Find where signups drop off

The question

Of the people who look at our pricing page, how many sign up and create their first deal within a week, and where do we lose the rest? You need a few weeks of events, and the developers’ events in place: Signed up from the app, Contacts imported and Deal created (from the browser or the server; the developer track sets them up).

1. Define activation

Activation is the first action that shows someone got value. For a CRM, signing up isn’t it; putting real contacts in and working a deal is. Acme settles on:

1. Viewed the pricing page ($pageview on /pricing, on the marketing site)
2. Signed up
3. Contacts imported
4. Deal created, within 7 days of step 1

Write the definition down where the team can see it (a Lexicon description, or a note on the board in step 10). A funnel is only useful if everyone reads it the same way.

2. Check the events exist

1. Open Lexicon (under Data) and search for each event name.
2. Check each has events in the last 30 days, and that there is only one spelling of each.

The pricing page is on acme.com and the rest is on app.acme.com, so this funnel crosses two sources. That only works if people are carried across. Open Sources › Identity and check the acme.com → app.acme.com hop reads 80% or more.

3. Build the funnel

1. Open Funnels and click New funnel.
2. Pick $pageview for the first step. The picker lists custom events from the last 30 days and autocaptured ones.
3. On that step click + Filter, choose Page, tick or type /pricing, and click Apply filter. The step now reads Page is any of /pricing.
4. Click + Add step and pick Signed up. Add Contacts imported and Deal created the same way. A funnel has 2 to 8 steps.

There are templates under Or start from a template, such as Signup → activation, but their steps are example event names. Replace every step with your own events if you start from one.

If a step matched no events in the last 30 days, saving is blocked and the step is named. Fix the event name, or click Widen to 90 days if it’s rare.

4. Set window, count and order

• Conversion window: 7 days. Someone has a week from viewing pricing to create a deal. The choices are 7, 14, 30 or 90 days.
• Count: Unique people. You are asking about people, not visits.
• Steps must happen in order: leave it on. Importing contacts before signing up isn’t possible anyway, and order makes the drop-offs meaningful.

5. Save it

Name it Activation: pricing → first deal and click Save funnel. Funnels are shared with your whole team.

6. Read the result

Four cards sit at the top:

• Overall conversion: pricing view to first deal. This is your activation rate.
• Converted: how many people finished, out of how many entered.
• Median time to convert, with the 90th percentile underneath. If the 90th percentile is close to 7 days, the window may be cutting people off.
• Biggest drop-off: the pair of steps where most people are lost. Start there.

Because this funnel spans two sources, the top of the results also shows how well identity is carrying between them. If it’s low, the drop between step 1 and step 2 is partly people counted twice, not people lost.

The table under Conversion by step lists each step with people, conversion from the previous step and from the start, how many dropped, and the median time to reach it.

7. See who dropped off

1. Click the drop before step 3 (on its bar, or on its drop percentage). The drawer opens as Dropped before step 3.
2. Open What they did instead: the last thing each person did, including Nothing — session ended.
3. Open Where they were: the top pages and events they reached afterwards.
4. Open People and click a few people to read their timelines.

From there, Heatmaps and Friction show what happens on that page.

8. Break it down

Is the drop the same for everyone? Split it:

• Break down by › Device: does mobile activate worse? Each value gets its own card with its difference in points from the combined baseline.
• Break down by › UTM source: a funnel breakdown reads the value on each person’s step 1 event itself, not the visit’s. Here that is the /pricing pageview, so only people who landed on pricing from a tagged link get a value; the rest are left out of the split. To compare campaigns fairly, use segments (below and in the acquisition walkthrough).
• Compare segments: tick up to five saved segments, plus Other, such as people on the Growth plan against everyone else.

The chart draws the top 3 breakdown values that have at least 500 people entering; smaller ones are listed in the table only. A breakdown and a segment comparison replace each other.

9. Watch it over time

Switch Conversion by step to Over time to see conversion by the day people entered the funnel, and to Time to convert to see how many finished within an hour, a day, a week or longer. Turn on Compare in the filter bar to show the previous period next to Overall conversion.

The most recent days always look worse: people who entered yesterday haven’t had their 7 days yet.

10. Add it to a board

1. On the funnel, click Add to board.
2. Under Tile, choose Linked, so the tile follows the saved funnel if you edit it.
3. Under Filters, choose Follow the board, so it uses the board’s date range.
4. Under Board, pick New board and call it Acme weekly.

Common mistakes

• Trusting a cross-domain funnel with weak identity. If the carry-over rate is low, people appear to vanish between the site and the app. Check Identity first.
• A window that’s too short. If the median time to convert is days, a 7-day window undercounts slow teams. Compare with a 14-day copy before deciding.
• Counting events instead of people. With Events, each person-and-session pair counts separately, and the results say Counts are sessions. For activation, count people.
• Reading the last few days. Recent entrants haven’t had the full window yet.
• Expecting page filters to be saved. Add filter in the filter bar applies to the page you’re looking at only. Save a step filter in the builder with + Filter.
• An event that started mid-range. If Contacts imported was only added two weeks ago, earlier sign-ups can’t convert, and the funnel looks worse than it is. Start the range after the event went live.

Next

Activation tells you who got started. Whether they stay is the next question: Are users coming back?