Most Amazon Ads accounts don’t underperform because the team lacks knowledge. They underperform because the work happens inconsistently: search terms get reviewed when someone remembers, bids get changed reactively after a bad week, and nobody can say why a budget was raised three Tuesdays ago.

A weekly review SOP (standard operating procedure) fixes that. It defines exactly what gets checked every week, in what order, against what thresholds, and what gets documented. Run it consistently and small optimizations compound; skip it and you end up firefighting.

This guide gives you the complete weekly review process — account summary, exceptions, search terms, budget pacing, bids and placements, and the weekly action log — plus how to turn the SOP from a document into a workflow your team can run on schedule.

Quick Answer

A weekly Amazon PPC review should evaluate spend, sales, ACOS, TACOS, budget pacing, search terms, bids, placements, campaign structure, and major changes from the previous period. The goal is to identify what needs action, what needs monitoring, and what should be documented for future decisions. A good weekly review SOP defines the data inputs, decision thresholds (such as click limits for negation and ACOS bands for bid changes), the outputs (an action list and summary), and who approves changes. Brands and agencies that run the same review every week catch budget and search term problems early instead of discovering them in a monthly report.

Who This Is For

This SOP is written for two readers:

• Brand operators managing their own Amazon Ads account who want a consistent weekly routine instead of ad-hoc tweaking.
• Agency account managers and leads who need every account on the roster reviewed against the same standard, regardless of who runs it.

It assumes campaigns are already live. If you’re still setting up campaign structure, start with a setup guide first and come back when you have at least a few weeks of data.

Why Weekly Is the Right Cadence

Daily checks should be lightweight — spend pacing and anomalies only. Monthly reviews are too slow: a search term can burn a meaningful share of budget in four weeks, and the best-performing terms from three weeks ago may have already shifted.

Weekly is where tactical optimization lives. As a rule of thumb, the cadence should scale with spend: an account spending $10/day doesn’t generate enough new data to justify a weekly deep dive, while an account spending $1,000+/day demands one. For most active accounts in between, a 30–60 minute structured weekly review is the right default, with a deeper strategic review monthly.

One caution: weekly does not mean reactive. Bid changes need enough data behind them. The review surfaces what has statistically changed; it shouldn’t trigger panic adjustments off two days of noise.

The Weekly Review Workflow, Step by Step

Each step below has an input, logic, and an output. That structure matters — it’s what makes the SOP executable rather than a vague checklist.

Step 1: Account-level summary

Pull the last 7 days versus the prior 7 days: total spend, ad sales, ACOS (Advertising Cost of Sale — ad spend divided by ad-attributed sales), and TACOS (Total Advertising Cost of Sale — ad spend divided by total revenue including organic). ACOS tells you ad efficiency; TACOS tells you whether advertising is carrying more or less of the business over time.

Output: one or two sentences of status. “Spend flat, ACOS up 4 points, TACOS stable” is enough to direct the rest of the review.

Step 2: Exception detection

Instead of scanning every campaign, look only at what changed materially week over week:

• Campaigns with spend up or down more than ~25% with no known cause
• Campaigns whose ACOS moved outside their target band
• Campaigns repeatedly hitting their daily budget cap
• Campaigns with impressions collapsing (possible suppressed listing, lost Buy Box, or out-of-stock issue)

Output: a short list of campaigns that need a closer look. Everything else gets left alone — restraint is part of the SOP.

Step 3: Search term review

Pull the search term report for the trailing 7–14 days and classify terms:

• Harvest: the term converts at or better than your efficiency threshold → add it as an exact-match keyword in a dedicated campaign, then negate it in the source (auto/broad) campaign so the two don’t compete.
• Negate: the term has accumulated roughly 15–20 clicks with zero conversions, or its spend has exceeded your unit margin with no orders → add as negative exact in the ad group where it appears.
• Monitor: early signal but not enough clicks to act → log it and revisit next week.
• Ignore: low-volume noise.

The exact click threshold depends on your conversion rate: if the product converts at 10%, ten clicks without a sale is one missed expectation, not proof of failure. Use 1–3x your expected clicks-per-order as the negation threshold, tighter for mature products, looser for launches.

Step 4: Budget pacing

Compare month-to-date spend against the plan:

• Over-pacing and inefficient → reduce budgets or bids on the offending campaigns.
• Over-pacing and efficient → a good problem; consider reallocating budget from weaker campaigns before raising total spend.
• Under-pacing and efficient → raise budgets or bids; you’re leaving demand on the table.
• Under-pacing and inefficient → don’t force spend; diagnose targeting, bids, or conversion issues first.

Step 5: Bid and placement review

For targets with enough data, apply a simple ACOS-band heuristic that practitioners commonly use:

• ACOS at 1–2x target → reduce bid ~10%
• ACOS at 2–3x target → reduce bid ~20%
• ACOS above 3x target → reduce bid ~30%
• ACOS below target with sales → increase bid ~10%

Then check the Sponsored Products placement report (Amazon breaks performance out by Top of Search, Rest of Search, and Product Pages). If Top of Search converts dramatically better, a placement modifier is usually more efficient than raising base bids across the board. Review placement modifiers every week or two, not daily.

