Most Amazon sellers check their numbers constantly and understand them poorly. They open the app at breakfast, glance at yesterday’s sales, feel briefly good or briefly anxious, and close it again. That’s monitoring, not measurement. It produces a lot of emotion and almost no decisions.
The fix is boring and it works: one fixed report, same metrics, same day, every week. Not because weekly data is magic, but because it’s the shortest interval at which Amazon’s numbers are stable enough to act on and the longest you can go without a problem compounding out of control.
Here’s the structure worth copying, what each number actually tells you, and how to stop assembling it by hand.
Why Weekly Beats Daily and Monthly
Contents
Daily data on most accounts is noise. A single ASIN doing 20 units a day will swing 40% between a Tuesday and a Saturday for no reason at all. Change your PPC bids on three days of data and you’re not optimizing, you’re chasing randomness — and you’ll do it again three days later in the opposite direction.
There’s also an attribution lag. Amazon’s advertising data continues to settle for several days after the fact as sales get attributed back to the click that caused them. Yesterday’s ACoS is not yesterday’s ACoS. It’s a draft.
Monthly reporting has the opposite problem. A stockout that starts on the 4th and gets noticed on the 30th has already cost you the ranking you spent six months building — and organic position doesn’t come back when the inventory does, not at the same rate.
Seven days is where the tradeoff lands. Enough data volume to see a real signal, short enough that nothing rots.
The Five Blocks
A useful weekly report has five sections and fits on one screen. If it runs longer than that, you’ve built an archive, not a decision tool.
Block 1 — Revenue and Profitability
- Gross sales and units sold
- Average selling price (gross sales ÷ units)
- Refund rate as a percentage of units
- Net margin after fees — referral fee, FBA fulfillment fee, storage, ad spend, COGS, freight
That last one is the whole point, and it’s the one most sellers skip because Seller Central won’t hand it to you assembled. Revenue is the number that feels like the business. Margin after everything is the number that is the business. Plenty of accounts have grown revenue 30% year over year while net profit went sideways, because fee increases, rising CPCs, and a creeping refund rate ate the difference quietly.
Track average selling price separately from revenue. If revenue is flat but ASP is falling, you’re discounting to hold volume, and that’s a different problem with a different fix than a traffic decline.
Block 2 — Traffic and Conversion
Pull these from Seller Central > Reports > Business Reports > Detail Page Sales and Traffic by Child Item:
- Sessions — unique visits to your detail page
- Page views
- Unit Session Percentage — your conversion rate, and the most diagnostic single number on the page
- Featured Offer percentage (the metric formerly labeled Buy Box percentage)
The diagnostic logic is simple. Sales down, sessions down, conversion flat: that’s a traffic problem — ranking, ads, or a competitor outbidding you. Sales down, sessions flat, conversion down: that’s a listing problem — a new negative review, a price change, a competitor undercutting you, a suppressed image, or a lost Featured Offer.
Those two situations look identical on a revenue chart and have nothing in common as problems. Sellers who only watch revenue guess. Sellers who watch conversion know.
Featured Offer percentage deserves its own line because a drop can be invisible for weeks. If you’re sharing a listing or you’ve been price-matched, you can be paying for ads that send traffic to a page where someone else gets the sale.
Block 3 — Advertising
- Ad spend, ad sales, ACoS, ROAS
- TACoS — total ad spend ÷ total sales, including organic
- Search terms harvested this week (converting terms promoted to exact-match)
- Negative keywords added
- Top 5 spend-no-sale search terms
ACoS alone is a trap. You can drive ACoS down by turning campaigns off, and your report will look like an improvement right up until organic rank decays and total sales follow.
TACoS is the honest number. If TACoS is trending down while total revenue climbs, your organic engine is getting stronger and advertising is doing its actual job — buying velocity that converts into rank. If TACoS is climbing while revenue is flat, you’re renting sales, and the rent is going up.
The two action lines — terms harvested, negatives added — matter more than the ratios. They’re the only part of the report that records what you did rather than what happened to you. A week with zero of both is a week nobody managed the account.
Block 4 — Inventory and Supply Chain
- Days of cover per SKU at current velocity
- Sell-through rate
- Capacity utilization — how much of your allotted FBA capacity you’re using
- Inbound shipments and their status
- Stranded and unfulfillable units
Days of cover is the one that prevents disasters, and it has to be read against your total lead time — production plus freight plus Amazon receiving, which can stretch well past 90 days on an ocean shipment from Asia. If a SKU shows 60 days of cover and your real replenishment cycle is 100 days, you are already late. You just don’t feel it yet.
On storage: Amazon’s Inventory Performance Index still exists and still influences what you’re allowed to send in, but the threshold has moved several times over the years, and the constraint that actually binds most accounts now is the FBA capacity limit shown in the Capacity Monitor, measured in cubic feet. Check the monitor directly rather than trusting a number you read in a blog post two years ago — including this one.
