The weekly business review

A single week's number tells you almost nothing - it could be noise, a holiday, a one-off. The review always shows a metric two ways: trailing-twelve-weeks so you see the recent trend, and year-over-year so you see it against the same week last year and strip out seasonality. A number that looks alarming in isolation often looks perfectly normal against last December. Context is not a nicety here; it is what makes the read trustworthy.

• Trailing 12 weeks shows the shape of the recent trend, not a single dot.
• Year-over-year controls for seasonality - retail in December, B2B in summer.
• Both together let you separate a real deviation from a predictable rhythm before anyone panics.

The most important rule of a weekly review is that the person who owns a branch speaks to it. The leader does not narrate; the leader listens and asks. When conversion moved, the product owner explains it. When churn moved, the retention owner explains it. This does two things: it puts the explanation in the hands of the person closest to the number, and it makes ownership real - you cannot own a driver you never have to explain.

• Every branch has one owner who narrates when it moves.
• Silence on a moved branch is a finding - it means the driver is unowned or unwatched.
• The leader's job is the question, not the answer - which driver, which segment, versus what baseline.

Amazon made the weekly business review famous: a metrics-first meeting where the deck is mostly charts, the tone is direct, and the discussion follows the numbers rather than a narrative someone wanted to tell. The genius is not the format, it is the frequency. By reading the trees every week against trailing and year-over-year context, deviations surface while they are cheap. Many operating companies have since adopted the rhythm because it converts metric decomposition from a one-time diagram into a living habit.

The slide tells the story someone wanted to tell. The tree shows the story the numbers are actually telling. Start from the top-line metric, walk down the branch that moved, and let the tree route the conversation - not a pre-written narrative. When a metric moves, ask the same three-part question every single time: which driver, which segment, versus what baseline? That question is the whole method, applied out loud.

• Which driver - use the tree to name the lever that moved.
• Which segment - split that driver to find where: which channel, region, cohort.
• Versus what baseline - is this a real deviation from trend and seasonality, or noise?

Not every wiggle deserves a discussion. A metric that bounces inside its normal range is noise; a metric that steps outside it is signal. Good reviews use simple control limits and known seasonality to decide which is which, so the hour is spent on the two branches that genuinely moved rather than the twenty that jiggled. Chasing noise is as costly as missing signal - it trains the room to ignore the dashboard.

Most failed reviews fail the same four ways. Name them out loud in your own review and they lose their grip.

• Status theatre - everyone reports that things are fine, nobody reads the tree, the hour produces no decision.
• No baselines - numbers shown as single dots with no trailing trend or year-over-year, so nobody can tell signal from noise.
• No owner - a branch moves and the room looks around; if nobody speaks, nobody owns it.
• Admiring the problem - twenty minutes discussing why a metric is down and zero minutes deciding what to do, with whom, by when.

These two cadences are not competitors - they are a pipeline. The weekly review generates a running record of what moved, what was decided, and what is still open. The monthly summit consumes that record and rolls it up: instead of re-litigating every driver, it tells leadership the strategic story - are we on plan, what moved and why, what we are doing about it. If the weekly is done well, the monthly writes itself.

• Weekly - driver-level, operational, fast; catch and correct deviations.
• Monthly - top-line, strategic, narrative; roll the weekly detail into a decision-grade story.
• The handoff - a clean weekly cadence is the raw material for a sharp monthly summit. Skip the weekly and the monthly becomes archaeology.

A useful rule: review a metric on the cadence at which it can actually move and be acted on. Fast, driver-level leaves - traffic, add-to-cart, pipeline coverage - reward a weekly look because you can act on them within days. Slow, strategic roots - market share, ARR trajectory, program outcomes - suit a monthly altitude because a weekly wiggle is mostly noise. Watching a slow metric weekly breeds panic; watching a fast one monthly breeds blindness.