Activation is the share of new users who do the one thing that shows they got value from your product. To measure it you need three things: a way to tell users apart, an event that marks the value, and a clear definition of "new". This guide shows how to choose the event and read the number. I am building JustOneCX (disclosure), so the setup uses it, but the thinking applies to any analytics tool.
What is an activation event?
It is one action that means a user got value. Not "logged in" and not "viewed the dashboard". Examples: created a first project, invited a teammate, ran a first report, connected a data source.
A test: if a user did this and nothing else, would you expect them to come back? If yes, it is a good candidate.
How do you choose it?
- List the first actions a new user can take.
- Look at the users you know are happy. What did they do in their first sessions?
- Pick the earliest action they all did.
- If you cannot decide, pick the one you would be most upset to see users skip.
Pick one. You can change it later, but a number with two definitions is not a number.
What do you need to send?
For activation you need two things, from the analytics definitions in the docs:
identify()with a user id, so a person can be counted once.- Your activation event, sent with
track()when the action happens.
import { identify, track } from "@justonecx/sdk";
identify(user.id, { plan: user.plan });
// when the user creates their first project
track("project_created");
How is the number counted?
- A signup is a person's first
identify()call, counted from the day we first saw them. - New identified users are people first seen in the chosen period who have a user id.
- Activated are the ones who then sent the activation event. Events sent before a person first appeared do not count.
- Activation rate is activated users divided by new identified users.
Two details change the number. First, if you call identify() late, users appear later and their earlier events do not count. Second, a person with no user id is not counted at all.
Why is my chart empty?
An empty chart names what is missing. Usually it is one of these: no identify() call, no activation event, or an event name that does not match what you set. Check the event log to see which events really arrive. Then check the spelling of the event name.
What do you do with the number?
- Compare it week to week, not against another company.
- Split it by segment, for example by plan or by sign-up source, to see who does not activate.
- Change one thing, such as a tour that points at the first step, and compare the next group of new users.
Do not tune the event to make the number look better. If the rate is low, the event is probably right and the onboarding needs work.
What next?
You can also see where people drop out between steps with a funnel of two to four steps. Read the analytics definitions for every number and what it needs.
If you want to see the analytics page first, you can book a demo.