Marketing KPI Design · Funnel-Stage Metrics Framework
Marketing KPI Design by Funnel Stage
What is marketing KPI design?
- Three axes at every stage: volume, conversion rate, speed
- Write the definition in one sentence that Sales agrees to
- About 7 KPIs per team; individuals own only what they can influence
When you design KPIs by funnel stage, the first thing to change is the conversation in the marketing team's weekly meeting. If there are only two metrics, 'website visits' and 'won deal value,' no one can explain what happened in between. With a metric at each stage, you can see which stage the numbers leak from, and the meeting shifts from 'why did it fail?' to 'which stage should we fix?' Write down the number of stages and each stage's definition in the same wording Sales uses. That is where the design starts, and if the two don't match, no metric will be trusted.
The funnel is usually set up in five layers. Awareness covers impressions, visits and search traffic; interest covers content downloads, webinar registrations and newsletter subscriptions; the lead stage covers the number of MQLs and the pass rate against the lead-scoring threshold; the opportunity stage covers the number of SQLs sales has accepted and the acceptance rate; and the closing stage covers the number and value of deals that started with marketing. For each layer, it works best to set metrics along three axes: volume, conversion rate and speed: how many leads there are, what percentage move to the next stage, and how many days the move takes. Marketing ROI is the result at the very end of this framework, converted into money, and KPI design is the work of choosing the intermediate metrics that come before it. If you draw it as a tree, with two or three lower-level metrics under each top-level metric, it also fits your reporting structure.
Three criteria decide which metrics to keep: whether the marketing team can actually move the number, whether it is statistically linked to sales results, and whether it can be pulled the same way every week. Keep metrics that can change behavior, not metrics that merely look good. When you set next year's KPIs for the year-end evaluation season, write each metric's definition, formula, data source, and owner on a single KPI definition sheet. A metric with no written definition will inevitably end up in a dispute over interpretation at year-end.
The point where things most often go wrong in practice is having too many metrics. Five per stage, twenty-five in all, and no one looks at them; once they feed into individual evaluations, you get behavior that only inflates what can be counted. If the goal is the MQL count, people lower the score threshold to manufacture MQLs, and Sales sends them back saying 'only useless leads come through.' If you track only top-of-funnel exposure metrics, the opposite happens and only ad spend grows. Cut team KPIs to about seven, including stage-to-stage conversion rates, and have individual evaluation metrics take just the two or three of those that the person can move; that is the approach that lasts in practice. When you change a standard, record the date of the change so comparisons with the prior year don't cause confusion.
John KimInstructor John Kim, author of "B2B Marketing Design" and CEO of Nanum Consulting
Jongsung EunExpert in AI, marketing, business models, and management strategyFunnel-stage metric frameworks took hold in practice when classic buying-stage models like AIDA were combined with data from marketing automation and CRM. Once counting leads separately as MQLs and SQLs became widespread, managing marketing by conversion rates between stages became something like a standard in B2B organizations.
Consider a hypothetical case. Company A sells enterprise security solutions. It has 150 employees, KRW 28 billion (≈ US$20,948,700) in annual revenue, and a six-person marketing team. The team’s KPIs were just two: website visits and closed-won value. In the first half, visits rose 40% while closed deals stayed flat, and no one could explain why.
- Choi, head of marketing, sat down with the head of Sales, drew the funnel in five stages, and wrote the definition in a single shared sentence: “An MQL is a lead with a score of 60 or higher and a confirmed contact job title.”
- When they recounted the past year's data stage by stage, the visit-to-lead conversion looked healthy at 1.8%, but Sales' acceptance rate for MQLs was only 31%, which exposed where the numbers were leaking.
- Choi set metrics along three axes at each stage (volume, conversion rate, and speed) but cut the team KPIs to seven, and gave team member Jung just two metrics within Jung's control, such as 'webinar registration to MQL conversion rate.'
- They wrote the formula, data source, and owner on the KPI definition sheet and signed it together with the Sales division. Each MQL that Sales rejected was tagged with one of three reason codes, and the team reviewed them together at the Monday meeting every week.
- Four months later, Sales' acceptance rate rose from 31% to 58%, and with the same ad spend, the number of opportunities passed to Sales grew from 14 to 26 a month.