Updated recently
Last updated:
Quick Answer: what every CEO should measure weekly: Six numbers, reviewed in the same order every week: cash (collected, not invoiced), revenue quality (what actually landed, and from whom), pipeline (weighted, with a commit definition), retention (churn and repeat rate), delivery (time to deliver and quality), and one diagnostic number — revenue per person or contribution margin per customer, whichever is more binding. Everything else can wait for monthly.
Key takeaways
- Review numbers in a fixed order with a fixed format. A weekly review that changes shape each week produces conversation rather than decisions.
- Cash means collected, not invoiced. This is the distinction that prevents the most expensive surprise in business.
- Revenue quality beats revenue. A month with 20 percent revenue growth that came from one customer who will not renew is a worse month than flat revenue from a healthy base.
- Pipeline needs a commit definition or the number is an opinion with a decimal point.
- The diagnostic number explains the rest. Revenue per person or margin per customer usually accounts for whatever else is moving.
What this guide covers
- The six numbers, why each one, and what a good trend looks like
- Cash: the definition that prevents surprises
- Revenue quality: what actually landed and who paid
- Pipeline: weighted, with a commit definition
- Retention and delivery: the two operational numbers
- The weekly review format that turns numbers into decisions
Six numbers, fixed order, same format
Most weekly executive reviews are long and produce nothing. The problem is usually that they mix levels — strategic, operational and diagnostic — into one conversation, so a delivery problem gets discussed as a strategy problem and no decision is made.
The fix is a small, fixed set of numbers reviewed in the same order every week, with the same format, in about forty-five minutes. Six is enough to cover a healthy business. The order below is deliberate: money first, then demand, then quality of what exists, then the operations that deliver it.
1. Cash
What to look at: collected cash in, total cash out, and the resulting balance — with a thirteen-week rolling forward.
Why it is first: because it is the only number that cannot be revised. Everything else — revenue, pipeline, forecast — can be adjusted, argued about or re-estimated. Cash is what the bank says.
What a good trend looks like: a balance that covers at least the period you would need to fix a bad quarter, with the pessimistic case also funded. That is the buffer that allows you to make decisions calmly.
The trap: counting invoiced revenue. A business can be profitable on paper and insolvent, because growth consumes cash before it produces it and payment terms stretch precisely when buyers get cautious. The mechanism is covered in what nobody tells you about growing a business.
2. Revenue quality
What to look at: what actually landed this month — recognised or collected, depending on your model — and a one-line characterisation of where it came from.
The distinction between revenue and revenue quality is where most weekly reviews should spend their time. A month with 20 percent revenue growth that came entirely from one customer who has signalled they will not renew is a worse month than flat revenue from a stable base — and the report will show 20 percent growth unless somebody asks.
Three questions to attach: Is this recurring or one-off? Is it from customers we already have or new ones? Does any single customer account for a large share? Those three answers turn a number into an assessment, and the assessment is what should drive the week’s decisions.
3. Pipeline
What to look at: total pipeline value by stage, weighted by your own historical conversion rates, plus a separate committed figure.
Why weighted and raw both: the raw number tells you how much opportunity exists, the weighted number tells you what is likely, and the gap between them tells you where the confidence is thin. A large gap is usually a sign that early-stage deals are being valued optimistically.
Why the commit definition matters: “committed” is the number most likely to be wrong, because it is usually a mood. A committed deal has an approver who has agreed, a date, and paperwork in motion. Without that definition, the forecast drifts upward every week, which is why forecasting needs its own discipline — covered in sales forecasting made simple.
The weekly question: which deals moved, which stalled, and which moved backwards. The third category is the one worth attention, and it is the one nobody reports.
4. Retention
What to look at: churn and repeat rate, tracked monthly but reviewed weekly for direction, plus the count of at-risk customers.
Why it belongs here and not in a monthly review: churn is slow to appear and expensive to fix, so a week of delay in noticing it costs a month of churn. The leading indicators are what you want weekly — customers whose engagement has dropped against their own baseline, and customers who have gone quiet ahead of a renewal.
What a good trend looks like: flat or improving, with the at-risk list shrinking. A rising churn number is the single most important signal on this page, because it arrives in the revenue line eight weeks later as an unavoidable drop.
5. Delivery
What to look at: two numbers — time to deliver, and quality as measured by complaints, rework or a small customer-facing score.
Why it matters for a CEO specifically: the other five numbers are lagging or semi-lagging, and delivery is the one where a CEO can act before the damage shows up in revenue. Quality slipping at volume is the leading indicator of a future churn wave, and it is visible only if someone is looking at delivery rather than at the bank balance.
The pairing matters: time to deliver rising alongside quality holding usually means capacity is short. Time to deliver stable with quality falling usually means people are working around a broken process.
