This is part of a series on the Low Data Google Ads Optimisation Framework. Start with the flagship post for the full picture.
Here’s a piece of Google Ads advice that sounds completely wrong until you think it through: if you’re managing a low data account, stop segmenting.
I know. Everything we’ve been taught says the opposite. Tighter ad groups mean better ad relevance. More granular campaigns mean higher quality scores. Segmentation is how you show Google exactly what you want it to do.
All true — when you have the data to support it.
When you don’t, fragmentation is one of the fastest ways to make a struggling account perform even worse.
Why fragmentation kills low data accounts
Google’s machine learning needs conversion data to optimise. The more it has, the better its decisions. The less it has, the more it guesses.
When you split a low data account into multiple campaigns — each with its own budget, its own keywords, its own learning curve — you’re not multiplying your data. You’re dividing it. Instead of one campaign with 20 conversions a month to learn from, you’ve got five campaigns with four conversions each. None of them have enough signal to do anything meaningful with.
Fewer campaigns equals faster learning. That’s the rule for low data accounts, and it’s the opposite of what you’d do with a high-volume account.
The garden centre that structured its campaigns like its website
I see this pattern regularly, and it’s almost always well-intentioned. A client takes the logical structure of their website — pages for every product or service — and mirrors it in their Google Ads account. One campaign per offering. Seems organised. Seems thorough.
In practice, it can be a disaster.
I worked with a garden supplies company that had done exactly this. They stocked all kinds of products: compost, tools, planters, seeds, bulbs — you name it, they had a page for it, and a campaign for it. The problem was their total daily budget was £10 across the whole account. Once you divided that across every campaign, some were running on under a pound a day. At those levels, they weren’t generating any impressions worth speaking of, let alone clicks or conversions. The data simply wasn’t there.
When I dug into the conversion data they did have, the picture was clear: people weren’t searching for “buy compost online” or “garden planters.” They were searching for “garden supplies near me,” “garden centre [location],” “local garden shop.” Generic terms. The granular campaign structure wasn’t matching how their customers actually searched.
We consolidated everything into a single, simpler campaign. Conversion data started coming in, CPAs dropped, and the client — who had been frustrated for months — finally saw results.
The segmentation instinct is hard to resist
The temptation to add structure is real, and I understand it. When an account isn’t performing, it feels like doing something. Creating a new campaign, splitting out a new ad group — it feels proactive.
But with low data accounts, more structure usually means more problems:
Each new campaign needs its own budget. Split a modest budget five ways and nothing has enough to learn.
Each campaign goes through its own learning phase. Every time you make a significant change — new campaign, new bid strategy — Google resets. For a low data account that might take weeks to gather meaningful conversion data, resetting the learning phase repeatedly is costly.
Competitor campaigns are a common culprit. I often have clients who want to bid on competitor terms because their competitors are bidding on theirs. Sometimes that makes sense. But it means a separate campaign, a split budget, and often very high CPCs for clicks that convert poorly. And more often than not, the competitor they want to target has ten times the budget and can sustain that strategy far more easily. Don’t pay to play someone else’s game if you can’t afford the stakes.
What to do instead
Start consolidated and earn your segmentation. Begin with the simplest structure that covers your core offering. One campaign, tight keyword list, sensible budget. Once conversion data starts flowing and you can see where it’s coming from, then you can consider whether splitting makes sense — and you’ll have data to guide that decision rather than gut feel.
Let the data tell you where the conversions actually come from. In the garden centre example, the keyword structure the client thought would work bore no resemblance to how people were actually searching. Your search terms report will show you the reality. Use it before you build your structure, not after.
Resist the pressure to add campaigns for the sake of coverage. More campaigns don’t mean more reach. They mean more budget dilution. If something isn’t in your keyword list, that’s a decision you can revisit — but adding a whole new campaign with a fraction of your budget isn’t the answer.
For brand new campaigns, consolidation matters even more. You have zero history, zero data, and Google has nothing to go on. Give it the best possible chance by keeping the structure simple and the budget concentrated. One campaign learning from 50 clicks is infinitely more useful than five campaigns learning from 10 clicks each.
The mindset shift
Managing a high data account and managing a low data account require genuinely different instincts. On a big account, the marginal gains from better segmentation, tighter ad groups, and more granular targeting add up to something significant. The volume is there to support the complexity.
On a low data account, complexity is a liability. Every extra campaign is another place for your budget to get lost and your data to get diluted.
The goal isn’t to build the most sophisticated account structure. The goal is to give Google enough signal to do its job — and the best way to do that is to keep things consolidated until the data justifies doing otherwise.
Preserve the signal you have. Don’t dilute it.
Next in the series: Mine Intent from Your Search Terms Report — how to use your search terms data to find positives, spot patterns, and guide Google when conversion data is thin.
Leave a Reply