Start learning free

Case study · saas · activation

The empty dashboard: Why trial users never come back

A fictional, interactive SaaS onboarding case study: why new trial users stall on a blank dashboard, and how WXM and PXL turn that into a tested activation plan.

WXM 1 · Wonder (Look)WXM 2 · Architect (Plan)WXM 3 · Validate (Test)WXM 4 · Expand (Grow)
Fictional demonstration

An invented business and wireframes used to teach a method. Any numbers are hypothetical.

By Phil the CRO Guy

Phil the CRO Guy

Mentorele's fictional CRO guide

Phil is the friendly, fictional guide who walks you through every lesson and case study. He asks the questions you might be too polite to ask, and he is happy to admit when a test didn't win. Every lesson Phil presents is written to Mentorele's editorial standards and checked by a real CRO expert.

Learn more about Mentorele Meet Phil

Expert-reviewed by Filip Doneski

Filip Doneski

CRO Expert and Strategist · Partner and Lead CRO Strategist at Wavesy

10+ years in conversion optimization and experimentation for DTC, retail and SaaS. Reviews every Mentorele lesson for accuracy before it’s published.

A/B testingConversion rate optimizationDigital psychology and persuasion

Filip’s expertise at Wavesy LinkedIn profile

Updated · How we write and review

Slide 1 of 9

Chartwise has sign-ups, but not users

Chartwise is a fictional reporting tool with a 14-day free trial. Marketing celebrates rising sign-ups. Product is worried: most trial users log in once and never come back.

A sign-up is the start of the conversion, not the end. Let's see what day one looks like.

Slide 2 of 9

WXM 1 · Wonder (Look)

Wonder: what predicts a user who stays?

The team compares early behaviour of trial users who later paid with those who didn't.

  1. Users who connected a data source and opened a first report within 7 days paid far more often.
  2. Most users who never paid never connected data at all.
  3. That makes a candidate activation event: connected data and viewed a first report within 7 days.
  4. Correlation isn't causation, so the event will be validated by testing whether helping people reach it helps conversion.

Slide 3 of 9

WXM 1 · Wonder (Look)

Wonder: the first screen today

After sign-up, users land on an empty dashboard. Recordings show long pauses, a few random clicks on the menu, then the tab closes.

Control (A)
Fictional first-run dashboard, current version Original wireframe, not a screenshot.
  1. Blank canvas. Nothing shows what the product can do or what to do first.
  2. Weak next step. 'New report' fails without data, which users only discover after clicking.

Slide 4 of 9

WXM 1 · Wonder (Look)

Wonder: what new users say

A one-question in-app survey on day two ("What's stopping you from building your first report?") and five short interviews.

  1. "I wasn't sure which data source to connect first."
  2. "I wanted to see what a report would look like before giving access to our data."
  3. "I didn't have time to set it up properly on day one."

Uncertainty, trust and effort. Three different reasons; one blank screen makes all three worse.

Slide 5 of 9

WXM 2 · Architect (Plan)

Architect: the hypothesis

Because day-one recordings, the survey and interviews show users unsure what to do and hesitant to connect data blind, we believe a first-run screen with a sample report and a two-step checklist (connect data, open your first report) will increase the share of trial users who activate within 7 days, because it shows the value first and breaks setup into small, visible steps.

Your turn: What is the best primary metric for this test?
Show Phil’s answer

Activation within 7 days. Clicks measure engagement with the new UI, not progress towards value. Sign-ups happen before the change. Activation is the outcome the change aims at; trial-to-paid is a guardrail and a longer-term check.

Slide 6 of 9

WXM 2 · Architect (Plan)

Architect: PXL score

It's a big, visible change on the first screen every trial user sees, backed by recordings, survey and interviews. It takes engineering time, which lowers the ease score.

PXL score: 10 / 14

Is the change above the fold?Yes1
Is it noticeable within 5 seconds?Yes2
Does it add or remove an element?Yes2
Is it designed to increase user motivation?Yes1
Does it run on high-traffic page(s)?Yes1
Does it address an issue found in user testing?Yes1
Does it address an issue found in qualitative feedback (surveys, polls, interviews)?Yes1
Does it address insights found in digital analytics?Yes1
Is it supported by mouse-tracking heatmaps or eye tracking?No0
Ease of implementationMore than 2 days0

PXL by CXL. A high score ranks an idea; it does not predict a win.

Slide 7 of 9

WXM 3 · Validate (Test)

Validate: control versus variation

The variation keeps the same product but changes the first screen.

Control (A)
A: empty dashboard Original wireframe, not a screenshot.
Variation (B)
B: sample report plus a two-step checklist Original wireframe, not a screenshot.
  1. Value first. A clearly labelled sample report shows what users will get.
  2. Small steps. Two steps with visible progress make setup feel achievable.

Slide 8 of 9

WXM 3 · Validate (Test)

Validate: plan and guardrails

Written down before launch.

  1. Primary metric: activation within 7 days of sign-up.
  2. Guardrails: trial-to-paid conversion, support tickets per new user, data-connection errors.
  3. Randomization unit: the account, so teammates see the same experience.
  4. Duration: enough sign-ups for the planned MDE, in whole weeks, and long enough for the 7-day window to close for every user in the analysis.

Randomize by account, not by user. Otherwise two colleagues might see different products and compare notes.

Slide 9 of 9

WXM 4 · Expand (Grow)

Expand: what happens next

Chartwise is fictional, so there's no result to report. The plan for each outcome:

  1. Activation up, guardrails fine: roll out, then test the onboarding emails against the same activation event.
  2. Activation up, trial-to-paid flat: the activation event may be a weak predictor; revisit the definition in Wonder.
  3. No change: check whether users saw the checklist; interview users who saw it and still didn't connect.

What this case teaches

  • Define activation from data, then validate it with experiments.
  • An empty state is a design decision, and often the most viewed screen in a trial.
  • Randomize at the right unit (the account) for B2B products.

Chartwise is fictional. No real company, data or result is described.

Sources

  1. WXM: the Wave Experimentation Model. Wavesy. Accessed 4 November 2026.
  2. PXL: A Better Way to Prioritize Your A/B Tests. CXL. Accessed 4 November 2026.

Learn the principles behind this case

Search lessons, guides, case studies and templates. Press Esc to close.