Skip to main content

Automated bid optimization

Once you turn on automation for a campaign, AdTrix looks at each target's performance once a day and makes a small adjustment to the bid — lower when the target is overspending, higher when it's underexposed, and no change most of the time. You don't have to click anything. You don't have to watch it.

This page explains what the optimizer is doing behind the scenes, why bids move the way they do, and when you might want to pause it.

Looking for the manual button?

If you want to optimize bids on-demand from the Target Manager (pick presets, review recommendations, apply them yourself), see Bid optimization explained. The two features do the same math from different angles — automation runs daily in the background, the manual flow runs when you click.

Prerequisites​

  • At least one Amazon account connected
  • A Target ACOS set at the account or campaign level
  • Automation enabled on at least one campaign

How to enable it​

See Your first optimization for the full walkthrough. In short, there are two entry points in the Campaign Manager:

  • Per-campaign — click Optimize Campaign in the Automated column, set the Target ACOS in the drawer, click Enable Automation.
  • Bulk — select multiple campaigns and choose Apply Automation from the Actions menu, or click Activate Automation → above the table to enable automation on every eligible campaign at once.

When it runs, and what data it looks at​

When

Once per day per campaign, between 9 AM and 4 PM UTC (roughly 4 AM to 11 AM Eastern). Each campaign is evaluated once a day, not multiple times.

What data it considers

A 14-day performance window on each target. Why 14 days — long enough to be statistically stable (a typical target needs at least 5 clicks to avoid noise), short enough to reflect recent performance instead of lagging old data.

A couple of signals also look at 30 days — specifically to catch "slow bleeds" on targets that have spent a little bit over a long period with zero sales.

You can sanity-check the data the optimizer sees. Open the Target Manager, set the date range to the last 14 days, and the Spend, Sales, and ACOS you see there are the same numbers the optimizer is using.

What it does each day​

The optimizer asks one question per target: is this bid working? The answer lands in one of three buckets.

The "optimal bid" formula​

When AdTrix talks about the "efficient" or "optimal" bid for a target, it's computing:

Optimal bid = Revenue-per-Click × Target ACOS

Revenue-per-Click is Sales ÷ Clicks over the 14-day window. A target that made $200 in sales from 100 clicks has a Revenue-per-Click of $2.00, and at a 30% Target ACOS the efficient bid is $0.60. Every decrease decision is a move toward that number.

Your bid goes down when a target is wasting spend​

Three cases, same goal — stop the bleeding:

The ACOS is too high. 14-day ACOS is above your Target ACOS. AdTrix moves the bid 10% of the distance between the current bid and the efficient bid.

Worked example: Target ACOS 30%, 14-day data shows $200 in sales from 100 clicks, current bid $1.00. Optimal bid = $2.00 × 30% = $0.60. Step = 10% × ($1.00 − $0.60) = $0.04. New bid: $0.96.

Tomorrow the same math runs on fresh data — the bid keeps stepping down 10% of the remaining distance until the step rounds to zero, landing within a few cents of $0.60 after about 25 days. Small steps protect you from a single noisy week.

It's spending with zero sales. Target has spent more than one full target cost-per-order and made no sales.

Example: Target ACOS 30%, product's average order value is $50 — so one sale is "worth" $15 of spend. Target has spent $18 with zero orders. AdTrix drops the bid to the marketplace minimum (typically $0.02) in one step. No convergence — there are zero sales to protect.

It's been quietly bleeding for a month. Target has spent $3 or more over 30 days with zero sales. Same one-step drop to minimum.

Your bid goes up when a target isn't getting a fair chance​

Two cases. Both use a fixed dollar step rather than a percentage, so high bids don't inflate faster than low ones.

A proven target is getting ignored. A target with a history of sales is now getting far fewer clicks than similar targets. AdTrix adds between $0.01 and $0.10 per run, sized to the current bid.

A brand-new target has had no chance. A target less than 90 days old with zero clicks gets a flat $0.02 per run so it has a chance to gather performance data. If it has impressions but still zero clicks, the step is capped at $0.05.

Your bid stays the same when the data says to wait​

  • The target's ACOS is already close enough to your Target ACOS.
  • The target is less than 10 days old with fewer than 10 clicks — not enough data to act on yet.
  • The bid is already within a few cents of the efficient bid.

Most targets, most days, fall into this last bucket. That's normal — the optimizer is meant to be quiet.

Safety limits​

Here's what AdTrix will not let happen to your account:

LimitWhat it meansWhy
Gradual convergenceDecreases move the bid ~10% of the distance to the efficient bid per dayA single bad week can't halve a good bid. Full correction takes ~25 days.
Daily increase capRoughly $5 of total bid increases per campaign portfolio per dayIf many targets qualify at once, AdTrix raises the lowest bids first (highest discovery value per dollar) and the rest wait for tomorrow.
Dynamic bid ceilingNo target gets raised above roughly 7× your account's typical cost-per-click (with a floor of $5 and ceiling of $20)Stops runaway bids even when the formula suggests higher.
Marketplace limitsAmazon's per-marketplace minimums and maximums (e.g. $0.02 minimum for Sponsored Products) are always enforcedRequired by the Amazon Ads API.

You can trace any bid change back to these rules. If a bid moved, one of the triggers above matched on 14 days of real data.

How it works alongside dayparting​

The automatic optimizer decides your baseline bid — the "true" bid for that target. If dayparting is also on, the baseline gets multiplied by an hourly multiplier before it's sent to Amazon.

If your baseline is $0.50 and dayparting bumps evening hours to 1.5x, your bid on Amazon at 8 PM is $0.75. The optimizer isn't fighting dayparting — they hand off to each other.

Seeing what changed​

Open the Target Manager and look at the Bid column. Use the 14-day date range and sort by Spend or ACOS to find the targets the optimizer worked on most recently.

When to turn it off​

"I'm launching a new product and want to be more aggressive than my Target ACOS allows." Turn automation off on just the launch campaigns, give yourself 2-4 weeks to buy ranking, then turn it back on.

"I'm running a Prime Day or holiday push." Same answer — turn it off for the short window, turn it back on after.

If you turn automation off for months and then back on, the optimizer may make larger adjustments at first while it catches up to current performance. Give it a couple of weeks to settle.

Common questions​

Why did my bid change when I didn't change my Target ACOS? AdTrix reacts to how the target is actually performing, not just to the target you set. If last week's ACOS went up, your bid comes down even if your goal is the same.

Why is the bid moving so slowly? Just set it to the right number. The "right number" depends on performance data that shifts week to week. A small daily step gets you to the right place over a few weeks and never overshoots on a bad week.

Can I have different goals on different campaigns? Yes. Each campaign has its own Target ACOS. Branded campaigns at 15%, discovery campaigns at 40% — that's fine.

Will this ever turn my bid off completely? No. Bids go to the marketplace minimum at the lowest, never to zero. Targets you truly want off, pause manually.

I enabled automation yesterday and nothing's changed. Is it broken? Probably not. Most targets land in the "no change" bucket on any given day. Check again in a week.

What's next​