6. The diagnostic number
One number that usually explains the others: revenue per person, or contribution margin per customer, depending on where the business is.
If revenue per person is falling while revenue grows, the business is adding complexity faster than output — which is the classic symptom of premature scaling. If contribution margin per customer is falling, the business is selling more of something it should not be selling. If both are healthy and revenue is still not growing, then demand is the constraint and the other five numbers were never the problem.
The reason this number earns a slot in the weekly review: it is the shortest explanation of what is happening in the other five, and a CEO who only knows one number should know this one.
| Number | Tells you | If it is moving the wrong way, look at |
|---|---|---|
| Cash | Whether the business survives | Collections, payment terms, spending |
| Revenue quality | Whether last month’s number repeats | Customer concentration, one-off deals |
| Pipeline | Whether next quarter is funded | Stage conversion, dwell time |
| Retention | Whether the base is growing or shrinking | Onboarding, service, engagement drops |
| Delivery | Whether quality is holding at volume | Process, capacity, ownership |
| Diagnostic | What explains the rest | Hiring, mix, pricing, demand |
The weekly review format that produces decisions
Forty-five minutes, same order every week, same three questions per number.
1. What is the number? Stated, with the trend, and nothing else. No commentary yet.
2. What changed, and why? One or two sentences. If nobody knows why, that is itself the finding — and it usually means a number is not being managed.
3. What will we do this week? A named action with a date. If there is no action, the number is not worth reviewing weekly.
Two rules that make the discipline hold. First, one number gets the week’s attention — the one that is furthest from where it should be. Not all six, because six is a briefing. Second, the review happens at the same time every week and is not cancelled; a cancelled review becomes a monthly review, which is a different and much weaker instrument.
Two things to keep out. Opinions about the market, which cannot be measured weekly and produce speculation. And any number that has not been reviewed for three months — if a metric has not been looked at in a quarter, it has lost its meaning, and it should be dropped rather than kept for the reassurance of having it.
The measure of a good weekly review is not how much it covered. It is how many decisions came out of it, and whether those decisions were made early enough to matter.
Related reading
- Sales Forecasting Made Simple
- What Nobody Tells You About Growing a Business
- How Data Can Improve Your Sales Results
Frequently asked questions
What are the most important metrics for a CEO to review weekly?
Six, in this order: cash collected, revenue quality (what actually landed and from whom), weighted pipeline with a separate committed figure, retention and churn direction, delivery time and quality, and one diagnostic number — revenue per person or contribution margin per customer. Everything else belongs in a monthly or quarterly review. The ordering matters because cash is the only number that cannot be revised, and revenue quality prevents a misleading headline number from setting the agenda for the week.
Why does revenue quality matter more than revenue?
Because the headline number can be misleading for a quarter or more before the difference shows up. A month with 20 percent growth from one customer who will not renew is worse than flat revenue from a stable base, and the report will show 20 percent growth unless somebody asks whether the revenue was recurring, whether it came from existing or new customers, and whether any single customer is a large share. Three questions turn a number into an assessment, and the assessment is what should drive decisions.
How often should a CEO look at sales pipeline numbers?
Weekly, in the same review and in the same place, because pipeline problems are cheap to fix early and expensive later. Look at total pipeline by stage, a weighted figure using your own historical conversion rates, and a separate committed figure defined strictly — approver agreed, date set, paperwork moving. Also watch which deals moved backwards, which is the category nobody reports and usually the most informative. The mechanics of making these numbers reliable are in sales forecasting made simple.
What is the best single metric for a small business?
Revenue per person if the business is people-dependent, contribution margin per customer if it sells a mix of products or segments. Both are diagnostic because they usually explain what the other numbers are doing. Falling revenue per person alongside rising revenue means the business is adding complexity faster than output — the classic symptom of premature scaling. Falling contribution margin means the business is selling more of something it should not be selling. If both are healthy and revenue is still flat, demand is the constraint and nothing else was the problem.
Should a weekly review include revenue forecasting?
Yes, but the forecast should be a brief output of the review rather than the review itself — committed, weighted and best case, each with one named risk. The purpose of the weekly meeting is to change something, and a forecast discussion tends to consume the time without producing an action. Present the forecast in five minutes, then spend the remaining forty on the numbers that explain it and the decisions that follow. The approach is developed in how data can improve your sales results.
What if the numbers are bad every week?
Look at the diagnostic number first, because it usually identifies whether the problem is structural or situational. If revenue per person or margin per customer is falling, the business is operating badly and the weekly review will not fix it — that needs a change to the model, the pricing or the team. If the diagnostic is healthy and cash is the problem, the issue is timing rather than performance, and it needs working-capital decisions. If the diagnostic is healthy and revenue is flat while pipeline looks reasonable, the constraint is in the conversion between pipeline and revenue, which is where the detail is in how data can improve your sales results.
