Start with the decision on your desk
Sales are up. Everyone is busy. The report looks good. Before calling it a good week, I want to know what it took to produce those sales and what the business has left to show for them.
That’s where I’d start a weekly review: with a question that affects the business. Did the promotion earn its place? Are customers waiting too long for a response? Can we deliver the extra volume we’re trying to sell? Pick the question, then bring the numbers that help answer it.
I’d start with three to five measures. Keep the fuller reports available for investigation, but give the weekly conversation a focus. If you can’t explain what you would do differently when a number changes, ask whether it belongs in that conversation.
Agree on what you’re counting
Before debating performance, make sure everyone means the same thing by “sales.” An order placed, a completed job and a payment received can land on different dates. Mixing them makes a report hard to act on.
Shopify is a good example. Its net sales figure subtracts discounts and sales reversals from gross sales. Its sales reports don’t measure cash received from customers; payments need their own view. The right report depends on the question you’re asking. [1]
For each measure, write down the calculation, where it comes from, the dates it covers and who updates it. Compare complete periods on the same basis. If a figure is missing, label it pending. If a rate can’t be calculated because its denominator is zero, label it unavailable. A blank with an explanation is better than a misleading zero.
Look at what’s left after the sale
Consider this hypothetical ecommerce business. We’re comparing two full weeks. Net sales exclude sales tax and already account for discounts and sales reversals. The variable costs include all costs that vary with these sales, and fixed costs are unchanged.
In Week A, $10,000 in net sales minus $5,000 in variable costs leaves $5,000. In Week B, $12,000 minus $7,000 also leaves $5,000. Sales grew 20%, but there’s no additional contribution toward fixed costs. Contribution as a share of net sales fell from 50.0% to 41.7%.
Contribution means sales less variable costs. Fixed costs still have to be covered, so this isn’t net profit or cash flow. The SBA’s break-even guidance explains how these costs fit together. [2]
I’d want to see the product mix, discount levels and variable costs behind the second week before increasing the promotion budget. The numbers tell us where to look. They don’t yet tell us what caused the change.
| Measure | Week A | Week B |
|---|---|---|
| Net sales | $10,000 | $12,000 |
| Variable costs | $5,000 | $7,000 |
| Contribution | $5,000 | $5,000 |
| Contribution / sales | 50.0% | 41.7% |
Follow the number into the business
When a result moves, look at the steps that produced it. If a store’s revenue slips, examine transaction count and the value of those transactions. Then check availability, promotions and returns. If sales inquiries increase but bookings don’t, look at response times, follow-up and the reasons people gave for declining.
The same approach works beyond marketing. A restaurant can compare demand by daypart with staffing and wait times. A hotel can examine booking channels alongside commissions and cancellations. The question is where the result points you next, and what evidence would support changing something.
Set targets around your own margins, capacity and customer promises. An industry average can prompt a question, but I wouldn’t adopt it as a target without understanding how it was measured and whether the businesses are comparable.
Give AI a narrow job
AI can help prepare the review. Give it an approved table and ask it to list the largest changes, show its calculations and flag missing information. Ask for questions to investigate rather than an explanation of why the business performed a certain way.
For example: “Compare these two complete weeks. Show the dollar and percentage changes. Separate what the table demonstrates from what we would need to investigate. Don’t invent reasons for the changes.”
Check the output against the source before sharing it. Use only information you’re permitted to put into that tool, and include checking and corrections when judging whether it saved time. You still need someone who understands the business to decide what happens next.
Leave Monday with one decision
Give this twenty minutes at your next weekly meeting. Choose one result that deserves attention, confirm the figures, and decide what you need to find out before making a change.
Using the example above, the action might be: the ecommerce manager compares product mix, discounting and variable costs for both weeks by Thursday. The owner reviews that breakdown before deciding whether to extend the promotion. That’s specific enough to follow up on.
Write down the person responsible, the deadline and the next review date. If the figures can’t be reconciled, make fixing that the first action. Next week, start with what happened to the decision you made this week.
Bring it to your next review.
Use your own revenue, acquisition costs and customer count to work through CAC, MER and contribution. Then decide what your team should investigate.
Sources and context
- Shopify Help Center: Sales reports. Net sales definitions and the distinction between sales and payments.
- U.S. Small Business Administration: Plan your business. See the Break-even point section for fixed costs, variable costs and contribution margin.
Sources checked October 7, 2026. The figures are hypothetical. The review process reflects my approach; adapt it to your business.