Google's official guidance on campaign structure has shifted substantially over the past few years. Where the old best practice involved granular segmentation — one campaign per brand, per category, per match type, per device — the current recommendation is simpler: fewer campaigns, more data per campaign, let the algorithm work.
This advice is correct. But it comes with a significant caveat that Google's documentation glosses over: consolidation helps when you merge similar things. It actively hurts when you merge dissimilar things. The distinction between those two scenarios is where most of the mistakes happen.
Why the Consolidation Advice Is Right
Smart Bidding — Google's umbrella term for automated bid strategies including Target ROAS, Target CPA, and Maximise Conversions — requires conversion data to function well.
The minimum threshold that most practitioners use as a guide is around 30 to 50 conversions per campaign per month. Below that, the algorithm is operating without enough signal to make reliable bid decisions. It can't distinguish which auctions to compete in aggressively and which to pull back from, because it hasn't seen enough outcomes to form a reliable model.
A business running 10 separate brand campaigns, each getting 20 to 30 conversions a month, has a problem. Each individual campaign is data-starved. Merge those 10 into 1 campaign, and suddenly the algorithm has 200 to 300 conversions a month to learn from. That's a qualitatively different dataset — more signal, less noise, better decisions.
This is the logic behind consolidation, and it's sound. Splitting campaigns too finely fragments your conversion data and prevents Smart Bidding from doing what it's supposed to do.
Where Consolidation Goes Wrong
The mistake comes when consolidation is applied as a blanket rule without considering what's being merged.
Imagine a retailer with two distinct product ranges: a premium line with 40% margins and a budget line with 12% margins. Both sell well. Both get conversions. Someone reads the consolidation advice and merges them into a single campaign.
What happens? The algorithm now sets bids based on blended data across both product groups. The target ROAS it's optimising toward is a weighted average of both margin profiles. It doesn't know — because you haven't told it — that the premium products can support a lower ROAS target (and therefore more aggressive bidding) while the budget products need to be held to a tighter return threshold.
In practice, this often results in underbidding on premium products (leaving profitable sales on the table) and overbidding on budget products (spending more to acquire conversions that aren't worth as much). The algorithm is doing its best with the data it has, but the data is blended in a way that doesn't reflect your actual business goals.
Consolidation in this case has made the account worse, not better.
The Principle That Actually Matters
The right principle isn't "fewer campaigns" — it's "consolidate within similar groups."
Products with similar margin profiles belong together. Products with similar funnel stages belong together. Products serving similar audience segments, or with similar conversion value potential, belong together.
Products with materially different economics, or fundamentally different customer journeys, don't.
A practical example: if you sell across five product categories and all five have roughly similar margins and similar average order values, consolidating to one or two campaigns is likely sensible. If one category has an average order value of £30 and another has an average order value of £300, those two categories have different economics and probably shouldn't be optimised toward the same target — even if merging them would technically improve the conversion count.
The question to ask before any consolidation is: would a single ROAS or CPA target make sense for everything in this proposed campaign? If the answer is yes, consolidation is likely beneficial. If the answer is no — if different products in the merged campaign actually require different return thresholds to be profitable — consolidation will cause the algorithm to optimise toward an average that doesn't serve either group well.
Segmenting Without Splitting: A Better Approach
There's a third option that often gets missed in the consolidation debate: using campaign structure to guide the algorithm rather than to segment it.
Product custom labels, audience signals, and asset group structure within Performance Max allow you to communicate information to the algorithm without creating separate campaigns. You can label your premium products as a distinct group, give them their own asset group with specific creative and audience signals, and let Google understand the difference — while keeping all the products within a single campaign that's getting enough total conversion data to function well.
This approach — segment, don't separate — often delivers the benefits of both strategies. The algorithm has sufficient data at the campaign level. Your strategic intent about different product groups is communicated through structure and signals rather than through campaign fragmentation.
It's more sophisticated to set up, but it scales better than either pure consolidation or excessive splitting.
A Simple Test to Run on Your Current Account
Take your lowest-performing campaign — the one with the worst ROAS or the highest CPA relative to your targets — and check how many conversions it's getting per month.
If the answer is fewer than 20, you've likely found the problem. The campaign is operating in a data environment that Smart Bidding can't work with. It's making bid decisions based on a sample size too small to be reliable.
The solution isn't necessarily to consolidate it with every other campaign in the account. The solution is to merge it with a related campaign — similar products, similar margins, similar customer intent — so that the combined entity has enough data to make good decisions.
If the low-performing campaign is getting 20+ conversions but still underperforming, the issue is elsewhere: target settings, product quality, landing page, feed quality, or something structural. Consolidation won't fix that.
The Underlying Point
Google's consolidation advice was a reaction to over-segmentation — accounts that were so fragmented that no individual campaign ever had enough data to perform well. In that context, consolidation was the right correction.
But "fewer campaigns" was never meant to mean "one campaign for everything." The nuance that got lost in simplification is that the algorithm needs coherent signals, not just volume. Merging campaigns with incompatible economics doesn't give the algorithm more useful data — it gives it more confused data.
Consolidate thoughtfully, within logical groupings. Keep separate the things that genuinely require separate optimisation targets. Use structure and signals to communicate strategy rather than relying on campaign fragmentation to do the job.
For a review of your current campaign structure and consolidation opportunities, the Shopping campaign management service covers full account analysis and structural recommendations.