Automatic Campaigns on Amazon: What They Do and When to Fix Them

Automate or manage by hand? A rule you can check

The question hides two choices: how your ads are targeted (an automatic campaign, where Amazon picks the search terms) and how your ads are managed (by hand, or by software or a service). Pick each one separately.

Start with the number every bid depends on: your break-even bid, which is what is left per order before ads, times your conversion rate. Example: a $40 product with $14 left before ads and a 10% conversion rate gives $14 × 10% = $1.40. Any click bid above $1.40 loses money on average at that conversion rate. Checking that across one ASIN is easy by hand. Checking it across dozens of ASINs every week is where most manual accounts slip.

• Manage by hand when: you run a handful of ASINs, have the hours for a weekly search term review, and enjoy the work.
• Automate when: the catalog is large, margins differ by SKU, and the weekly cycle of harvesting, negating and re-bidding keeps getting skipped.
• Keep a human on: launches, pricing, creative and inventory risk, whichever way you go.

Whichever you choose, ask the same of a person or a tool: each change should come with its evidence, a measurement plan and a rollback trigger.

Two different things get called an "automatic campaign"

Mixing these up is the first mistake. The first is a setting inside Sponsored Products: at campaign creation you choose Automatic targeting, and Amazon's algorithm decides which search terms and product pages your ad shows against instead of you writing a keyword list. The second is automated or autonomous campaign management — software or a service that adjusts bids, budgets, and negatives for you on a schedule or by rule, sitting on top of whichever targeting type you picked.

Most people typing this phrase mean the first. Anyone comparing tools or agencies usually means the second. This page covers both, because a complete answer needs to.

How Amazon's automatic targeting actually decides who sees your ad

Amazon reads your listing — title, bullets, backend keywords, category — and matches your ad against search terms and product pages it judges relevant. It sorts what it finds into four match types, and they don't behave the same. Treating them as one bucket is how accounts lose money without noticing.

Close match and loose match come from search queries; substitutes and complements come from other product pages. Your bid applies across all four unless you split them into their own campaigns — which is worth doing once you can see a pattern in the search term report.

A worked example: 30 days of automatic spend, and what to do with it

Say you launch a new ASIN with one automatic campaign at a $40 daily budget — about $1,200 for the month — and a modest starting bid. After 30 days you pull the search term report and split spend by match type. A pattern this common is worth walking through: close match runs 40% of spend and carries most of the sales; loose match is 35% of spend, with a handful of terms burning budget and converting rarely; substitutes are 15% and roughly break even; complements are the remaining 10% with almost no sales.

The fix isn't turning the campaign off. It's reading the report line by line: negate the loose-match terms that spent without converting, and for the close-match terms that did convert, add them as exact match in a manual campaign where you set the bid yourself instead of the algorithm's default. Automatic keeps discovering; manual captures what's proven.

The common mistake: treating automatic as "set and forget"

The mistake we see most — and one we've made ourselves on new launches — is checking an automatic campaign once a quarter instead of every week. Left alone, it doesn't get worse in a straight line; it drifts, usually in loose match and substitutes, as Amazon keeps testing adjacent traffic that looked related at launch and stops converting once the honeymoon period ends. A campaign that looked efficient in week one can be spending a third of its budget on searches that never should have qualified by week six.

The discipline that fixes this isn't complicated, but it has to happen on a schedule: pull the search term report weekly, harvest the winners, negate the losers, then look at the bid. Every proposed change carries three things before it runs: the evidence behind it, a measurement plan, and a rollback trigger — that discipline holds whether a person is doing the harvesting by hand or a system is doing it for them. Skip any one of those three and a good automatic campaign quietly turns into a bad one.

When the number looks wrong: troubleshooting an automatic campaign

If ACOS is climbing, spend is concentrated in one match type, or sales are flat despite budget headroom, check these before touching the bid:

• Budget capping early in the day: if the campaign runs out by midday, Amazon can't show it during your best-converting hours. Raise budget before touching bids.
• Negative list is empty or stale: an automatic campaign with no negatives after 60 days is almost always letting decent close match pay for loose match's sins.
• Bid strategy setting: dynamic bids down-and-up behaves differently than down-only; confirm which is set before crediting or blaming a bid change.
• You're reading campaign-level ACOS, not match-type ACOS: a healthy blended number can hide a substitutes bucket losing money every day.

If none of that explains it, the listing itself may not give Amazon enough signal — automatic targeting is only as good as the content it's reading.

Where autonomous management fits

Everything above is manageable by hand, and plenty of sellers do it well at a handful of ASINs. It gets harder at scale — dozens of products each needing that same weekly harvest-and-negate cycle, across automatic and manual campaigns at once. That's the second meaning of "automatic campaign": management that runs the cycle for you.

Dr. PPC is built for that layer. It reads the whole ad account, writes a strategy per product against that brand's real economics, and proposes each change with the evidence, measurement plan, and rollback trigger described above — the client sets the autonomy level, from every change waiting on a click to routine changes running automatically with the bigger ones queued for review. Inventory risk, pricing, new launches, and creative always go to a human regardless of setting. It's $299/month plus 2% of ad spend, month-to-month, with the first 30 days for $1, and Orbit — the analytics and tracking suite most competitors sell as a separate subscription — is included. It's operated by Full Circle (100+ brands managed, $500M+ in Amazon revenue managed). None of that replaces understanding what an automatic campaign does; it just means the harvesting happens on schedule whether or not you have the hours for it.