Glossary

87 terms, each defined once and used consistently everywhere else. Where the industry overloads a word, the definition says which meaning is in play.

A

• ACOSAdvertising Cost of Sale Ad spend divided by ad-attributed sales, expressed as a percentage. A measure of advertising efficiency.
• Ad group A set of keywords or product targets pointed at a set of products.
• Alert Something happening now that may need you today. Alerts never change anything.
• AOVAverage Order Value Average revenue per order. Sets how much a conversion can afford to cost.
• Auto campaign An Amazon Sponsored Products campaign using automatic targeting. Amazon matches the advertised product to relevant searches and products.

B

• Bid The maximum you will pay for a click. You usually pay less — that is your CPC.
• Bid change Move a bid toward what the evidence says the target can afford.
• Bid floor and cap Hard limits no proposed bid may cross, whatever the math says.
• Brand term Your own brand names. Never negated, never bid down.
• Break-even ACOS Contribution before advertising divided by revenue. The approximate ACOS at which an attributed sale leaves no contribution.
• Budget brake A temporary budget cut on a campaign running far above target. A stopgap, not a fix.

C

• Campaign The top level of an Amazon ad structure. Holds the daily budget.
• Campaign group A named set of campaigns sharing a target — usually one product family.
• Capped under target A campaign hitting its daily budget while beating its ACOS target — profitable and throttled.
• Change history Every change AutoPPC has made, what it was before, whether it landed, and whether it worked.
• Clamp The most a bid may move in a single run. Stops one run from restructuring your account.
• Conquest winner A competitor's product page where your ad converts. Effective, and a strategic call.
• CPCCost Per Click What you actually paid per click — not what you bid.
• CVRConversion Rate Attributed orders divided by clicks, expressed as a percentage. Compare like-for-like reports and settled date windows.

D

• Daily budget The most a campaign can spend in a day. When it caps, your ads stop showing.
• Data throughThe last day of data a run measured The last day of Amazon data a run measured. The most recent days are trimmed until attribution settles.
• Day-5 judgment The verdict on whether a change actually helped, taken once attribution has settled.
• Decision window How many days of evidence a decision uses. Longer is more reliable, shorter is more current.
• Dedup against live state Checks your live account before proposing, so you never get a change that already exists.
• Direct apply Applying approved changes through Amazon’s API. This delivery method is not available in AutoPPC’s current beta.
• Discovery layer The auto or broad campaigns whose job is finding new search terms to harvest.

E

• Earned CPC The CPC a search term actually paid when it converted — used as a harvest's starting bid.
• Estimated impact What approving a set of proposals would be worth per week, based on frozen evidence.
• Evidence The exact numbers that triggered a proposal, frozen at the moment it was made.
• Evidence bar The minimum evidence before a decision may be made — usually a click count.
• Export The Amazon bulk files containing everything you approved, ready to upload.

F

• Family check AutoPPC will never negate a term if any variant of it has orders anywhere in your account.
• Fluke One click, one order, at a low CPC — a coincidence that looks like a great keyword.

G

• Gate A check every proposal must pass before it can reach you. Most pass silently.
• Generic term A broad category phrase like 'cutting board' — high traffic, low intent, converts poorly.
• Generic winner A broad term that converts anyway — rare, valuable, and worth its own campaign.

H

• Hands-off campaign A campaign AutoPPC never proposes changes to. Monitors still watch and report.
• Harvest Give a proven search term its own keyword so you can bid it deliberately.

I

• Input lock The fingerprint of every input a run used, so the run can be reproduced exactly.
• Intensity How much rope the engine gives each target — from Profit-first to Aggressive.
• Invariant A rule nobody can switch off — including us. Enforced in the engine, not the interface.

J

• Junk term A search term for something you do not sell at all. Pure waste, high confidence.

K

• Kept The day-5 verdict: the change stayed because it worked, or was reverted because it didn't.
• Keyword A phrase you choose to bid on, with a match type controlling how loosely it matches.
• Kill gate Nothing is paused until its ACOS clears a multiple of your target across a full evidence window.

L

• Later Defer a proposal. It comes back next run with fresh evidence.
• Learning Something the engine worked out about your account from its own history.

M

• Margin-weighted Deriving each group's target from real product margins instead of one chosen number.
• Marketplace One Amazon country. Kept entirely separate — different currency, competition, and targets.
• Match type How loosely a keyword matches shopper queries — exact, phrase, or broad.
• Materiality floor Proposals worth less than this are withheld — not worth your attention.

N

• Negate Block a search term so it stops matching and stops costing you money.
• Norm A campaign's learned normal daily spend, used to spot anomalies.
• Not buyable You are paying for clicks on something that cannot currently be bought.

O

• One lever per run Bid and placement modifier never move on the same campaign in one run.
• Opportunity A structural finding that needs a decision, not an urgent problem. Persists until acted on.
• Own-catalog exclusion AutoPPC will not auto-harvest your own ASINs. Self-targeting is a deliberate choice.
• Owner decision A trade-off AutoPPC will not make for you. The row arrives unticked and waits for your call.
• Owner pauses are sacred If you paused something, AutoPPC will never propose turning it back on.

P

• Pause Stop a target or ad from serving. Reversible, but not automatically.
• Placement Where your ad appeared — top of search, rest of search, or product pages.
• Placement modifier A percentage that raises your bid for one placement. Effective bid = base × (1 + modifier).
• Product economics Marketplace-specific product prices, costs, fees and inventory context used to assess margins and supported safety checks.
• Product target Bidding on a specific ASIN or product category instead of a search phrase.
• Proposal One suggested change, with its reason and frozen evidence. Nothing until you approve it.
• Protected family A term family marked never to be negated or bid down. Seeded from your own converting data.
• Provenance Where a number came from — which report, which pull, which run, who last touched it.

R

• Re-audit Re-checking a settled change against fresh data to confirm the verdict still holds.
• Receipt coverage Whether retained report days reconcile to the immutable receipt captured when Amazon data was ingested.
• Reject Stop AutoPPC proposing this change to this item in this direction — durably.
• Reverse file The undo. Generated before anything ships, containing every prior value.
• Review queue Where proposals wait for your approve, later, or reject. Nothing ships without a tick.
• ROASReturn on Ad Spend Ad sales divided by ad spend — the inverse of ACOS. A 4× ROAS is a 25% ACOS.
• Run One pass of the engine over your account, producing a reviewable set of proposals.

S

• Sanity cap The maximum changes any one campaign can receive in a single run.
• Search term What a shopper actually typed. Your keyword is what you bid on; this is what matched it.
• Settle window The wait after a change before judging it, so Amazon can finish attributing sales.
• Source block A conditional negative in a separate discovery lane after the harvested exact serves.
• Specific term A precise phrase describing exactly what you sell. Lower traffic, much higher intent.
• Spend anomaly A campaign spending well above its learned norm today.
• Stock cover How many days of inventory remain at current velocity. Gates every bid raise.

T

• TACOSTotal Advertising Cost of Sale Ad spend divided by total sales for the same dates, scope and currency. It is not a direct measure of profit or causation.
• Target ACOS The efficiency you are aiming for. Every bid decision is measured against it.
• Target source Where your governing ACOS number comes from: account-wide, per group, or from margins.
• Total sales All sales in a marketplace — advertised and organic. Required to compute true TACOS.

V

• Verified Confirmation from the next data pull that a change actually landed on Amazon.

W

• Write mode How approved changes reach Amazon — bulk files you upload, or direct API apply.