Amazon PPC in the Rufus era: 5 structural fixes before you touch bids
Rufus changed how buyers search. Most Amazon PPC accounts are still built on 2024 assumptions — and their ACoS is quietly climbing. Five structural changes that matter more than bid adjustments.
By WAYAMZ Team
We keep seeing the same pattern on new accounts: ACoS is creeping up month over month, the team responds by raising bids and daily budget, and the curve stays the same. The bids aren’t the problem. The structure is.
Rufus (Amazon’s AI shopping assistant) has shifted how buyers actually type — and the PPC scaffold most accounts are running on was built for a pre-Rufus world.
Five fixes, in order of how much they move the needle:
1. Keywords: stop feeding the short heads
Rufus-era queries are conversational. A buyer asking Rufus “lightweight marathon shoe for rainy days” converts roughly 3× better than the same buyer typing “running shoes” into the search bar — because the intent is explicit, not broad.
If your manual campaigns are still weighted toward short-head keywords, you’re paying a premium for the generic traffic and missing the long-tail natural-language queries that actually close.
Fix: Audit your keyword list. Anything under 3 words gets a fresh look — is there a long-tail version of this intent you should be bidding on separately?
2. Campaign structure: separate brand, non-brand, competitor
If your brand term, non-brand discovery terms, and competitor-conquest terms live in one campaign, your budget allocation is lying to you. Amazon’s auto-optimizer will pour spend into whichever bucket is easiest to win — usually brand defense — and starve the discovery bucket that’s supposed to grow the business.
Fix: Three campaigns minimum: Brand / Non-brand / Competitor. Each with its own daily budget. Data stays clean, you actually see where ROAS comes from, and you can throttle one without cannibalizing the others.
3. Match types: don’t mix them
Exact, Phrase, and Broad intent-mean completely different things. Putting them in the same ad group obscures performance — a broad-match winner masks exact-match losers, and you can’t tune bids with confidence.
Fix: One match type per ad group. If you want a keyword running on all three, it gets three ad groups.
4. Day-parting: not every hour converts the same
Conversion rate across the day is not flat. Mornings are browsers, evenings are closers, weekends skew differently by category. Running 24/7 at one bid is leaving money on one side of the day and wasting it on the other.
Fix: Pull hourly conversion data from the last 30 days. Identify your top 4–6 converting hours per day. Concentrate budget there. Lower bids or pause entirely during the hours that don’t convert for your category.
5. Sponsored Display retargeting: the line most accounts don’t run
The buyer who saw your detail page but didn’t buy is radically cheaper to convert than a cold impression. Sponsored Display retargeting runs at a fraction of the CPC of discovery traffic and closes out the mid-funnel demand you’re already paying for on the top.
Yet most accounts we inherit have zero Sponsored Display spend.
Fix: Allocate 15–25% of your PPC budget to Sponsored Display views-remarketing and purchase-remarketing against your own ASINs. It’s one of the highest-leverage structural moves in the account.
The read
2026 PPC isn’t about spending more. It’s about having the structure to know where to spend. Doubling a budget on top of a broken scaffold just doubles the waste. Fix the five above, then talk about scale.