PPC · Bid optimizer · Rules · Harvesting · Analysis
Amazon PPC bid optimizer for Bulk Operations files. Your ads never change until you've read the file.
A full Amazon PPC optimizer that runs on Amazon's own Bulk Operations spreadsheets — the same files Seller Central understands natively. It works for every seller from day one. Nothing touches your account until you upload the file back at Amazon yourself, and Amazon's uploader shows you its own per-row accept/error results. It's part of the Pro plan — one more standalone bill you can cancel.
7 days free · the full product on your real data · card required · cancel any time
- Mechanism
- Amazon's own bulk files — the format Seller Central reads natively
- Setup
- none — day one, every seller
- Who applies changes
- you, always
- Coverage
- keywords and product/auto targets, all four ad types
- The bill
- part of the Pro plan — no separate bid-tool subscription
- Their logic
- unpublished, unreadable, uneditable
- Their changes
- applied overnight, on their judgment
- Their skips
- silence — you never know what wasn't touched
- Their apology
- in the reviews, after the budget's gone
- Ours
- a file you read, row by row, and upload yourself
You've heard it. Maybe you've lived it.
An autopilot ad tool rewrote bids overnight and torched a week of budget before anyone looked. The bidding logic wasn't yours to read or edit. The category sells set-and-forget, then apologizes in the reviews. We inverted it. The tool proposes. You dispose. That's not a policy we promise — every change is a row you've read in a file you upload yourself, so nothing can happen overnight without you.
- File
- Amazon's bulk .xlsx, to 25 MB, drag-and-drop
- Window
- 60 days recommended, yours to change
- Review
- every row, every input, before anything exists to upload
- Apply
- you, at Amazon, with Amazon's per-row accept/error results
- Freshness
- ad data is as current as the file you downloaded — we won't call a spreadsheet "live ad data," because it isn't
Download. Optimize. Review. Upload.
One: download your Bulk Operations file at Amazon Ads and drag it in — .xlsx up to 25 MB, the recommended 60-day window suggested on screen, dates editable. Two: your rules, thresholds, and portfolio targets compute proposals with every input number in view. Three: accept, reject, or hand-edit any row — and read the "left unchanged" list, which has reasons. Four: upload the export back at Amazon yourself and read Amazon's own per-row results. The honest sentence, printed on the page: you upload the file yourself. That's work — a few minutes per cycle. It's also why nothing can ever change without you.
- Control
- bulk accept/reject per rule; filter by rule, sheet, portfolio; type over any bid
- Hand edits
- edit any bid or flip any row enabled/paused, even ones the rules wouldn't touch — out-of-limits edits warned before download
- Floors
- Amazon's per-ad-type minimum bids enforced (SP $0.02, SB $0.10, SB video $0.25, SD $0.02) — no illegal rows in the export
- Export
- changed rows only, Amazon's exact sheets and column order
- Verification
- Amazon's uploader is the final referee, per row
The rule-grid workflow sellers pay a standalone tool for. In your subscription.
Proposals for keywords AND product/auto targets across all four Amazon ad types — computed only for rows that are actually running (target, ad group, and campaign all enabled), because "optimizing" paused rows is how other tools generate garbage. Every proposed bid sits next to every input that produced it: impressions, clicks, orders, spend, sales, ACOS, CPC, CTR, CVR, campaign and portfolio names. Check the tool's homework on the same screen it's shown. And the export is a true delta: only the rows you changed, in Amazon's exact template — never a whole-account rewrite of thousands of no-op rows that buries the one error in a haystack you created. Open the file. Count the rows. That's the check.
- Unchanged rows
- a reason per row
- Default-bid targets
- what they really bid, surfaced
- Zero-impression campaigns
- named
- No-change runs
- explained, never blank
- Conservation
- every dollar in exactly one bucket — the report always sums
We also tell you what we didn't touch. With the reason. Per row.
An empty results table makes you wonder if the tool worked or silently skipped half your account. So everything the optimizer did NOT change gets a stated reason — including targets with no bid of their own (and what they actually bid via ad-group default), and campaigns switched on that got zero impressions. When the right answer is "no changes," it says why instead of showing you a blank. The same policy runs through the whole suite: anything a bulk file genuinely cannot show — organic sales, sessions — is labeled unavailable with the reason, never fabricated. And every dollar in your file lands in exactly one analysis bucket, so the totals always add back to the file, where some audit tools silently drop spend their categories don't recognize.
- Guard rails
- per rule: min bid, max bid, max increase %, max decrease % — a rule can never move a bid further than you allowed
- Portfolios
- account defaults + per-portfolio overrides: target ACOS, raise/cut %, bid limits, step-up, click and impression thresholds
- Formula bids
- revenue per click × your target ACOS, on day one
- Overrides
- five scopes, permanent until deleted; a live hand edit still wins on the day
- No file loaded
- saved rules and portfolio settings stay visible and editable
Ordered. First-match-wins. Not one threshold hard-coded.
