Amazon PPC Help: Find the Actual Problem First

What kind of help do you actually need?

“Amazon PPC help” covers three very different problems, and the fix for one won't touch the other two. First: you don't know if your campaigns are working at all — nobody told you what to look at. Second: you know something's off — ACOS climbing, sales flat — but you don't know which lever caused it. Third: you already tried something — cut a bid, added a negative keyword, raised the budget — and it didn't move the number you expected. Work out which one you're in before you open the console.

• No visibility: you're only checking total spend and total sales. Start with the search term report and the placement report, not the bid.
• Visible problem, unknown cause: you can see the number is wrong. The four levers below tell you where to look.
• A fix that didn't work: either it hasn't had enough clicks yet, or it wasn't the real cause. Covered further down.

The four levers that move Amazon PPC performance

Every Amazon PPC problem traces back to one of four things: structure (how campaigns and ad groups are organized), targeting (which keywords or ASINs trigger the ad), bids (what you pay per click), and budget (how much you'll spend before the campaign stops serving for the day). Most sellers only ever touch bids, because it's the field right in front of them. It's usually not where the problem lives.

Here's a worked example. Say a product sells for $40, and after cost of goods and Amazon's fees, $16 of that is margin — a 40% breakeven ACOS. Push ad spend past that and the unit loses money on the ad line even though it sold. Say the account is running at 55% ACOS on $500 a day of spend. That's not a bidding problem you fix by nudging every keyword down 10%. It's a targeting problem: pull the search term report for the last 60 days and look for terms that have taken meaningful spend with zero orders. In accounts we've audited, that's almost always a handful of broad-match keywords pulling in loosely related searches — the kind of thing where “sneakers” matches “running shorts” — spend that never converts but keeps getting impressions because the bid is high enough to win the auction.

Fix the targeting first — move the converting terms to exact match, negative out the ones that spent without converting — before touching the bid on anything else. Cutting bids account-wide when three keywords are the real problem starves your winners along with your losers, and total sales drop with it.

When the fix doesn't work

Three checks before you assume a fix failed. First, did it actually apply — bid and budget changes can take a few hours to fully propagate, and checking the same day means you're looking at a mix of old and new data. Second, did it get enough data — a keyword needs a real sample of clicks before its conversion rate means anything; judging a change after eight clicks is judging noise. Third, did something else move at the same time — a competitor's price drop, your own stockout, a listing edit — that would swamp whatever the PPC change was supposed to do.

The discipline that avoids most of this: decide the measurement and the reversal point before you make the change, not after. Every proposed change carries three things before it runs: the evidence behind it, a measurement plan, and a rollback trigger. If you cut a bid, write down what ACOS or click volume you expect to see, by when, and the number that means you were wrong and should put it back. Most PPC accounts don't have a bad strategy — they have no record of what they already tried, so the same fix gets tested three times a year, forever.

Mistakes even experienced accounts make

• Killing a keyword after one bad day. We've made this one too — pausing something after a single 24-hour window because ACOS spiked, before it had enough clicks to say anything real. A keyword with 40 clicks and no orders is a different problem than one with 4 clicks and no orders.
• Chasing ACOS with no reference to margin. A 25% ACOS on a product with 20% margin is a loss. A 45% ACOS on a product with 55% margin is fine. The target has to come from the P&L, not a benchmark from a blog post.
• Treating Sponsored Products, Sponsored Brands, and Sponsored Display as one lever. They do different jobs — capturing existing search demand, building brand presence, and reaching audiences off your listing. Judging all three against the same ACOS target misreads what each one is for.
• Automatic and manual campaigns bidding on the same keywords. This inflates your own cost per click without anyone noticing, because the account is competing against itself.

Where this fits

Everything above is doable by hand, and plenty of sellers do it well — most Amazon PPC content stops at definitions because the diagnostic work is repetitive, not because it's secret. Where it gets hard is doing it consistently, every week, across a catalog with more than a handful of SKUs, without a system that remembers what was already tried.

Dr. PPC is built for that part. It's not a retainer agency and not a dashboard you have to operate yourself — it reads the account, writes a strategy per product against that product's real margin, and proposes each change with the evidence, the measurement plan, and the rollback trigger built in, before anything runs. You choose how much of it happens automatically; a few things — inventory risk, pricing, new launches, creative — always come to a human regardless of setting. The software underneath, Orbit, covers search-term profitability, inventory, BSR, buy box, and fee tracking, and it's included rather than sold separately. It's operated by Full Circle, which has managed over $500M in revenue across 100+ brands, and pricing is $300 a month plus 3% of ad spend, capped, month-to-month, first 30 days free. If you're running the account yourself, the diagnostic table above works regardless of who's doing it.