Three months on: What the Google Ads target ROAS change taught us.

What we saw after moving from targets to Maximise Conversion Value.

In August we wrote about changes to Google’s Target ROAS and Target CPA bidding which came into effect on 17th August, and why campaigns that were outperforming their targets needed attention.

Before 17th August, campaigns could significantly outperform their ROAS or CPA targets. Following the update, Google is optimising more consistently towards the target you have set. This means a campaign that has historically exceeded its target could begin moving closer towards it.

Since then, we’ve been putting that advice to the test across three of our ecommerce accounts. In this blog, we’ll cover:

• What we changed, and why
• What happened to volume, ROAS and CPA
• Why the three accounts told three different stories
• How we’re deciding where to reintroduce targets

What we changed.

Rather than guess at new targets, we moved campaigns to Maximise Conversion Value (or Maximise Conversions where it made more sense). The aim was to hold conversion volume steady and build a clean set of post-update data. Once we had that, we could set Target ROAS and Target CPA only where the numbers showed a campaign could hit them without losing volume.

Volume held, but efficiency moved.

Across the three accounts, conversions rose from 1,178 in August to 1,249 in October, an increase of 6%. Spend over the same period rose by 16%. That pushed blended cost per conversion from £11.20 to £12.27. Blended ROAS fell from 846% to 561%. That drop is largely down to one account, which we cover below. So, the first lesson is that removing targets did what we intended: volume was maintained, and in two of the three accounts, it grew. But with no efficiency target in place, spend and returns drifted.

Three accounts with three different stories. 

Account A: the one to watch. Spend, conversions and conversion value all fell over the three months, and cost per conversion rose by 36%. ROAS held up better than volume did, which suggests the account may need more than a bidding change. We want to understand why before putting a target on it.

Account B: efficiency gained but value fell. This account was running at 1,149% in August and peaked at 1,445% in September. That is the situation our original article described, with a campaign far above any sensible target. By October, cost per conversion had fallen by 39% and conversions were up a third. But conversion value dropped sharply because average order value fell, from roughly £188 in September to around £63 in October. [Add the reason, for example product mix, promotions or seasonality. This is why ROAS on its own can mislead. The account became cheaper to convert and still looked far less impressive on ROAS.

Account C: volume at a price. Spend almost doubled between August and October. Conversions grew by 22% and conversion value by 64%, so the campaigns clearly found more demand. But each extra conversion cost more. Cost per conversion rose from £10.02 to £15.83, and in October conversions were flat on September despite spend rising by 80%. This is the sort of account where a target stops spend running ahead of returns.

How we're deciding where to reintroduce targets.

The data gave us a simple test for each account. Has it held volume, and is the efficiency it's delivering something the client would genuinely be happy to repeat?

• Account A stays on Maximise Conversion Value for now while we investigate the volume decline.
• Account B has the clearest case for a target. A cost per conversion of £9.41 and growing conversions gives us a solid base. We'll set the target below recent performance, not at it, to leave headroom.
• Account C is the strongest case for tightening efficiency. We'll introduce a target to pull cost per conversion back towards where it was in September, then reassess volume.

What this means for your Google Ad campaigns.

Three accounts is a small sample, and the picture will vary by sector, seasonality and budget. October also leads into peak trading, so we'll keep watching as Q4 develops. But a few of the learnings apply to almost any account:

• Dropping targets can protect volume but expect efficiency to move.
• Look at ROAS, cost per conversion and average order value together, because looking at them in isolation can mislead.
• Set targets from performance you've achieved over a meaningful period, not from the number a campaign launched with.
• Not every account is ready for a target at the same time.

Need help reviewing your Google Ads bidding?

If you're unsure whether your campaigns are ready for a target, or whether a target is still the right approach, our team can review your account and help you decide. Get in touch to talk through your Google Ads strategy.