KPIMarketing

A/B Test Velocity

Number of A/B or multivariate tests run across campaigns, emails, and landing pages in a period, measuring marketing experimentation cadence

What is A/B Test Velocity?

A/B test velocity is the number of controlled experiments a marketing team starts in a given period. It counts A/B tests, where two versions of something are compared, and multivariate tests, where several elements are varied at once, across campaigns, emails, landing pages and other channels.

It measures how quickly a team is learning rather than how well any single campaign is performing. Teams that run more well-designed experiments find what works sooner, and that learning compounds over time into better conversion rates and lower acquisition costs.

How this KPI is measured

Function: Marketing
Sub-function: Campaign Performance
Industry: All Industries
Type: Leading
Impact: Operational
Unit: #
Measurement frequency: Monthly / Quarterly
Formula: Count of A/B or multivariate tests started in period

How to calculate it

Count every test that was started during the period, whichever channel it ran in. A test counts once, when it launches, regardless of how long it runs or how it turns out.

Decide in advance what counts as a test. A valid experiment has a clear hypothesis, a control and at least one variation, a defined success metric and a fixed audience split. Changes made without a control, such as a redesign that replaces the old page for everyone, should not be counted.

Most teams track the figure monthly to see the pace, and review it quarterly alongside the share of tests that produced a clear winner.

Where it is used

Building an experimentation culture
Set a target number of tests per month so that experimentation becomes a routine part of how the team works.
Spotting bottlenecks
A low count often points to constraints in design, development or analytics support, or to a backlog of ideas that has never been prioritised.
Reporting on team effectiveness
Show leadership how much the team is learning, which is a leading indicator of future improvements in conversion and revenue.

How to improve it

Keep a ranked backlog of ideasScore test ideas by expected impact, confidence and effort so the team always has a strong next test ready.
Remove the slowest stepFind where tests stall, such as design, build, approval or analysis, and shorten it with templates or standing approvals.
Run tests in parallelUse different audiences or channels so several tests can run at the same time without interfering with one another.
Standardise the set-upUse a shared template for hypothesis, sample size and success metric so each new test takes less effort to launch.
Review results on a scheduleHold a regular session to record what was learned and feed it into the next round of ideas.

Tips and common mistakes

Track quality alongside the count. Many low-quality tests are worth less than a few well-designed ones.
Check that tests reach enough people to give a reliable result. An underpowered test wastes the slot it occupies.
Record the outcome of every test, including those with no winner. A negative result is still learning.

More Marketing kpi library

← All kpi library

Build better plans, faster

Join the waitlist for early access to curated insights, flexible templates, and rapid iteration across strategy, planning, and project management.