A dashboard shows twice as many leads, so the campaign looks successful. Sales then reports fewer useful enquiries and fewer customers. Both reports can be correct. The problem is the choice of KPI: the team rewarded a recorded action without checking whether it supported the business goal. A useful KPI needs a definition, a decision and a limit on what it can prove.
Separate delivery from the change you want
Use three levels to describe the work. Outputs are what the team delivers, such as a new enquiry form or a reporting tool. Outcomes are the changes that follow, such as more suitable prospects completing an enquiry. Business impact concerns the wider value, such as additional contribution from new customers after the relevant costs. The boundaries depend on the project, so write down how you use these terms.
HM Treasury’s Magenta Book separates monitoring delivery and outcomes from evaluating the additional impact of an intervention. That distinction also helps in analytics. A completed form is an observed outcome in a lead-generation process. It does not establish that the form change caused additional profitable customers. Activity measures, such as hours spent building the form, answer another question: whether the planned work happened.
None of these measures is automatically useless. An output measure can confirm delivery. A form completion rate can identify friction. Contribution can support a commercial decision. The mistake is to treat them as interchangeable. First decide which level the team can act on, then keep enough supporting measures to check the assumed connection to the goal.
Check what a higher lead count hides
Consider two hypothetical campaign cohorts. Each has the same follow-up period and qualification rules. Cohort A produced 100 unique leads, 40 qualified opportunities and eight customers. Cohort B produced 200 unique leads, 30 qualified opportunities and six customers. The lead count doubled, but qualified opportunities fell by 25%, from 40 to 30. Customers fell by 25%, from eight to six. These numbers illustrate a metric choice, not a real campaign result.
