What's the difference between an Amazon PPC strategy and PPC management?

Strategy is the plan: how campaigns are structured, which match types feed which, how negatives flow, and how budget ties to inventory. Management is the weekly execution against that plan. You can have great management of a bad strategy and still lose money, which is why the plan comes first.

How long before a new PPC strategy shows results?

Two to four weeks to see directional ACoS and impression share changes, six to eight weeks before the harvesting cycle has produced enough exact-match campaigns to drive meaningful efficiency gains. Anyone promising 30% ACoS cuts in week one is either lucky or cutting spend on converters.

Should auto campaigns run forever or get paused once I've harvested terms?

Keep them running at reduced budget. Auto is your discovery engine. Amazon adds new search behaviour every quarter, and an auto campaign with tight negatives catches it. Pausing auto entirely means you stop finding new keywords.

What's a reasonable starting ACoS target?

It depends on margin, not category. Work out your break-even ACoS (gross margin minus referral fee minus FBA) then target 60-70% of that for profit campaigns and up to break-even for launch or ranking campaigns. Blanket 20% ACoS targets without margin context are meaningless.

Do I need separate campaigns for Sponsored Products, Brands, and Display?

Yes, and separate budgets for each. The ad types serve different stages of the funnel and bid against different inventory. Mixing them into one campaign structure makes attribution impossible and leads to over-spending on whichever type is cheapest rather than whichever is working.

Amazon PPC Strategy — When to Build Your Own vs Hire It Out

91 vetted agencies · Updated May 2026 · Amazon Ads Verified Partner

A real amazon ppc strategy isn’t “raise bids on converters, lower bids on losers.” That’s tactical bid management, and it’s downstream of everything that actually matters. The strategy is a five-layer framework: campaign architecture, match-type ladder, keyword harvesting cycle, negative layering, and budget pacing tied to inventory velocity. Get those right and bid optimisation is almost automatic. Get them wrong and no amount of rule-based automation will save you.

This page walks the framework end-to-end, then helps you decide whether to run it yourself or bring in a partner from our vetted partner network.

The 5 layers of an Amazon PPC strategy

Think of the five layers as a stack. Each one depends on the one below it. If campaign architecture is wrong, your match-type ladder leaks money. If your negatives are wrong, your harvesting cycle pollutes your exact-match campaigns. Order matters.

Campaign architecture. The split between auto, manual broad, manual phrase, manual exact, and branded defense. This is the container everything else lives inside.
Match-type ladder. How search terms flow from discovery (auto/broad) down to conversion (exact), with bids increasing as match type tightens.
Keyword harvesting cycle. The weekly or bi-weekly process of promoting converting search terms up the ladder and demoting losers.
Negative layering. Campaign-level negatives to block unprofitable terms, ad-group-level negatives to prevent match-type cannibalisation.
Budget pacing. Daily and monthly budgets tied to inventory velocity, not fixed spend targets.

Most accounts we audit get layers 1 and 5 wrong, which means 2, 3, and 4 can’t function even if they’re set up correctly.

Campaign architecture: auto, manual, research, branded

The baseline split that works for most catalogues has four buckets.

Auto discovery. One auto campaign per ASIN or tight ASIN group, low-to-medium bids, running as a keyword discovery engine. Four ad groups inside each auto campaign: close match, loose match, substitutes, complements. This lets you pull separate search term reports and treat each differently.
Manual research (broad and phrase). Manual campaigns using broad and phrase match on keywords harvested from auto. Bids sit slightly above auto to catch volume before pushing terms to exact.
Manual performance (exact). Exact-match campaigns for proven converters. Highest bids, tightest budgets, cleanest ACoS. This is where the money gets made.
Branded defence. Exact-match campaigns on your own brand terms plus competitor brand terms where policy allows. Low bids, high conversion rates, protects against competitor bidding on your name.

If you’re selling under 10 SKUs, you can run this with four campaigns per ASIN. Over 50 SKUs and you’ll want to group by parent ASIN or category. The campaign structure primer on the blog goes deeper on naming conventions and portfolio setup.

Match type ladder and harvesting funnel

The ladder is how keywords graduate. Auto finds them, broad tests them at volume, phrase tightens intent, exact locks in the converters.

