Search ROI is monthly uplift divided by monthly cost. Model uplift from just two levers — search usage rate and search conversion rate — holding visitors and order value constant. Everything else, from support savings to retargeting efficiency, stays out of the number so the claim remains defensible.
Baseline search revenue is visitors × search usage × search conversion × average order value. Apply a target usage rate, then a target conversion rate, and the difference from baseline is monthly uplift. ROI is that uplift over plan cost; payback is cost divided by uplift, expressed in days.
| Input | Where to find it |
|---|---|
| Monthly visitors | Analytics — sessions |
| Search usage rate | Analytics — site search behaviour |
| Search conversion rate | Analytics — search-session segment |
| Average order value | Analytics — sales |
Conversion improvements are relative and bounded — a strong relevance project might move search conversion by tens of percent. Usage improvements are absolute and often multiples, because most sites start from a search bar hidden behind an icon. Moving usage from 3% to 20% changes the revenue base itself.
Reduced support volume, lower retargeting spend, better paid-traffic efficiency and catalogue insight from the query log are all real and all harder to attribute cleanly. Leaving them out keeps the headline number conservative — and means the actual return is likely higher than the model shows.