Trial X
20 FIRST CHARGESFewer families convert. Most of them stay.
100paid months
20 + 18 + 17 + 16 + 15 + 14
20 + 18 + 17 + 16 + 15 + 14
FIELD REPORT 13 · SUBSCRIPTION · PICNIIC
The free trial
redesign
Picniic is a family organizer app with a paid premium tier. I ran a redesign of its free trial. Subscription revenue per trial start rose 25%, net of churn, and the lift held as the cohorts aged.
+25%subscription revenue,
net of churn, sustained
net of churn, sustained
- OLD TRIAL COHORTS
- NEW TRIAL COHORTS
- THE GAP
01 THE BRIEF
A trial has one job: get the family to the value before the charge.
Picniic is a family organizer: a shared family calendar, to do and grocery lists, meal planning and an information locker. The free tier covers the basics. Premium sits behind a free trial.
The trial sat between a family's first session and its first charge. What happened inside it decided who paid, and who stayed.
I ran the redesign and graded it on one number: subscription revenue per trial start, net of refunds and cancellations, tracked month over month by cohort.
COMPANYPicniic, family organizer app, free and premium
THE CHANGEThe free trial, from day 0 to the first renewal
THE SCORERevenue per trial start, net of churn
02 THE MECHANISM
The gap opened at the first charge and stayed open.
Switch between gross and net, then scrub the months. Both views are cumulative subscription revenue per trial start, indexed so the old trial reaches 100 at month six, net. The curve shapes are illustrative. The 25% gap is the real result.
● COHORT CURVESINDEX · OLD TRIAL, MONTH 6, NET = 100
NET OF CHURN
Refunds and cancellations taken out. Both curves bend as families leave, and the new trial stays 25% above the old one as the cohorts age.
- OLD TRIAL
- NEW TRIAL
- THE OTHER VIEW
- WHAT CHURN TOOK FROM THE NEW TRIAL
OLD TRIAL100.0Index, per trial start.
NEW TRIAL125.0Index, per trial start.
GAP+25.0 · +25%The ratio does not move as the cohorts age.
CHURN TOOK31%Of gross, gone by month 6, in both trials.
Gross assumes every family that paid once kept paying. Net takes out refunds and cancellations. In this sketch both trials lose the same share to churn, which is what a sustained 25% implies: the gap in index points grows as the cohorts age, and the ratio stays put.
03 THE TRIAL
Value before the ask, and no surprise at the charge.
What changed, stage by stage. The old trial runs on the top track, the new trial below it.
DAY 0
OLD TRIAL
The trial starts at sign up. The premium tools wait behind upgrade prompts.
NEW TRIAL
Every premium tool opens. Guided setup builds the shared calendar and a first list. The family picks its plan here.
DURING THE TRIAL
OLD TRIAL
The family finds the value on its own, or it does not.
NEW TRIAL
Value moments by design: a first shared event, a first grocery list, a first week of meals.
BEFORE THE CHARGE
OLD TRIAL
No reminder.
NEW TRIAL
A reminder with the date, the amount, and what the family has set up so far.
FIRST CHARGE
OLD TRIAL
The charge arrives. For some families it is a surprise.
NEW TRIAL
A trial end screen recaps the calendar, lists and meals. Then the charge, on the plan the family picked.
RENEWAL
OLD TRIAL
Renews until the family cancels.
NEW TRIAL
Renews on the plan chosen on day 0, with a household already built.
- STAGE, OLD TRIAL
- STAGE, NEW TRIAL
- VALUE MOMENT
- REMINDER
- CHARGE WITHOUT WARNING
04 WHY NET OF CHURN
A trial that converts more can still lose money.
A worked example with made up numbers, per 100 trial starts. Each bar counts the families paying that month. Trial Y wins the first charge. Trial X wins the half year.
Trial Y
30 FIRST CHARGESMore families convert. 8 ask for a refund, and the rest leave faster.
81paid months
22 + 16 + 13 + 11 + 10 + 9
22 + 16 + 13 + 11 + 10 + 9
- FAMILIES PAYING THAT MONTH
- FIRST CHARGES REFUNDED (8 OF 30)
GRADED ON FIRST CHARGES
Y wins by 50%: 30 against 20.
GRADED ON SIX MONTHS, NET
Y loses by 19%: 81 paid months against 100.
That is why the Picniic trial was graded on revenue per trial start, net of refunds and cancellations, month after month. The first charge only starts the count.
05 HOW I READ IT
One number, read the same way for every cohort.
COHORTS OF TRIAL STARTS
Every trial start joined a cohort by the month it started. The old trial's cohorts were the baseline. The new trial's cohorts were the test.
REVENUE PER TRIAL START
Cumulative subscription revenue divided by the cohort's trial starts, updated month over month, so a small cohort and a big one sat on the same scale. Old and new cohorts were compared at the same age.
NET OF REFUNDS AND CANCELLATIONS
Refunds came out of revenue, and cancelled families stopped adding to it. A charge only counted if the money stayed.
GUARDRAILS
Refund rate, early cancels and billing support contacts, each read against the old trial's cohorts. A trial that bought revenue with refunds or angry tickets would have shown up there first.
06 THE RESULT
Subscription revenue rose 25%, and it held.
Per trial start, net of refunds and cancellations, the new trial's cohorts ran 25% ahead of the old trial's, and the gap held as the cohorts aged.
+25%subscription revenue per trial start, net of churn
125index against the old trial at 100
Heldas the cohorts aged, month over month
OLD TRIAL100
NEW TRIAL125
The 25% is from my own reporting at the time, read net of refunds and cancellations. Curve shapes, the six month window and the worked example are illustrative. Raw subscriber and revenue figures stayed with Picniic.
07 WHAT CARRIES OVER
Grade every trial on what stays.
01
Remind before you charge. A family that knows the date and the amount decides before the card is hit, and a cancel costs less than a refund.
02
Put the value before the ask. Open the paid tier on day 0 and walk the family into one real use before the trial ends.
03
Show people what they built. A trial end recap turns the first charge into a receipt for work they have already done.
04
Grade on revenue per trial start, net of refunds and cancellations, by cohort age. Trial starts and first charges flatter the wrong variant.