Could the blanks in your CRM flip the winner?
Change the totals and watch the verdict move. Counts only, no names, one currency.
The readout
The apparent winner is not settled.
Possible range
160.00 to 200.00
Possible range
142.86 to 250.00
6 additional outcomes from Campaign B's 15 unresolved records would make it cheaper, if the other campaign gains none.
The range runs from none of the unresolved records converting to all of them. It gives no probability, and it says nothing about profit or what the next dollar will do.
Everything runs in your browser. Nothing you type is uploaded or saved. Normal site analytics still run.
How it works
A range tells the truth. A single number hides the blanks.
Confirmed cost = spend ÷ confirmed outcomes. The lowest possible cost = spend ÷ (confirmed + unresolved). The highest remains the confirmed cost if no unresolved record becomes an outcome. With no confirmed outcome, that upper limit cannot be expressed as a finite cost.
This is a sensitivity check on how complete your attribution is. It is not a confidence interval, and it does not prove one campaign caused more sales. An unresolved record is a lead whose outcome nobody has recorded yet. It is not a lost lead.
- Use the same currency and equivalent outcomes. Do not compare booked calls with paid contracts.
- Give both cohorts equal time to mature. A younger cohort can look worse merely because its outcomes have not happened yet.
- Deduplicate. A record can contribute at most one outcome and belongs to only one campaign.
- Keep value and margin in the next decision. Cheaper outcomes are not automatically more profitable.
Read the worked example and decision method, or use the five-check attribution guide to find the broken link.