RANGE MODEL / GROUP
What behaves differently?
Build groups that can share sensible assumptions. Separate work when search intent, success measures or delivery need a different forecast.
Fictional example. Change a group and see what needs checking.
01 / SORT
Same assumptions—or separate?
Start with a few meaningful bundles of pages or work. A group does not have to match a site folder.
02 / CHECK
Your grouping plan
4 work items → 3 assigned forecast groups. Matching labels are a starting point, not proof that the work behaves alike.
Commercial pages
- Running-shoe category pages
- Running-shoe product pages
These labels match. Check page type, demand, search mix and observed CTR or conversion before sharing assumptions.
Buying guides
- How to choose running shoes
These labels match. Check page type, demand, search mix and observed CTR or conversion before sharing assumptions.
Brand searches
- Shoe-shop brand searches
These labels match. Check page type, demand, search mix and observed CTR or conversion before sharing assumptions.
When should I split or combine groups?
Split when one assumption would hide a meaningful difference: brand versus non-brand search, different conversion behaviour, page types, seasonal demand or launch dates. Combine only when evidence supports comparable behaviour and the extra detail would not change the decision.
Keep assisted conversions separate from direct sales. They may overlap. Combine compatible outputs only for the same period and after checking for double-counting.
03 / TAKE IT FORWARD
Keep the groups with your forecast.
Use a separate baseline and supported assumptions for each group. This plan does not calculate uplift or transfer entries automatically.