First, decide what decision needs to be made and then choose indicators to support that decision. Rather than the name of the indicator, the team needs a shared understanding of its calculation method, scope, review schedule, and owner.
Link goals and observations
If your goal is to increase B2B inquiries, your judgment will be different when you look only at the number of form submissions and when you look at the number of valid inquiries suitable for a sales conversation. In the case of commerce, even if the number of orders increases, profits may vary if cancellation, refund, or discount costs are reflected. First agree on what results will be considered success.
Document each metric definition
Record the indicator name, calculation formula, data source, inclusion/exclusion conditions, comparison period, review schedule, and person in charge. “Conversion rate” is also a different number depending on whether it is per user or per session. If conditions have changed, it must be indicated whether the previous figures can be compared on the same basis.
Outcome metrics and diagnostic metrics
Sales or valid inquiries show results. Search impressions, content visits, and button clicks can help you see what's causing the change. Rather than increasing all numbers equally, we also look at cost and quality indicators that may move in the opposite direction of our goals.
When using AI exposure as a KPI
First fix the scope of the observed questions, the services, language, and measurement method. Citation rates calculated from our own observations are internally managed metrics and should be differentiated from search engine official scores. We also check the relationship between impressions and actual customer inquiries separately.
Key takeaways
- Choose the metrics you need for your goals and decisions.
- Define calculation formula, scope, source, and person in charge.
- Distinguish between outcome indicators and diagnostic metrics.