Fullscript Autoship
A dormant subscription feature stuck at single-digit adoption. I ran the research and redesign that turned it around.
- Role
- Product Designer — Patient Experience
- Timeline
- January 2021 – December 2022
- Team
- 1 Product lead · 1 Designer (me) · 1 UX researcher · Titan engineering squad
- Scope
- Subscription cart entry · Pre-checkout flow · Manage Autoship portal · Cancellation instrumentation · Patient research program · Web, desktop and mobile
My role
- Ran the Q1 2022 research program: a 1,485-response in-app survey, follow-up interviews, and a scored baseline audit of every subscription flow.
- Designed the full-page and dropdown cart experiment that moved subscribing into the purchase flow, and the pre-checkout frequency step that followed.
- Designed the Manage Autoship portal: skip, reschedule, frequency change, and cancellation as first-class actions.
- Designed the cancellation-reason dialog and its taxonomy — instrumentation the retention team still runs on 4 years later.
- Shadowed customer support calls and turned recurring patient complaints into the research agenda.
TL;DR
In short
Problem
Patients prescribed the same supplements every month were reordering by hand — the subscription feature had existed since 2018, and fewer than 1 in 16 orders used it.
Approach
Grounded the redesign in a 1,485-response patient survey and a scored baseline audit, then moved subscribing into the cart and rebuilt management around the control patients asked for.
Outcome
More than 1 in 5 orders now arrive on subscription — activations grew 6.4× inside the redesign year, inflecting at both ship dates.
Context and problem
The situation we walked into
Business context
Fullscript's revenue is practitioner-prescribed supplements, and most prescriptions run for months — but reordering was manual. The subscription feature that should have carried repeat revenue had sat dormant since 2018 at single-digit adoption. Nothing about it was on fire, so nothing about it got prioritized.
User context
Survey write-ins named the hesitation in patients' own words: "I'm not sure how often I'll need a refill," fear of ending up with excess supplements, recurring charges they'd forget about. 13% of respondents didn't know the feature existed. Patients wanted to never run out — without surrendering control.
Constraints and complexity
The cart is the most contested surface in any commerce product, so any subscription placement risked hurting all conversion to protect one program. And there was no dedicated patient analytics at the time — evidence had to be built by hand through the survey, interviews, and a step-by-step scored audit.
Goals and success metrics
What success looked like
Make subscribing the default way repeat patients reorder — a measurable adoption lift with no harm to cart-to-checkout conversion.
Success criteria
- 80%+ confidence of no harm to cart → checkout conversion under A/B.
- A statistically significant lift in Autoship adoption (+2% target).
- No measurable impact on cart loading speed.
- A/B rollout to 50% of stores before full release.
Approach
How we got there
- 01
Shadowed support before sketching
I sat in on customer support calls in my first months — 4 patient calls in a single hour, none about checkout mechanics. The friction was hesitation: refill timing, excess supply, forgotten charges. Those calls set the research agenda before any screen was touched.
- 02
Surveyed 1,485 patients in-app
A card on the Autoship page plus an email drew 1,485 responses. The surprise: 78% of existing users rated the feature 4 or 5 out of 5, while 13% of respondents didn't know it existed. Adoption, not quality, was the problem — and the write-ins listed the exact controls patients were missing.
- 03
Scored the baseline, moved the roadmap
A step-by-step audit — the team's "rainbow spreadsheet" — graded every existing flow so the redesign had something concrete to beat. Survey plus audit convinced product and engineering to bet a full roadmap on a feature nobody had prioritized: that quarter became the busiest in Autoship's 9-year history.
- 04
Shipped in 3 waves
April 2022: subscription choice moved inside full-page and dropdown cart variants. July: the Manage Autoship portal — order now, skip, reschedule, and the cancellation dialog. September: pre-checkout frequencies. Activation growth inflected at the April and September ships and reached 6.4× January's volume by November.
Key decisions
Forks in the road
Cart placement vs a subscription page
Autoship lived as a diverging path from how patients actually order: browse, select, check out. A dedicated subscription page was the safe option — and the invisible one. Doubling the interface's calls to action was considered and rejected as cognitive load. That left the riskiest, most visible option: the cart itself.
- Decision
- Subscribing became a selection inside the cart, where purchase intent already lives — with the risk written into the hypothesis and A/B guardrails agreed before ship.
- Tradeoff
- We spent contested cart real estate and accepted a measured risk to overall checkout conversion, and we studied, then declined, Amazon's basket-maximizing cart patterns.
Flexible exits vs locked-in subscribers
Retention orthodoxy says make leaving hard. The survey said the opposite: patients asked, in their own words, to skip, pause without losing settings, delay a shipment, and deliver one now. We treated those as transcription, not feature ideas.
- Decision
- Skip, reschedule, and frequency change shipped as first-class actions with the same prominence as cancel.
- Tradeoff
- Painless pausing could cannibalize orders short-term — we bet that timing problems, not lost value, drove churn.
Instrumented exits vs frictionless cancel
Adding a step to the most sensitive moment in the lifecycle reads as hostility, and dark-pattern cancel flows poison trust in a healthcare product. But leaving silently teaches the team nothing. The compromise was one tap: a single structured question with 3 honest, plain-language reasons.
- Decision
- Cancelling asks why — once, with 3 tappable reasons: enough supply, cost, or a product issue.
- Tradeoff
- One extra step at the exit, in exchange for turning every departure into retention data.
Solution
What we shipped
Outcomes and impact
What it moved
Orders on subscription
1 in 5
up from 1 in 16 in January 2021
Activation growth, 2022
6.4×
January to November, inflecting at the April and September ship dates.
Survey responses
1,485
Cited as the team's patient-feedback benchmark for years after.
Cancellation taxonomy
3 → 9 reasons
Collecting continuously since June 29, 2022.
Business impact
Autoship's share of patient orders roughly doubled across the redesign window and has compounded every year since — from fewer than 1 in 16 orders in January 2021 to more than 1 in 5 today. A later growth team built its subscribe-and-save experiment directly on these flows.
User impact
Four years of cancellation data confirmed the founding bet: the top reason patients cancel is deliveries arriving too early — a timing problem — and skip and reschedule events now rival cancellations in volume. Patients manage their subscriptions instead of abandoning them.
In their words
What people said
“Emile's energy is infectious… his attention to detail and passion for getting things right is flawless… to the point where Titan calls it 'roasting Emile's designs.'”
Harry Verran — Product Lead, Fullscript “Stay hungry! Your hunger and enthusiasm are energizing. Working with you is an absolute pleasure. Your detail orientation, creativity and thoughtfulness are superpowers.”
Ben Leger — Director of Design, Fullscript
Learnings and reflections
What I’d take with me
Dormant features hide the biggest wins
Nothing about Autoship was on fire, which is exactly why nobody had looked closely at it. One survey later, a feature that had idled for 4 years got the busiest engineering quarter in its history.
A dialog can be instrumentation
The highest-value screen in the project is one most patients see once. Its reason data started flowing on June 29, 2022, and it still sets the retention roadmap — the data stream turned out to be part of the deliverable.
Patients leave over timing, not value
The top cancellation reason today is deliveries arriving too early, and skip and reschedule volumes rival cancels. Absorbing a timing problem keeps the relationship; punishing it ends the subscription.