An advertising report should help you decide what to do next. A page of green arrows cannot tell you whether a campaign earned its budget, whether product costs changed, or whether a profitable month depended on one unusually strong ASIN.
In his reporting breakdowns, Alfredo Roselli starts outside the campaign dashboard. He looks at the business trend, then connects advertising to units, revenue and product economics. This guide follows that sequence. Use it with an agency report, a spreadsheet or a connected brand review. The goal is the same: understand the evidence and choose the next action.
A weekly review in five steps
- Identify a finding. Start with a change that matters to the brand’s goal.
- Inspect the evidence. Match the brand, product, marketplace and period. Keep source dates visible.
- Ask with context. Include the finding, supporting figures and the saved brand goal in the investigation.
- Decide. Name the person who will check or approve the next action.
- Review. Return to the decision on a set date and record what changed.
Keep these steps in the report. A reader should be able to trace a recommendation back to its evidence without opening every export.
Agree on the question before reviewing the numbers
Write down the account's current objective. Are you protecting contribution, introducing a product, or increasing sales within a spending limit? The same increase in advertising cost can be acceptable for a planned launch and unacceptable for an established product whose sales are flat.
Give the report a consistent date range, marketplace, currency and product scope. Record when it was exported and whether recent advertising conversions are still developing. Amazon explains that attribution metrics remain incomplete until the lookback window ends. Use the attribution window for the actual report and ad type; do not assume one window applies to every campaign. Compare like periods and add seasonal context. A holiday week and the following January week are not a clean performance experiment.
Separate the report period from its freshness
Record three dates: the period covered, the export or collection time, and the last date for which the source has usable data. A report collected today can still describe an earlier week. An inventory snapshot is a point-in-time observation, not an average for that week.
Rivet Rocket’s documented refresh runs daily at 9 a.m. Pacific. Amazon reports can arrive two to four days later, and advertising conversions can continue to develop within their attribution window. A daily refresh does not make every source current. Check the dates shown with the evidence before acting.
Keep sources separate: Seller Central provides seller-account evidence; Amazon advertising provides ad-report evidence. Public marketplace or search observations describe what was collected at a particular time. Public observations do not reveal another seller’s private sales, and public ranking is different from a completed weekly search-performance report.
Start with sales, units and the selling price
Read revenue alongside units and orders. Revenue can rise because more people bought, because prices increased, or because shoppers bought a different mix of products. Each explanation changes the advertising decision.
For example, $30,000 from 1,000 units gives $30 revenue per unit. Next period, $33,000 from the same 1,000 units gives $33 per unit. Revenue grew 10%, but unit volume did not. Check prices and product mix before describing the change as stronger demand. Revenue divided by units is average selling price per unit; revenue divided by orders is average order value. Label them separately.
Connect advertising efficiency to the business
| Measure | Question it helps answer |
|---|---|
| Ad spend | How much did we invest, and where? |
| ACOS | How does spend compare with attributed ad sales? |
| TACOS | How does spend compare with total sales? |
| Contribution after ads | What remains after the variable costs included in our model? |
| Ad-attributed sales share | Is more of our reported revenue receiving advertising credit? |
Use product-specific ACOS targets rather than one number for the whole catalog. Then ask whether total contribution dollars and contribution per order support the campaign result. A lower ACOS with fewer profitable orders can still be an undesirable outcome.
Read Amazon ads attribution before reconciling advertising and sales exports. Different reporting scopes and crediting rules can create apparent discrepancies. A rise in attributed revenue is not, on its own, evidence that advertising created that much additional demand.
Investigate the cost lines that changed
In Alfredo's P&L walkthrough, refund costs remained relatively steady while revenue fell after the holiday period. That made refunds a larger share of current revenue. The useful finding was in the relationship between the lines, not the advertising ratio alone.
Make a short bridge from last period's contribution to this period's. Separate changes in selling price, units, product mix, cost of goods, fees, discounts, refunds and advertising. Check when costs posted before blaming a new campaign. Refunds recorded this month can relate to earlier sales; the exact treatment depends on your reporting method.
Keep gross and net revenue definitions explicit. If refunds or discounts already reduce your revenue line, do not subtract them a second time. Use the same treatment in both comparison periods.
Look below the portfolio average
Start with the products responsible for most contribution and the products with the largest changes. Do not automatically begin with the campaign that has the worst-looking percentage. A tiny campaign and a major product decline have different financial consequences.
Add three monthly checks to your weekly review: advertising dependency, catalog concentration, and cost and margin drift. Together they test whether the headline result hides rising dependence on ads, fewer dependable products, or higher selling costs.
Build a cross-brand report without hiding the details
Use one row per brand and marketplace, with a consistent period and currency. Keep the report source and export timestamp beside the metrics. If currencies or attribution scopes differ, keep the rows separate until you have an explicit conversion or reconciliation method.
