When to Pause an Amazon PPC Campaign (and When to Keep It Running)

Every PPC account has at least one campaign sitting in a grey zone — spend is going out, sales aren’t quite there, and you’re staring at the dashboard trying to decide whether one more week will tell you something, or whether you’re just feeding money into a keyword that was never going to convert.

Most advice treats this as a clean math problem: hit a target ACOS, kill it. In practice it’s rarely that tidy. I’ve run campaigns that lost money for five months before turning into one of a business’s best-ranked keywords, and I’ve killed campaigns in week two that never had a shot no matter how long they ran. The difference wasn’t patience or impatience — it was whether anyone actually checked the signals that mattered, and whether the business could genuinely afford to wait.

Short answer: there’s no universal ACOS or time threshold for killing a campaign. The right call depends on your conversion rate and click-through rate trend, whether the campaign is still surfacing usable data, and — the part most people skip — whether your margin and cash flow can actually absorb the wait. Here’s the framework I actually use, with two real cases where it went both ways.

The framework: what actually decides kill vs. bleed

There’s no single number — it depends on what the campaign is doing and the context around it:

• Auto/broad, pure research — if you’ve already harvested the profitable keywords and this particular term isn’t converting, kill it.
• Auto/broad still surfacing new keywords, with some profitable and some not, and terms graduating to exact actually performing — let it bleed.
• Exact match — judge it against product maturity and positioning. A premium, five-star, higher-priced product shouldn’t be chasing the highest-volume, most generic head term in its category. Wrong fit, and it’ll drain budget regardless of how much data you let it gather.
• A term you tested deliberately and it failed — kill it.

Conversion rate and click-volume thresholds shift with product price, niche, average niche conversion rate, and margin — there’s no number to borrow from someone else’s account.

Price also changes how fast you get an answer. A £300 outdoor post box can afford to bleed a bit longer — volume is low, but one or two sales completely changes the picture. A £9 water bottle gets you an answer much faster, because click volume is so much higher. That’s not really a choice you’re making — it’s just the math at that price point.

If you’re deliberately bleeding a campaign to build organic rank, don’t just watch ACOS — track actual rank movement. DataDive’s Rank Radar is built for exactly this: it tracks daily organic rank alongside your PPC data, so you can see the two side by side rather than guessing whether the spend is actually moving the needle. If a month or two passes with no rank movement despite the spend, more patience isn’t the fix. Go audit the listing itself: price positioning, imagery, missing keywords. You cannot PPC your way out of a listing that isn’t pulling its weight.

When bleeding it out paid off

One case: an exact-match campaign on “electrolyte sachets” — high search volume, the main head term for that category.

This only worked because of who the client was. They had real funding behind them and weren’t chasing short-term growth, so losing money on individual ad sales, or roughly breaking even overall, was something they could genuinely absorb. That’s not a green light for every account — it was specific to a client with that kind of runway.

We let the campaign run continuously for three months with nothing to show for it in rank. By months four and five, it climbed to somewhere between rank one and rank three for the keyword. Four to five months before the payoff showed up — and it only worked because the client’s financial position could support the wait.

When killing early was the right call

Earlier in my advertising work, I ran a campaign for a keyword that was genuinely relevant and a strong fit for the product — this wasn’t the “wrong keyword for the product” problem from the framework above. The issue only showed up once we zoomed out and looked at price and review count against the competition: the conversion rate was never going to get high enough to sustain the rank we were chasing, no matter how long we let it run.

The client would have been fine with breaking even — same tolerance as the electrolyte sachets case. But this campaign wasn’t breaking even. It was losing money and making no ground, so we cut it.

The deeper lesson here is different from the bleed case above: even if we had somehow paid our way to rank one, we’d have had to keep paying forever. Pull the ad spend back and the organic rank falls off instantly, because that rank was never earned through organic conversion strength — it was rented.

Before committing to a long bleed for rank, be honest about whether price, reviews, and competition actually make your target conversion rate achievable. Relevance alone isn’t enough.

The most common mistake

Two related mistakes, not one.

The first: letting exact or broad match campaigns bleed too long without actually checking the signals — conversion rate and click-through rate — that tell you whether continuing is justified. People let campaigns run on gut feel or attachment to the keyword rather than the data.

The second: not being honest — with the client, or with your own numbers — about whether margin and cash flow actually support bleeding. This cuts both ways. Killing too fast because cash flow genuinely doesn’t allow patience isn’t a mistake — that’s a legitimate constraint. But killing too fast despite having plenty of margin or no real cash flow pressure is the mistake: the patience was affordable, and it got abandoned anyway, usually because nobody ran the numbers to check.

The mistake isn’t bleeding or killing as a category. It’s making the call without checking the signals, and without an honest read on whether your financial position supports the choice either way.

FAQ

How long should I let a losing campaign run before killing it? No fixed number — it depends on product price, whether it’s still surfacing usable data, and whether your margin and cash flow can absorb the wait. A £9 item tells you fast; a £300 item can take weeks.

What conversion rate should I use as a kill threshold? There isn’t a universal one. Judge the campaign against your own account’s baseline for similar products, not an industry figure.

Is it ever worth bleeding a campaign just to build organic rank? Yes — but only if you’re tracking actual rank movement, not just ACOS, and the underlying price, reviews, and competition can sustain that rank once ad spend stops. If they can’t, you’re renting the rank, not earning it.

My client wants to keep spending on a losing keyword — is that ever reasonable? Only after an honest margin and cash flow conversation. If breakeven or a temporary loss is genuinely something they can absorb, patience can pay off. If it isn’t, killing early isn’t a failure of nerve — it’s the right call.

What’s the clearest sign a campaign should be killed, not bled? Conversion rate and CTR that stay flat no matter how much data accumulates. That’s not a “give it more time” signal — it’s a listing, pricing, or positioning problem that more ad spend won’t fix.