Step 6: Weekly summary and action log

Write down: what was changed, why, and what should be watched next week. This is the most skipped step and the most valuable one. Without a decision log, next month’s review starts from zero, and in an agency, the log is what makes account handoffs and client reporting defensible.

Example: One Pass Through the Search Term Step

• Input: 14-day search term report for a kitchen brand; target ACOS 30%; product conversion rate ~10% (so ~10 clicks per expected order).
• Workflow: filter to terms with ≥15 clicks. One term, “collapsible silicone funnel,” has 22 clicks, 4 orders, 18% ACOS. Another, “funnel cake mix,” has 19 clicks, 0 orders, $14 spend against an $11 unit margin.
• Expected output: harvest “collapsible silicone funnel” into an exact-match campaign and negate it in the source auto campaign; add “funnel cake mix” as negative exact where it appears.
• Business decision: the harvested term gets its own bid and budget control going forward; the negated term stops draining spend this week instead of at the end of the month.

Tools and Data Needed

• Search term report (Seller Central → Campaign Manager → Measurement & Reporting), trailing 7–14 days
• Campaign performance data with week-over-week comparison
• Placement report for Sponsored Products
• Your target ACOS/TACOS per product or portfolio, and unit margins
• A place to log decisions — a spreadsheet works at small scale

Common Mistakes

• Reviewing without thresholds. “Look at search terms” isn’t an SOP. “Negate at 15–20 clicks with zero orders” is. Vague steps produce inconsistent decisions across weeks and people.
• Changing bids on thin data. Weekly cadence with daily-sized samples produces whiplash. Let targets hit minimum click counts before acting.
• Harvesting without negating the source. If you promote a term to exact match but leave it live in the auto campaign, you bid against yourself.
• Treating every campaign equally. Exceptions first. A campaign that’s stable and on target needs no touch this week.
• Skipping the log. Undocumented changes make it impossible to learn from outcomes — or to explain them to a client.

Where the Manual SOP Breaks Down — and Where Trellis Fits

Run manually, this review takes 30–60 minutes per account when nothing is on fire. The breakdown comes with scale and consistency:

• Across ten accounts, the same SOP is a full day of repetitive report-pulling and filtering every week.
• Different team members apply thresholds slightly differently, and standards drift.
• The documentation step erodes first, because it’s the least urgent.

This is the natural fit for Trellis: the weekly review SOP you just defined — inputs, thresholds, classifications, approval steps — becomes a saved workflow that runs on schedule. The AI applies your harvest/negate logic and pacing rules, and your team reviews and approves the recommendations instead of rebuilding the analysis from scratch each week. For the reasoning behind that division of labor, see Why LLMs Should Build Amazon Ads Workflows, Not Run Them.

Run this review manually for a few weeks first — you’ll understand your account’s rhythms and sharpen your thresholds. Then make it repeatable: Turn your weekly review SOP into an automated workflow with Trellis →

FAQ

How often should I optimize Amazon PPC campaigns? Match cadence to data volume. Light daily checks for pacing, a structured weekly review for tactical changes (search terms, bids, budgets), and a monthly strategic review. Low-spend accounts (~$10/day) can run the tactical review monthly; high-spend accounts ($1,000+/day) need it weekly.

How long should a weekly Amazon PPC review take? With clean data and defined thresholds, 30–60 minutes per account. If it routinely takes longer, the SOP is missing thresholds or the account is carrying structural problems a weekly review can’t fix.

What should a weekly Amazon Ads report include? Week-over-week spend, sales, ACOS, and TACOS; the exceptions investigated; search terms harvested and negated; budget and bid changes made with reasons; and items being monitored for next week.

How do I review an auto campaign weekly? Focus on the search term report: negate non-converting terms, harvest converters into exact match (negating them in the auto campaign), and adjust bids on the auto targeting groups (close match, loose match, substitutes, complements) according to your ACOS bands.

ACOS or TACOS — which should drive weekly decisions? Use ACOS for tactical weekly moves (bids, negatives, budgets) because it’s directly attributable. Use TACOS as the monthly strategic gauge of whether advertising dependency is rising or falling. Weekly TACOS swings are often noise.

What’s a good negation threshold? A common rule is 15–20 clicks with zero conversions, or spend exceeding your unit margin with no orders. Adjust by conversion rate and lifecycle: tighter for mature, margin-focused products; looser for launches still gathering data.

Can AI run my weekly PPC review for me? AI is best used to build and run the structure — pulling data, applying your thresholds, drafting recommendations — with a human approving consequential changes. An unconstrained chatbot making live bid decisions has no memory of your strategy, guardrails, or history. See AI for Amazon Ads for the full picture.

Do agencies need a different SOP than brands? The steps are identical; the difference is governance. Agencies need the SOP standardized across account managers, with documented thresholds per client and a decision log that supports reporting and QA across the whole roster.

Conclusion

A weekly Amazon PPC review only works if it’s the same review every week: pull the same inputs, apply the same thresholds, produce the same outputs, and log the decisions. Start manually — six steps, 30–60 minutes — and tune the thresholds to your margins and conversion rates. Once the process is stable, stop re-doing the mechanical parts by hand.

Build your weekly review as a repeatable workflow with Trellis →