Amazon defines ACOS as ad spend divided by attributed ad sales, multiplied by 100. Define TACOS as ad spend divided by total sales on the revenue basis you specify, multiplied by 100. Show an undefined ratio as “not available” when its denominator is zero. Contribution after ads subtracts the variable costs included in your model and ad spend from the specified revenue basis; list the costs rather than implying it is net profit.
| Brand | Total sales | Ad spend | Attributed ad sales | ACOS | TACOS |
|---|---|---|---|---|---|
| Brand A | $20,000 | $2,000 | $8,000 | 25% | 10% |
| Brand B | $10,000 | $1,500 | $5,000 | 30% | 15% |
| Combined, same scope | $30,000 | $3,500 | $13,000 | 26.92% | 11.67% |
Compute combined ratios from summed numerators and denominators. Averaging 25% and 30% would give 27.5%, which does not describe this combined spend and attributed sales. Keep brand-level rows visible so a larger account does not hide a smaller account’s problem.
Work through a dated seller finding
Illustrative example: the following figures, brand goal, dates and explanation are invented. They show a review method, not live account data or a customer result. Both sales rows cover one SKU in the US marketplace, in US dollars, with the same reporting definitions.
| Evidence | Period or snapshot | Finding and limitation |
|---|---|---|
| Seller Central sales | September 28–October 4: $2,000; October 5–11: $1,800 | Total sales fell 10%. Check units, price and product mix before explaining why. |
| Advertising report | Same weeks: spend $200 → $250; attributed sales $800 → $750 | ACOS 25% → 33.33%; TACOS 10% → 13.89%. Recent attribution may still change. |
| FBA inventory evidence | October 11 snapshot: 120 available units; planning rate 10 units/day | Simple cover is 12 days at that rate. Recheck current available stock and recent sales velocity. |
| Seller’s inbound estimate | Recorded October 11: expected arrival October 31 | 20 days from the snapshot, leaving an estimated 8-day gap. Arrival and availability are unconfirmed. |
The finding is that advertising efficiency weakened while the stock plan shows a possible gap. The figures do not prove that inventory caused the advertising change. Check conversion maturity, units, prices and timing before deciding whether to change spending.
Ask with the finding and brand goal
For this example, the saved brand goal is “protect contribution and avoid running out of stock.” Rivet Rocket’s contextual Ask AI carries the finding being viewed into the conversation. Saved brand context helps frame the question without restating the business goal.
Illustrative question: “Given our goal, what should we check before increasing spend on this SKU? Use the report periods and stock snapshot above, and flag missing evidence.”
Illustrative explanation: “Spend increased while attributed sales fell, but the latest attribution may be incomplete. The October 11 stock snapshot suggests a gap before the seller’s estimated arrival. First confirm current available units, recent sales velocity and inbound timing with operations. Then reconcile mature advertising results and product economics before approving a spending change.”
This explanation is written for the guide. It is not a live AI response or a promise that every answer will reach the same conclusion. Connected reviews help investigate evidence; people approve and make Amazon changes. The brand manager has read-only account access and does not manage PPC, change prices or place inventory orders.
Other capabilities can supply context when relevant. Perfect Price supports manual, rule-based US price tests. Search performance uses completed weekly Amazon reports to compare the brand’s query funnel with the market. Refund Logic organizes reimbursement-report cases; it does not explain every refund cost in the P&L.
NevahOOS provides US Seller Central/FBA stock-planning context. If you add a Buy Box observation, record its location and collection time separately using the educational guide. ASIN Fetch / Buy Box is currently paused; the review example does not depend on that product.
Download a reusable review record
Download the blank weekly brand review (CSV). Add one row per brand and period, then record the source dates, brand goal, question, decision, owner, approver and review date. Use the follow-up result to carry the decision into the next review. The template has column headers only; it does not import Amazon data or calculate metrics.
Agencies can use this structure in a spreadsheet today. For a connected workflow, explore Brands or Agencies. The agency example uses client brand context, custom reviews and private/team conversations to support a client explanation and follow-up. Confirm data access and review setup on the demo; this guide does not promise an agency dashboard, automatic exports or scheduled client delivery.
End the report with a decision log
Every important finding should produce a proposed action or a deliberate decision to wait. Record the evidence, the explanation you are testing, the person responsible, and the next review date.
- Decision on October 13: wait for the missing evidence before approving increased spend.
- Owner and next check: operations checks available units, sales velocity and inbound timing on October 14; the ads lead checks report scope and attribution maturity.
- Approval: the brand owner approves any resulting inventory, pricing or advertising action.
- Review on October 20: record updated stock evidence, the latest comparable ad results, the approved action and what remains unknown.
A useful report can say that the evidence is incomplete. That is better than attaching a confident explanation to every movement. The next report should revisit the decision log and show what happened, not simply replace last week's charts with new ones.
Bring your reporting question to a demo
Bring one finding, the report dates, your brand goal and the decision you need to make. The call covers relevant data access, the review workflow and team-assisted setup. See current audience pricing and confirm agency usage terms during the demo.
Book a Brands demo or book an Agencies demo. If you want to build your own investigation through MCP, start with the plain-language MCP guide and FBA Builders. AI and scraping are metered; external AI costs may also apply. Confirm rates, connections and data scope on the demo.
Watch the original breakdowns
This guide adapts ideas from Alfredo’s videos into a reading format. Original videos may mention earlier product versions or offers; confirm current availability with the team.
Updated October 13, 2026 with a dated illustrative seller review, source-freshness checks, a contextual question, a human decision and follow-up, and current audience offers. Examples are editorial; no live dashboard or AI output is shown.
