The Lesson: Not Every Keyword Deserves the Same Budget Logic
When a brand decides to go after rank or market share on Amazon, the usual approach is to raise bids across the board, accept a higher ACoS for a while, and wait for organic rank to catch up. The logic sounds fine. Paid sales drive sales velocity, sales velocity drives organic rank, and organic rank eventually makes the whole thing profitable.
The problem is that this logic only holds on some keywords. On others, you can spend as much as you want and the payoff never comes. If you treat both kinds the same, you end up overpaying on the terms that will never return the investment, and often underinvesting on the terms that would.
So I split keywords into two buckets: rankable and non-rankable. That one label changes how much I'm willing to lose on each term, and whether I'm willing to lose anything at all.
How It Works: Two Buckets, Two Rules
Rankable keywords
A rankable keyword is one where you can realistically land in the top ten organic results, because your product is what the shopper is actually looking for. The search intent and your product line up. When someone types that term and sees your listing, it's an obvious fit.
On these keywords, losing money on ads can be the right move. The sales you buy now earn you organic rank, and once you hold that rank, more of your sales come in without ad spend behind them. Your TACoS comes down over time. The early loss isn't waste. It's an investment with a payback you can reasonably expect.
Non-rankable keywords
A non-rankable keyword is one you'll never own. It might be only loosely related to your product. It might be too generic. Or it might be so expensive that even if you could climb, the ranking would never pay back what it cost. You can still make sales on these terms. You just can't take the spot.
Mansour Norouzi, Partner and Director of Advertising at Incrementum Digital, laid out this distinction in a LinkedIn post this summer, and his examples make it concrete. A left-handed mouse targeting "wireless mouse." A cast iron pizza pan targeting "baking sheet." Both can get clicks, and maybe a sale here and there. But the top ten for those searches is never going to belong to them, because most of the people searching aren't looking for that product.
His rule for the second bucket is simple, and I agree with it fully: on non-rankable keywords there's no payback, so every unprofitable sale is just money gone. Never lose money there.
The takeaway: Losing money on a keyword is only an investment if that keyword can be ranked. On rankable terms, a short-term loss buys organic rank and lower TACoS later. On non-rankable terms, there is nothing to buy, so hold them to a profitable target.
How I'd Apply It
This framework is easy to agree with and easy to skip. The work is in actually labeling your keywords before you touch bids. Here's how I'd go about it.
Start with the shopper, not the search volume
For each keyword, ask a plain question: if a shopper searched this, would my product be what they were hoping to find? Not "could they buy it," but "is this what they meant?" High volume makes generic terms tempting, but volume doesn't matter if the shopper behind the search wants something else. If your honest answer is "sometimes" or "only a small slice of them," treat the keyword as non-rankable.
Be honest about cost
Some keywords are a perfect fit and still don't belong in the rankable bucket, because the price of climbing is so high that the organic position would never pay it back. Relevance is necessary, but it isn't enough. If you can't see a realistic path where the rank earns back what you spent to get it, label the keyword non-rankable and treat it that way.
Set targets by bucket
Once every keyword has a label, the ACoS targets mostly set themselves. Rankable keywords get a target that allows a deliberate, time-boxed loss while you push for rank. Non-rankable keywords get a target that has to be profitable on its own, every week. I'd rather keep a non-rankable keyword running at a modest, profitable level than cut it entirely, since those sales still count. They just can't be bought at a loss.
Watch the payback on the rankable side
Accepting a loss on rankable terms only makes sense if the payback actually shows up. Track organic rank and TACoS on those keywords as you invest. If you're climbing toward the top ten and TACoS is starting to come down, the strategy is working. If the loss continues and rank isn't moving, revisit the label. You may have put a keyword in the wrong bucket.
What to Do This Week
- Export the keywords you're currently bidding on for your main products.
- Label each one rankable or non-rankable, based on shopper intent and whether ranking could ever pay back the cost.
- Find any non-rankable keyword running at a loss and bring it back to a profitable target.
- Pick the rankable keywords where you're willing to invest, and decide how long you'll run at a loss before reviewing.
- Track organic rank and TACoS on those rankable terms so you can see whether the investment is paying off.
Want this playbook applied to your brand?
We help brands grow across Amazon, TikTok Shop, Walmart and DTC.
Get a Free Consultation →The rankable vs. non-rankable framework and examples above come from a public LinkedIn post by Mansour Norouzi.