Stranded inventory is the quiet money-burner. Units sitting in a fulfillment center attached to a listing that’s gone inactive earn nothing and accrue storage fees. It’s a two-minute check that gets skipped for months.
Block 5 — Account Health and Compliance
- Account Health Rating
- Order Defect Rate — Amazon’s target is under 1%
- Late Shipment Rate (under 4%) and Pre-Fulfillment Cancel Rate (under 2.5%) for seller-fulfilled orders
- Valid Tracking Rate and On-Time Delivery Rate
- Policy violations, suppressed listings, open cases
If you’re pure FBA, several of these barely move, since Amazon carries responsibility for fulfillment defects on orders it ships. That’s exactly why the block gets skipped — and why a policy violation on a single ASIN can sit unread for three weeks.
The Account Health Rating is a score, not a pass/fail. Watching the trajectory matters more than the absolute number: a rating drifting downward over four weeks is telling you something well before it turns into a deactivation email at 2 a.m.
Open cases belong in the report with an age count. A case opened 19 days ago that nobody has followed up on is a real cost sitting in a queue.
Three Columns, Not One
Every metric gets three columns: this week, last week, and a four-week average.
A single number is meaningless. “ACoS 24%” tells you nothing. “ACoS 24%, up from 19% last week, four-week average 20%” tells you something changed and roughly when. The four-week average is what stops you from panicking over a holiday week or a Prime Day distortion.
Add a fourth column if you can: year-over-year for the same week. Amazon is deeply seasonal, and comparing December to November will make a good month look like a collapse.
The Part Everyone Skips: The Decision Log
At the bottom of the report, three lines:
- What changed this week and why we think it changed
- What we’re doing about it — with a name and a date attached
- What we changed last week, and whether it worked
Without this, you have a data archive. With it, you have a management system. Point three is what turns reporting into learning — it forces you to check whether last week’s fix actually did anything, which is a question almost nobody goes back and asks.
Keep an annotation log alongside it: price changes, new images, coupon starts, campaign launches, competitor price drops. Three months later, when you’re staring at a conversion rate that fell off a cliff in week 7, that log is the difference between knowing the answer and inventing one.
Who Should Actually Own This Report
Somebody’s name has to be on it, and it can’t be “whoever has time.”
For a single-SKU seller, it’s you, and it’s forty minutes on a Monday. Past roughly 20 SKUs, or once you’re running meaningful ad spend, the assembly work alone becomes a part-time job — and it competes directly with the work that actually grows the account.
This is where the build-versus-outsource question usually surfaces. An in-house Amazon manager is a real salary plus benefits before they’ve optimized a single campaign. The agency route trades that for a monthly retainer, typically with a percentage of ad spend layered on top. Both are legitimate; the deciding factor is usually catalog complexity and how much of your own time the account is currently eating.
If you go the outsourced route, fixed weekly reporting should be a contractual deliverable, not a favor — and you should see the actual template before you sign anything. ZonHack’s Amazon account management team publishes downloadable samples of the weekly report and dashboard it sends clients, which is a reasonable structure to benchmark against whether you hire anyone or build your own. The test for any provider is simple: if they can’t show you a sample report before onboarding, they don’t have one.
Automating the Assembly
Copying numbers between Seller Central tabs by hand is fine for a month and unsustainable after that. Three levels of automation, in ascending order of effort:
Level 1 — Scheduled exports into a spreadsheet. Seller Central lets you export Business Reports and advertising reports as CSVs. Drop them into a Google Sheet on a fixed schedule and build the calculated fields — TACoS, margin after fees, days of cover — as formulas that reference the raw import tabs. Ninety minutes to set up, then five minutes a week.
Level 2 — API-driven pulls. Amazon’s Selling Partner API and Advertising API expose the same data programmatically. A scheduled script can pull sales, traffic, and ad data into a sheet or database automatically, no exports involved. This needs developer time and API credentials, but it removes the manual step entirely and lets you keep unlimited history.
Level 3 — Off-the-shelf analytics tools. Plenty of platforms will assemble most of this for you. They’re the fastest path to a dashboard, and the tradeoff is that you get their opinion of what matters rather than yours — and profitability figures are only as accurate as the COGS and freight data you feed in.
Level 1 is right for most sellers. The main reason to move up is history: dashboards that recalculate on the fly will happily show you the last 90 days and quietly lose the annotated record of what you did and why.
A Quick-Start Checklist
- Pick a day. Monday morning or Friday afternoon — consistency matters more than which.
- Build the five blocks in a single sheet, one tab for raw exports, one for the summary view.
- Add the three comparison columns before you add a single extra metric.
- Write the decision log every week, even when the answer is “nothing changed.”
- Review the report structure itself once a quarter. Metrics you’ve never once acted on should be deleted.
The report isn’t the goal. A sheet nobody reads is worse than no sheet, because it creates the feeling of control without any of the substance. The goal is that every Monday, someone looks at five blocks of numbers and decides three things — and that a month later, someone checks whether those three things worked.
That loop, run consistently for a year, beats almost any tool you could buy.