Analytics
The weekly ecommerce report: eight numbers that matter
A one-page weekly report for an online store: the eight figures worth reviewing, where each one comes from, and the question each is meant to answer.
By CartKernel · Published
A weekly store report should fit on one page and take under ten minutes to read. Most do not, because they were assembled from whatever the tools export rather than from the decisions the week actually needs. Eight figures are enough to run an online store week to week. Everything else is a drill-down you open when one of the eight moves.
Here is the set, what each one is for, and where to get it without building a warehouse first.
The one page
| # | Figure | Source of record | The question it answers |
|---|---|---|---|
| 1 | Net revenue | Store platform | Did the business grow this week |
| 2 | Orders | Store platform | Was growth volume or price |
| 3 | Sessions | Analytics | Did demand arrive |
| 4 | Conversion rate | Orders divided by sessions | Did the site turn demand into sales |
| 5 | Average order value | Net revenue divided by orders | Did basket size hold |
| 6 | Total marketing spend and efficiency ratio | Ad platforms plus store revenue | Is growth being bought at a sustainable price |
| 7 | New customer share of revenue | Store platform | Is the customer base widening or leaning on regulars |
| 8 | Contribution after ads | Revenue, cost of goods, fees, spend | Did the week make money |
Two rules make the page trustworthy. Every figure comes from one named source, written on the report, and it never changes source. And every figure carries the same week last year alongside last week, because a store with any seasonality cannot be read week over week alone.
1. Net revenue, not gross
Net revenue means after discounts and after refunds, in the store’s own currency. Gross merchandise value flatters a discount-heavy week and hides a refund problem entirely.
Take it from the platform, not the analytics tool. The store is the record of what was sold. Analytics is a sample of what was measured, and the two will differ every week for reasons that have nothing to do with sales.
Where the week’s revenue is up, the next three lines say why.
2. Orders
Orders separate a busy week from an expensive week. Revenue up with orders flat means average order value did the work, which is usually a promotion, a bundle or a mix shift toward higher-priced products. Revenue up with orders up means more customers bought.
Exclude point of sale, wholesale and draft orders unless the report is meant to cover them, and say on the page which channels are included. This is the single most common reason two people quote different numbers in the same meeting.
3. Sessions
Sessions are the demand that arrived. Take them from analytics, split into a small number of channel groups you actually manage: organic search, paid search and shopping, paid social, email and SMS, direct, referral.
Read sessions with the caveat that analytics undercounts. Consent choices, tracking prevention and blocked scripts remove a share of traffic from the report every week, and that share is not stable. What matters here is the direction and the mix, not the absolute count. If direct traffic swells with no explanation, why direct traffic is so high in GA4 covers the usual causes.
4. Conversion rate
Compute it as store orders divided by analytics sessions, and keep that definition forever. It mixes two sources, which is not ideal in theory and is the most useful version in practice, because it uses the reliable numerator and the directionally useful denominator.
A conversion rate that falls while sessions rise usually means the traffic mix changed rather than the site got worse. Check which channel grew before touching the product page. A conversion rate that falls while sessions are flat is a site or checkout issue, and conversion rate dropped suddenly is the triage list. Segment by device before anything else, because mobile and desktop rates move independently.
5. Average order value
Net revenue divided by orders. Track it because it is one of the four levers in the growth model, and because it is the fastest to move: bundles, thresholds, cross-sells and price changes all land within a week.
Watch it against the free shipping threshold. If a large share of orders cluster just above the threshold, the threshold is doing its job. If they cluster just below, it is set too high for the catalog.
6. Marketing spend and efficiency ratio
Add every dollar of media across every platform, then divide total store revenue by total spend. That ratio is the store’s marketing efficiency ratio, and it is the only marketing number on the page.
Platform-reported return on ad spend does not belong in a weekly report. Every platform claims conversions the others also claim, and the sum of the claims exceeds the store’s revenue. Reported return is a campaign management tool, useful inside the account, misleading at the top of a report. The MER calculator gives the figure, and why Google Ads reports more conversions than my store explains the arithmetic behind the double counting.
Show spend as its own line as well as inside the ratio. A ratio can hold steady while both sides double, and that is a different business.
7. New customer share of revenue
The share of the week’s revenue that came from people who had never ordered before. Most platforms expose this directly; where yours does not, first-order flags on the customer record will produce it.
This is the growth quality number. A week where revenue held up entirely on repeat buyers looks fine and is not fine, because the repeat base is finite and shrinks without new entrants. A week where new customer share spikes alongside a spend increase tells you what the money bought. Read it beside repeat purchase rate monthly rather than weekly, since repeat behaviour plays out over a longer window than seven days.
8. Contribution after ads
Net revenue, minus cost of goods, minus payment and fulfilment costs, minus media spend. This is the number that decides whether the week was worth having.
You do not need perfect costs. A blended cost of goods percentage, a payment fee percentage and a per-order fulfilment estimate get you within a few points, which is enough to see the trend. Precision can come later; the habit of putting a profit line on the weekly page should not wait for it.
Where to build it
For most stores, a spreadsheet with three tabs and a weekly paste is honest and takes twenty minutes. Automate it once the routine is fixed, not before, because the definitions are what take time to settle, not the plumbing. Where you want it live, a dashboard tool connected to analytics and the store covers seven of the eight; contribution usually still needs a cost table you maintain. Is Looker Studio enough for ecommerce reporting covers the point at which a warehouse starts paying for itself.
Whatever the tool, the figures should be locked after the week closes. A live dashboard that keeps restating last week’s numbers as attribution windows fill makes it impossible to remember what you decided and why.
Three drill-downs to keep beside the page
You will not need these every week, and you will need them the week something moves.
- Channel detail. Sessions, orders and revenue by channel group, so line three and line four can be explained.
- Product movement. Top twenty products by revenue against the prior week, plus anything that went out of stock. Stock-outs explain more sudden revenue drops than any marketing change.
- Search visibility. Clicks and impressions by page group from Search Console, monthly rather than weekly, because organic movement over seven days is mostly noise.
What to leave off
Bounce rate, time on page, impressions, follower counts, email open rates and any metric that cannot change a decision. Every extra line lowers the chance the eight that matter get read. If a figure has never once prompted an action, it belongs in a drill-down or nowhere.
Set the report to arrive the same morning each week, before any meeting that discusses it, and keep the archive. Twelve weeks of one-page reports is a better record of a store than any dashboard, because it shows what you knew at the time. The wider set-up work, from event tracking to a source of record for every line, is what GA4 ecommerce tracking and attribution reporting are for.