The harvesting rule we use: a search term that’s produced 10+ clicks and converted at or below target ACoS in auto or broad gets promoted to its own exact-match ad group at a 20-30% higher bid. The same term gets added as a negative exact in the source campaign so you’re not paying twice for the same click.

Terms with 15+ clicks and zero conversions get added as negative exacts in their source campaign. Terms with spend over 2x target CPA and no conversions get negative-phrased if the root word is unprofitable across variations.

This cycle runs weekly for active catalogues, bi-weekly for slower movers. Skip it for a month and your exact campaigns stop receiving fresh converters while your auto/broad campaigns accumulate waste. For the keyword research side of this, the sponsored products keyword strategy piece covers the discovery end, and the sponsored brands keyword strategy guide handles Sponsored Brands specifically.

Negative layering: campaign vs ad-group level

Two types of negatives, two jobs.

Campaign-level negatives block entire search concepts. Wrong-intent terms (“free,” “cheap,” competitor brands you can’t legally bid on), unprofitable categories, irrelevant variations. These sit at the campaign level because they apply to every ad group inside.
Ad-group-level negatives prevent match-type cannibalisation. If “stainless steel water bottle” is a converter in your exact campaign, it should be a negative exact in your broad and phrase campaigns. Otherwise broad wins the auction sometimes, you pay more for the same conversion, and your exact campaign looks underperforming because it’s getting starved.

Most accounts we see have one or the other. Accounts that do both properly run 15-25% lower ACoS at the same spend level, just from preventing self-competition. There’s more detail on the mechanics in our PPC optimisation walkthrough and the companion keyword strategy piece.

Budget pacing tied to inventory velocity

This is the layer most strategies skip, and it’s the one that silently kills accounts going into Q4.

Fixed daily budgets make no sense when inventory is finite. If you’ve got 400 units and 60 days of supply, spending hard on ads that drive above-forecast velocity just means you stock out in week 8 and lose ranking you paid to build. If you’ve got 2,000 units and 180 days of supply, under-spending is leaving growth on the table.

The pacing rule: calculate days of cover (inventory / daily velocity), then set monthly ad budget as a percentage of revenue based on that cover.

Under 45 days of cover: cut ad spend 30-50%, preserve rank, don’t accelerate sellout.
45-90 days of cover: standard spend at target ACoS.
90-180 days of cover: increase spend 20-40%, push for rank and velocity.
Over 180 days: aggressive spend, consider launching new keywords and ranking campaigns.

Tie this to an inventory alert, not a calendar. Re-check weekly during Q4, monthly otherwise. Amazon’s own guidance on budget management covers the platform mechanics but doesn’t connect to inventory, which is where most sellers lose money.

DIY vs managed: when each makes sense

The framework above is the same whether you run it or someone else does. The question is whether you’ve got the time, tooling, and temperament to execute it weekly.

Run it yourself if:

You’ve got under 30 active SKUs.
You can give PPC 4-6 hours a week, every week, without skipping.
Your margin is thin enough that agency fees would eat the ACoS savings.
You want to learn the mechanics so you can scale it later.

Hire it out if:

You’ve got 50+ SKUs or multiple marketplaces.
Ad spend is above $15-20K/month (the point where a good manager pays for themselves in efficiency gains).
You’re scaling fast and PPC is the bottleneck for growth, not learning.
You keep skipping the weekly harvest because something else is on fire.

Neither choice is better in the abstract. The worst outcome is the middle ground, hiring someone cheap who runs the same generic playbook across 40 accounts, or trying to DIY while skipping three weeks out of four.

What to do next

Two paths from here. Pick the one that fits your situation.

Run it yourself. Use the framework above, pull your search term reports weekly, and work the harvesting cycle. Keywords.am handles the keyword research, rank tracking, and listing optimisation side of the stack. Sign up for the app and the free tier covers everything you need to start.

Hire a partner. If you want someone who already runs this framework across dozens of accounts, our partner directory is vetted specifically for Amazon PPC work. Browse vetted PPC partners or submit a partner application if you’re an agency that fits the framework.

Either way, the strategy doesn’t change. Only the hands on the keyboard do. For more on the service side, the main services hub lists the current catalogue.

This is operational research based on common patterns across Amazon advertising accounts, not legal or financial advice. Platform rules, auction dynamics, and category benchmarks change. Sellers with active account-health or policy issues should consult a qualified specialist before acting on any framework.