A launch does not end after four months. It ends when your margin passes your TACOS

Every seller who has run Amazon advertising for more than a quarter has been asked the same question by somebody: what is a good ACOS? The honest answer is that the question is missing a word. Good for what, and good when.

An account that launched three weeks ago and an account with four thousand reviews are not playing the same game, and a number that would be a disaster in one is a success in the other. So we stopped answering with a number and started answering with a phase.

There are three. What follows is what we plan against in each one, and the two events that move an account between them.

Before the numbers, the thing that matters most about them: these are planning ranges, not measured outcomes. They are where we set targets and what we expect to manage through, on our own model. They are not a study of the marketplace and not an average of anybody's results.

1. Launch2 to 4 months

   Profit

   Negative

   TACOS

   30–50%

   ACOS

   50–90%

   Reviews

   0–100

   Ends when: Margin passes TACOS

2. Growth6 to 24 months

   Profit

   Breakeven to 20%

   TACOS

   20–30%

   ACOS

   30–50%

   Reviews

   100–1,000

   Ends when: Position holds without the push

3. Profit24 months and beyond

   Profit

   15–40%

   TACOS

   10–20%

   ACOS

   20–40%

   Reviews

   1,000 and up

The phases are conditions, not dates

The months under each phase above are the ranges we tend to see. They are the least useful part of the picture and the part most likely to be wrong for your product, because a twelve dollar consumable and a four hundred dollar appliance do not move at the same speed and never will.

What actually ends a launch is a condition, and it is visible on the chart as the moment the orange band crosses zero.

Your product margin passes your TACOS. Below that line, advertising is buying you a position you could not otherwise hold, and you are paying for it out of capital. Above it, the same position starts paying for itself. Nothing about the calendar changes that; the only things that change it are the cost of the traffic going down, the conversion going up, or the margin going up.

The second event is quieter and easier to miss. The position holds without the push. When you can reduce spend on a ranking campaign and the organic position does not slide back within a couple of weeks, the rank has stopped being rented. That is the point at which spending less is a strategy rather than a risk.

Launch

The goal is not profit. The goal is information and position.

An ACOS of 70% in the first eight weeks is not a failure, it is the price of finding out which phrases convert for a product nobody has bought yet. The mistake we see most often is a seller cutting that spend in week three because the number looks frightening, and thereby buying no information at all.

The work runs in three moves.

Start. Identify the five phrases the product genuinely competes for, then build the full semantic core around them, usually eighty to two hundred and fifty terms. Our keyword analyzer is free and reads a term the way this step needs it read: how many products compete, who holds the top ten, and what they charge. Set up the campaigns, launch, and get the first reviews moving.

Optimise. Read what actually happened. Find the phrases that convert rather than the phrases that looked obvious in a spreadsheet, double the effort behind the ones worth ranking for, and feed the search term report back into the listing itself. Advertising data is the cheapest listing research there is, and most of it is thrown away. It is also where the difference between a phrase that gets clicked and a phrase that gets bought shows up, and those are not the same phrase more often than sellers expect.

Accelerate. Push budget towards the converting phrases and pull it off the rest, shift the emphasis from paid to organic sales, and drive TACOS down towards the product's margin. When it gets there, the launch is over.

Through all of this we test the traffic sources available inside Campaign Manager and stay off external traffic, unless the plan is deliberately aggressive and funded for it.

Growth

The goal changes from finding a position to taking share.

The account is now profitable, which means every decision has a second cost: money spent here is money not taken. The focus narrows to holding the positions that matter, growing the half-relevant phrases that are cheaper to win, and improving the things that lift every campaign at once rather than one of them.

Whether that work is done by a person or by software is a real decision at this point, and we have written about where each one stops being the right tool.

This is where external traffic earns a place. Live streams, DSP, working with influencers: tested properly, one at a time, with the account stable enough that you can tell what moved and why. Sponsored Products carries the weight of holding keyword positions while that testing happens.

And this is where reviews stop being a count and start being a rate. Holding a rating at 4.5 and above, and converting a higher share of sales into reviews, does more for advertising efficiency than most bid work. The listing gets reworked here too, against what the PPC reports say rather than against opinion.

Profit

The goal is to keep the position while paying less for it.

Notice that ACOS can sit higher here than in growth while TACOS is half what it was. That is not a contradiction, it is the whole point of the phase: the paid sales can stay expensive per click as long as they are a small enough share of a much larger total. TACOS is the number that says whether the account is winning. ACOS on its own never did.

The work moves off the campaigns and onto the economics. Supply chain, unit cost, external traffic run for margin rather than for discovery, and every ad type audited for what it still earns. Defending the brand becomes a real line of work rather than an afterthought.

The whole model on one screen

What this model does not tell you

It does not tell you that your account will follow it. We have not published a measurement of how many accounts do, over what period, or in which categories, and until we have, nobody should read these ranges as a forecast, including us.

It does not survive a thin margin. A product at 15% margin and a product at 45% margin cross the TACOS line at completely different levels of advertising efficiency, and the second one can afford a phase the first one cannot.

It does not replace the arithmetic on your own product. The ranges are where we set targets before we know an account. The first month of real data is worth more than any of them, and where the two disagree, the data wins.

And it is one agency's operating model, published because it is more useful in the open than in a deck. If you want the version applied to your own account rather than read in the abstract, that is what our PPC management is. It is not an industry standard and there is no authority behind it beyond the accounts we have run.