A Low ACOS on Amazon Ads Isn't Always a Good Thing: You May Not Be Saving Ad Spend, but Losing the Precise Visitors You Should Have Won

Every day, many Amazon sellers and operators open the Amazon ads console and the first thing they look at is ACOS. Campaigns with low ACOS get a quick glance and a scroll past — many people think this campaign is performing well and needs no attention, so they put all their energy into the campaigns with ugly ACOS.

Measured as a single metric, a low ACOS is not a problem in itself: it means ad spend is a small share and the return on spend is high, and it also shows that the targeting keywords and bids are fine. But if you make it the only standard for judging whether an ad is good or bad, and the only number left in your view is this one, then the data you should be analyzing most — impressions, clicks, click-through rate, conversion rate — gets ignored instead.

In my view, talking about whether ACOS is high or low while setting aside the ad's data volume is actually meaningless. ACOS is ad spend divided by ad sales; it is a result data point built from click after click and order after order. If the data volume is not enough, this number may be a false indicator.

For example, a campaign gets just two or three clicks a day and produces 1 order; judged only by the ACOS number, it looks beautiful. But what happens when the data volume scales up? The performance may quickly collapse. So neither a small amount of data nor short-term data can measure well whether an ad is good or bad.

So how should we analyze and judge our ad data?

First we still have to go back to the purpose of Amazon advertising — what advertising is actually for: acquiring precise traffic at scale, so our products show up in front of the right target users. Only when the volume of visitors comes up is there a foundation for steady orders. But ACOS only reflects the ratio between ad spend and sales; traffic scale is something it cannot reflect. Yet if our listing is to grow, that growth actually comes from visitors that keep growing steadily. Once you see this clearly, the cost of the move to blindly push ACOS down becomes obvious.

Generally speaking, the two easiest moves to push ACOS down are lowering bids or lowering budgets.

When the listing's conversion performance is still decent, if you don't dare raise the bid — or even lower the bid — just to make ACOS look good, then our ad placements can't move up, and we can't get higher impressions and clicks; if you don't dare raise the budget, click volume can't reach scale, so the daily visitors are forever just a handful. The end result is that a listing that was performing fine loses the traffic foundation to keep climbing, just to make ACOS look good. When you turn back later and want to scale up, no data or weight has accumulated on the core keywords, so it very easily loses the momentum for sustained growth and finally enters a downward channel, from which it is hard to come back.

So to judge whether ad performance is good or bad, you still have to go back first to the ad's goal and see whether it brought the listing enough precise visitors.

Impressions and clicks are two must-check items, and campaigns with different purposes have to be analyzed separately: for a campaign that picks up long-tail keywords, a low ACOS with a small budget and low bids to pick up long-tail orders — few clicks are perfectly normal.

If it is a campaign pushing the core big keywords, and ACOS looks quite good but there are only a few clicks a day, then the campaign is unqualified — it did not get the listing the visitors it should have gotten, and the growth opportunity simply leaked away.

Often a low ACOS most easily gives people a bit of a false sense of safety: the lower the ACOS, the more unwilling they are to touch it. But that ad spend you didn't spend may not really be saved: the budget that should have been spent wasn't spent, the more precise visitors you should have captured didn't come in, and the users were shut outside the door by you.

So the next time you open the ad report, we should analyze across multiple dimensions instead of looking at one single data point in isolation and then making a judgment and adjustments. When analyzing every campaign, ask one more question — did it bring our listing in front of our target users as much as possible?