Short answer: not yet, and probably not for another month. Target ROAS needs enough conversion history to make a real prediction instead of a guess, and most small-budget accounts flip the switch weeks before they have that history. The rule we actually use is simple: we don’t touch Target ROAS bidding until a campaign has logged at least 30 conversions in the trailing 30 days, and we’d rather see 50. Below that, the algorithm is pattern-matching on noise, and the “optimization” you’re paying for is closer to a coin flip with your budget attached.
That threshold is the part most guides skip, because “it depends on your account” is a less satisfying answer than a number. But the number is what actually decides whether Target ROAS helps or hurts, so that’s where we’re starting.
What Target ROAS Actually Optimizes For
Target ROAS (tROAS) tells Google’s bidding system: “for every euro I spend, I want this much back in conversion value.” The algorithm then raises or lowers bids in real time, trying to hit that ratio across the campaign. It’s not optimizing for volume, and it’s not optimizing for the lowest cost per click — it’s optimizing for a value ratio, which means it needs to know, with reasonable confidence, what a conversion from a given click is likely to be worth.
That confidence comes from data: past conversions, their values, and enough variety in auction outcomes for the model to separate signal from noise. A campaign running for three weeks on €25/day with eight sales doesn’t have that. A campaign running for four months with 200 sales does. Everything else in this article follows from that one difference.
The Data Volume Threshold We Actually Use
This is one of the first things we check when we audit a Google Ads account that’s already running tROAS with disappointing results — nine times out of ten, the target was set before the account had the data to support it.
We look at two numbers before recommending tROAS to a client:
- 30+ conversions in the trailing 30 days at the campaign level — the minimum we’ll consider testing at.
- 50+ conversions in the trailing 30 days — where we’re comfortable setting a target and leaving it alone for two full weeks without touching it.
These aren’t numbers Google publishes as a hard cutoff, and Google’s own guidance is vaguer than that on purpose — it varies by vertical, average order value, and how much conversion value differs between customers. But under 15–20 conversions a month, we’ve seen Target ROAS behave erratically often enough that we treat it as a real threshold, not a suggestion. If your account is under that line, Maximize Conversions or Maximize Conversion Value (with no target set yet) will get you to the data volume faster than tROAS will, because they’re not fighting the algorithm’s own uncertainty at the same time.
What if you sell something with a long sales cycle?
If your average time from click to conversion is measured in weeks rather than days — common for B2B services, high-ticket goods, or anything requiring a quote — stretch the lookback window mentally even if Google’s reporting doesn’t. Thirty conversions spread across 90 days of actual buying behavior is a different situation than thirty conversions in 30 days of impulse purchases, and the algorithm needs the underlying buying pattern to stay reasonably consistent to bid well against it.
When Target ROAS Backfires
We’ve watched three patterns cause the most damage, in order of how often we see them.
1. The target is set too aggressively on day one
The most common mistake isn’t turning on tROAS too early — it’s setting the target based on a margin spreadsheet instead of what the account has actually been achieving. If your account has been converting at a 350% ROAS under Maximize Conversion Value and you set a tROAS target of 500% because that’s what your margins need, the algorithm will cut spend hard to try to hit a number it has no evidence it can reach. Volume collapses, and the campaign looks broken even though the bidding is doing exactly what you told it to.
The fix: set the initial target at or slightly below the average ROAS the campaign has already been delivering, then tighten it gradually — a few percentage points every one to two weeks, not in one jump.
2. High variance in order value
If your average order value swings widely — a service business quoting jobs from €200 to €8,000, or an e-commerce store selling both €15 accessories and €600 furniture in the same campaign — the algorithm has a much harder time learning what a “good” click looks like, because the same click-quality signals can lead to wildly different outcomes. Segmenting these into separate campaigns by price tier, even at the cost of more manual setup, usually outperforms one blended tROAS campaign trying to average across both — it’s the same logic we use when we split a paid media test budget instead of running every hypothesis through one pool of spend.
3. Seasonal spikes right after switching
This is the same reasoning behind what we turn off in Performance Max campaigns before a big sales period — automated bidding of any kind needs stable patterns to learn from, and a spike breaks that pattern on purpose. Turning on tROAS heading into a demand spike — a holiday period, a launch, a press mention — is asking a model trained on “normal” weeks to make good decisions during an abnormal one. It will often under-bid during the exact window you need it to be aggressive, because the historical pattern it’s matching against doesn’t include that kind of demand. If you know a spike is coming, either delay the switch until after it, or budget for the campaign to underperform during it and correct course afterward.
How We Migrate an Account to Target ROAS Without Blowing It Up
- Start on Maximize Conversions (or Maximize Conversion Value if order values differ meaningfully) with no target set, and let it run until you hit the 30-conversion threshold.
- Check the account’s natural ROAS over that period — not your target margin, what it actually delivered.
- Set the tROAS target at or just under that natural number for the first two weeks. This should feel almost too easy. That’s intentional.
- Hold for a full two-week cycle before making any adjustment — Google’s own guidance recommends at least this long, and in our experience shorter windows just mean you’re reacting to daily noise.
- Tighten in small increments — 5–10 percentage points at a time — checking volume and ROAS after each change before tightening again.
The accounts that get into trouble are almost always the ones that skip step 3 and set the “real” target immediately, or skip step 4 and adjust after three days because volume looked soft.
Target ROAS vs. the Alternatives
| Strategy | Data needed | Best for | Main risk |
|---|---|---|---|
| Manual CPC | None | Brand-new accounts, tight manual control | Time-intensive, doesn’t scale bidding intelligence |
| Maximize Clicks | None | Pure traffic goals, awareness campaigns | No connection to conversion quality |
| Maximize Conversions | Low (15+ conversions helps) | Building initial conversion volume and data | Can chase cheap, lower-value conversions |
| Maximize Conversion Value | Moderate | Accounts with varying order values, no fixed margin target yet | No cap on cost per conversion |
| Target ROAS | High (30–50+ conversions/month) | Mature accounts with a real margin target to hit | Punishes thin data and aggressive targets hard |
Common Mistakes We See
- Switching to tROAS the same week a new landing page or offer goes live, so the algorithm is learning two new variables at once.
- Setting one tROAS target across a campaign with wildly different audiences (cold prospecting mixed with warm retargeting) instead of splitting them.
- Panicking and lowering the target after a single bad week instead of a full two-week read.
- Never revisiting the target once it’s set — margins change, seasons change, and a target set in January can be wrong by September.
Frequently Asked Questions
How many conversions do I need before switching to Target ROAS?
We look for at least 30 conversions in the trailing 30 days before testing it, and prefer 50 or more before treating the results as reliable. Below that, stick with Maximize Conversions or Maximize Conversion Value while you build up data.
What happens if I set my ROAS target too high on day one?
The algorithm will cut bids and spend aggressively trying to hit a number it has no history of achieving, and volume will usually collapse within days. Set the initial target at or below the account’s recent natural ROAS, then tighten gradually.
Should I use Target ROAS or Maximize Conversion Value?
Maximize Conversion Value with no target set is the right move if you don’t yet have a firm margin number to hit, or if you’re still building conversion history. Target ROAS makes sense once you have both sufficient data and a specific ratio you need the account to hit.
How long should I wait before judging whether Target ROAS is working?
Give it a full two-week cycle after any change — new target, new tightening step, anything. Google’s bidding systems need that window to relearn, and judging performance after three or four days almost always means reacting to normal daily variance rather than a real signal.