Quibi: A 90-Day Free Trial That Hid the Truth for 90 Days
Evidence: 90 days for launch-week users
HypothesisThe trial was meant to build a habit. Instead it delayed the only number that mattered, so Quibi spent three months optimising for downloads it couldn't monetise.
In 30 seconds
- What happened: Quibi raised $1.75B for premium short videos made for phones, launched in April 2020, and shut down in December 2020.
- The usual explanation: the pandemic kept its "on the go" audience at home.
- My read: the pandemic hurt, but a product decision made it fatal. A 90-day free trial meant Quibi couldn't learn whether people would pay until July, three months into a launch it had no time to waste.
- The number: Sensor Tower estimated about 8% of the earliest trial users converted to paid. Quibi disputed it, but never published its own figure.
The thinking, drawn
3 diagrams
What I'd Have Done Differently
- Get the payment signal in weeks, not months. Run trial lengths side by side (7, 14 and 30 days) from launch day, and read week-four retention by group.
- Watch week-two return rate before conversion. People who come back unprompted in week two are the ones who'll pay. Estimate a weak week-two return rate was likely visible by late April. Quibi never published it.
- Re-test the core bet once lockdowns started. "Watch on the go" stopped being true overnight. That was the moment to test the product on TVs, not months later. Quibi only added TV apps in October.
A generous free trial doesn't just cost money. It costs the time it takes to learn the truth.
The full deep dive
For interviewers who want every detail and source.
More on the signal
The underrated lesson isn't "8% is low." Disney+ lost most trial users too. It's that Quibi chose a trial length that made its most important metric arrive too late to act on.
What Happened
- Before launch: raised $1.75B. Priced at $4.99 a month with ads, or $7.99 without.
- April 6, 2020: launched with a 90-day free trial, just as lockdowns kept people at home.
- May 2020: shortened the trial to 14 days for new users.
- July 2020: the first 90-day trials ended. Sensor Tower estimated at most about 72,000 of 910,000 first-three-day sign-ups stayed, roughly 8%. For comparison, Disney+ converted about 11% in a similar period, on a 7-day trial.
- October 2020: announced shutdown. Service ended in December.
PM Deep Dive
Learning velocity: the concept Quibi got wrong
A launch is an experiment. The faster you get a trustworthy read on the metric that decides the business, the more chances you have to fix things. Quibi's trial length set its learning velocity to one read per 90 days, the slowest possible choice for a company that had raised money against an aggressive plan.
Metric tree
- North Star: paid subscribers still active in their third paid month.
- Input metrics: weekly active viewing days, episodes per session, week-2 return rate, trial-to-paid conversion.
- Guardrail: CAC per paid subscriber. Downloads were Quibi's headline number, but a download from a 90-day free trial cost money and proved nothing.
Activation: defining the aha moment
I'd define activation as finishing three episodes of the same show in week one. (Hypothesis.) A user who does that has found a reason to return; a user who sampled five first episodes has not. Content and recommendations would then be judged on how many new users reach it.
Positioning under a changed context
Quibi's job to be done was filling short gaps on the go: a commute, a queue. Lockdowns removed those moments almost overnight. And for short video on a phone, TikTok and YouTube were free substitutes. So Quibi was charging for a job that had disappeared, against competitors priced at zero.
Technology
Quibi's flagship feature, Turnstyle, switched video between portrait and landscape as you rotated the phone. It was clever engineering, and it also drew a patent lawsuit over that very technology. My read: rotating the phone was a problem few viewers felt, so engineering effort followed the technology, not the user.
Experiment I'd have run on day one
- Design: a multi-arm test randomised at sign-up: 7-day, 14-day and 30-day trials.
- Primary metric: paid conversion and month-2 retention per arm.
- Decision rule: pick the arm with the best payback per paid subscriber, not the most sign-ups.
- Benefit: the first read arrives in week two, not month three.