026 (from our Automated Reports page and changelog):

1. Build the scoreboard once. Ask Dot in the web app, Slack or Microsoft Teams: "revenue, orders and conversion for last week vs the prior week, plan and the same week last year, by region." Check the SQL it shows, then publish the answer as a dashboard if you want a fixed layout.
2. Schedule it to Slack, Teams or email, daily, every weekday, weekly or monthly, or type "schedule this weekly to me" in the chat.
3. Pick how much work each run does. Latest numbers refreshes the charts and tables for a fraction of a credit. Brief adds a short summary of what changed since last time. Fresh analysis has Dot re-investigate and write up new findings, which turns a scoreboard into a review that explains the moves.
4. Choose the format. A PDF that leads with the headline and key numbers, or, for the MBR or QBR, a PowerPoint deck on your uploaded template with native, editable charts, every slide checked for overlapping text, content off the slide and broken tables.
5. Send only when it matters. An "Only send when…" condition, such as "daily signups drop below 1,000", costs nothing on a quiet dashboard run when it can be checked against the numbers, and a cooldown stops repeats. A condition that needs judgment runs as a fresh analysis, billed as one even when it decides not to send. A work gate ("only run if new orders arrived today") skips a run when the data hasn't landed.
6. Reply to ask why. Recipients reply in the thread or to the email, and Dot answers with the report's context; a bigger "why did this drop?" gets a root cause analysis.
7. Keep it governed. Every SQL query carries a comment naming the user, the workspace and what triggered it; a scheduled job can run as a specific user with that user's group permissions; admins can restrict scheduled emails to approved domains.

House rules ("lead with the three biggest movers", "flag anything more than 5% off plan") go in a custom skill, a markdown playbook an admin writes once. Dot also sees previous runs, so it can comment on trends across weeks (more in Dot-Maxxing).

A four-week rollout

1. Week 1: list your output metrics and 10 to 20 inputs, name an owner for each, write the definitions down, and load the plan into the warehouse.
2. Week 2: build the scoreboard query or dashboard and check it against last week's manual deck, number by number.
3. Week 3: send the automated version alongside the manual one and note every difference.
4. Week 4: switch over. Owners keep writing their one-line explanations, now on top of a report that already exists.

Common mistakes

• Too many metrics. If a metric hasn't prompted a question or an action in a quarter, move it to the appendix.
• The review creates the work it should prevent. If every WBR triggers a round of one-off questions to the data team, see how to cut ad hoc data requests.

FAQ