Build conditions from any mix of impressions, clicks, orders, spend, sales, CTR, CVR, ACOS, CPC, and current bid — "clicks >10 and orders = 0 → cut bid" takes a minute. Actions go past raise/cut by % or $: set a bid straight from the formula — revenue per click × your target ACOS — or pause. Percentage nudges take months to converge; the math answer exists on day one. Drag rules to reorder; the order is the priority, stated right on the screen. Settings differ per portfolio, by name — your replens can afford 25% ACOS and your clearance can't afford 10%, and now the tool knows too. And Bid Overrides pin a bid at any scope — single target, campaign, portfolio, ASIN, or whole account — respected on every future compute until you delete it. The keyword you correct after every upload stops needing correcting.
- Thresholds
- yours, all three ways
- Evidence
- full metrics per flagged row, real bid included
- Pause
- opt-in per row, never automatic
- Dead rows
- already-paused campaigns and ad groups labeled — no puzzling
- Campaign switching
- full on/off control — the file is the mechanism
There's a keyword eating $60 a month with zero orders. This screen knows which one.
Flags keywords and targets burning money by YOUR thresholds — spend over $X with no orders, clicks over N without a sale, CTR under Y% at high impressions — with full metrics, campaign, and the real bid on every flagged row. A flagged row is only written into the file as paused if you tick it. Show-and-decide, never auto-pause: nobody's best keyword dies during its slow week because a robot got confident. Next to it, the Campaigns table shows every campaign in your file — state, targeting type, lifetime numbers, bid-row count — with one-click enable/pause riding the same export. Deliberately unfiltered, because the paused campaign you came to revive is invisible in every other tool's filtered view. Only flipped campaigns are written; undo all pending switches in one click.
- Harvests
- converting terms → their own targeting, destination suggested per term
- Negatives
- waste priced in dollars, per term, before the decision
- Thresholds
- minimum orders to harvest; minimum clicks and spend to negate — all yours
- Export
- both halves, one file, one upload
- The point
- stop paying auto-campaign prices for traffic you've already proven
"$214 spent, 0 orders" — next to a term you never meant to bid on.
Your auto campaigns already found the phrases that convert. The harvesting drawer splits actual customer search terms out of the bulk file: the converting ones become harvest candidates — each with a suggested destination ad group you can override, minimum-orders threshold yours — and the money-wasting ones become a block list with the wasted spend totaled in dollars before you decide. One export carries both — new keywords and negatives in Amazon's bulk format — because a postponed negative file keeps spending while you procrastinate.
- Split
- three placements, per campaign and account-wide
- Advice
- the exact +X%, derivation visible on screen
- Basis
- your own target ACOS × the placement's revenue per click
- Thin data
- refused, by name
- Replaces
- gut feel at +50%
"This placement justifies +X%" — and here's the math.
Performance by top of search, product pages, and rest of search — per campaign and account-wide — then what each placement can actually afford, derived from your target ACOS and its revenue per click, down to the exact suggested boost percentage. And when a placement doesn't have enough evidence, it says so: placements under the evidence threshold are named, not guessed at. A confident recommendation built on 12 clicks is how budgets die. We'd rather tell you we don't know.
- Structure
- negatives at both levels, done right by default
- Targeting
- keyword, competitor-ASIN (with expanded matching), or auto
- Names
- duplicate-checked against your file before the workbook builds
- Validation
- Amazon's uploader, per row
- Self-competition
- caught before it exists, by the Active Checker
One form becomes a complete, uploadable campaign.
Build a complete Sponsored Products campaign — campaign, ad group, product ad, targets, negatives, bidding strategy, placement adjustments — from a single form, downloaded as a ready-to-upload bulk file Amazon validates row by row. Three targeting types: keyword lists (Exact/Phrase/Broad in any mix), competitor-ASIN with expanded matching, or auto. Negative keywords set independently at campaign AND ad-group level, plus negative ASINs, portfolio, dates, and placement percentages. Before you build, the Active Checker: paste keywords or competitor ASINs and see how many running targets already cover each one, and where — so you never build the campaign that bids against your own and pays Amazon twice for the same click.
- Rollup
- by ad type, by placement, all metrics, running-only or everything-built
- Branded split
- your brands, typed once, priced in dollars
- Per-ASIN audit
- one click
- N-grams
- the one word poisoning twenty terms, found
- Housekeeping
- Harvest List + Renamer, previews before export
Account health, branded spend, and the words that drain — one tab.
Quick Insights: whole-account rollup by ad type and placement across every metric, with a branded vs. non-branded split — type your brand names and see how much ad money rides clicks you might have gotten organically; sellers pay audit services for that one number. Snapshot Audit: a one-screen health check — ACOS, CVR, ad sales and orders, CPA, CPC, CTR, spend, impressions, target counts, each ad type's share of sales and spend — filterable to a single ASIN, so one product's ads get audited in one click. N-grams: 1-, 2-, and 3-word analysis across all search terms with full metrics per phrase, filterable by ASIN and minimum impressions — the word "cheap" can poison twenty terms at once, and term-by-term review would never catch it. Harvest List: rule-based promotion candidates with a minimum-orders threshold and per-lane selection. Renamer: bulk-rename campaigns or ad groups from a template with variables — brand, ad type, old name, prefix/suffix — previewed old → new before anything exports. Three years of naming chaos, one afternoon.
The trade, in a table
Bring last month's bulk file.
Ten minutes in the trial: drop it in, read the proposals, read the "left unchanged" reasons. Upload nothing if you don't like what you see. That's the whole deal.