Granular campaign structure sounds sensible, and we're big fans of it here at Bidnamic, as you'll see across most of our content. The more segmented your Google Ads account, the more control you should have: less generalisation, and nothing slipping through the gaps. Separate campaigns for categories, margins, brands, products, regions and customer types can all feel like a smart way to manage performance.
But there's a trade-off. If you segment too far without enough conversion data, performance can look weaker than it really is, simply because each segment is too small. Sometimes you're better off grouping products and categories together to see their true performance. Let's break it down.
Why advertisers like granular structures
Granular structures give teams more control over, and insight into, what is driving their ad spend. They make it easier to set budgets, targets and reporting by product group or business priority, right down to the finest detail. For ecommerce retailers, this is useful when planning separate strategies for bestsellers, sale products, high-margin items, new arrivals or seasonal categories.
The catch is that Google's bidding systems need data to learn and report properly. If each campaign has too little conversion volume, the system may struggle to make reliable decisions, and so will you.
The data density problem
Smart Bidding works best with more conversion data. When you split one large campaign into ten smaller ones, you may gain product-level control (which we're all for), but you lose the data density needed to make clear decisions, because each campaign now has fewer conversions to learn from.
That can make bidding less effective, and you may start to see:
- Slower learning
- Volatile performance
- Overreaction to small changes in data
- Budget trapped in low-volume campaigns
- Strong products held back
- Weaker forecasting
- Less reliable automation
Granularity is only useful if the structure has enough data to support it. A handful of clicks and sales for a very small product group might be perfectly normal, but it's hard to make confident decisions from it.
The 15 to 30 conversions guide for granular campaigns
A common rule of thumb is that campaigns need enough recent conversions to give Smart Bidding a meaningful signal. Many advertisers use 15 to 30 conversions as a guide when deciding whether a campaign has enough data to stand alone. It isn't a perfect rule, but the principle holds: if a campaign doesn't generate enough conversions, splitting it further may reduce performance rather than improve control.
When segmentation works well
Segmentation can work well when there's a clear strategic reason and enough data behind it. For example, when:
- High-margin products need different targets
- New customer campaigns need separate goals
- Seasonal products need dedicated budgets
- Bestsellers need protecting
- Low-performing products need tighter control
- Different categories behave very differently
The key is to segment around meaningful commercial differences, not simply because more structure feels more sophisticated. Where your revenue and business priorities lie, use more granularity.
