LemFi · Case study
Designing LemFi eSIM, from a blank page to launch
Summary
I designed LemFi's travel eSIM from a blank page to launch, first for existing customers in the UK, US, Canada and the EU in early 2026, then as part of LemFi's public rebrand in October. It was LemFi's first product that wasn't about moving money, so the job was to make buying mobile data feel as quick and safe as sending money home.
- Role
- Senior Product Designer, sole designer on eSIM, end-to-end UX and UI
- Team
- Product manager, mobile and backend engineers, an eSIM network partner
- Timeline
- Designed late 2025; launched to existing customers in Q1 2026 (February to March); improved through 2026; wider public rollout with LemFi's rebrand in October 2026
- Scope
- Search, plan selection, one-screen checkout, install, eSIM and plan management, a website landing page, then post-launch fixes
- Outcome
- Over 1,600 orders in the first few months, about two thirds of eSIMs installed, and more than 8 in 10 customers who installed buying again
The problem
LemFi's customers live between two countries. When they fly home, or anywhere else, they land without data. That breaks the things they need most on arrival: messaging family, getting a ride, and using LemFi itself.
Roaming is expensive and buying a local SIM at the airport is slow. Other apps already sold eSIMs, but generically. Our angle was to be diaspora-first: show people the destinations they actually travel to.
The goal I set for the design: keep people connected from the moment they land to the moment they return, without leaving LemFi.
My role
I was the only designer on eSIM. I owned the end-to-end experience: the design doc, user flows, every screen and state, the error and fallback paths, the website landing page, and the handoff to iOS and Android. I worked closely with the product manager, who owned the product.
I produced the first end-to-end flow early, so the team had something concrete to react to. After launch I kept designing improvements, and during LemFi's rebrand I brought eSIM orders into the new activity list on our design system.
Research and framing
I started with real competitor eSIMs, not just screenshots. With colleagues who used them, I went through purchase, install instructions, switching plans and how each purchase showed up afterwards.
Three things stood out:
- Install is where people get lost. Buying took seconds everywhere; setting up the eSIM meant leaving the app for phone settings.
- Nobody reads the explainer. Onboarding that opened with "what is an eSIM" got skipped. People want the destination first.
- Plans pile up. Repeat travellers end up with several plans and no clear view of what is active.
I turned the research into a design doc with the PM. It set the targets the design had to hit:
| Customer goal | Business goal |
|---|---|
| Buy an eSIM in under 2 minutes | 90%+ of eSIMs installed within 24 hours |
| Simple install and setup | Very few orders needing support |
| Manage several eSIMs easily | Most buyers coming back for another plan |
Testing before launch was light: an internal test with a few loyal customers who opted in.
The journey
I designed the journey as six steps, each with one job.
- Discover. The first screen asks one question: "Where do you need an eSIM?" Search sits on top, with popular destinations and the full country list below.
- Choose a plan. Plan cards show data and validity, nothing else.
- Pay. One-screen checkout: plan details, payment method, terms and a single pay button.
- Install. One tap, using the phone's own eSIM install, so people never leave the app.
- Ready. A clear confirmation, with a reminder that the plan starts when they land and connect, not at purchase.
- Manage. A home for active and expired plans and data used.

When things go wrong
Most of the design time went on failure. Install depends on the phone, the network and the partner, so every dead end needed a way out. If install fails, people can try again or install manually. If it fails twice, manual install becomes the main action, with support as the backup. Manual install offers a QR code or step-by-step details with one-tap copy.
After launch
Once real customers used it, the gaps were clear: people browsed and left, didn't know if their phone was supported, and got stuck when payment didn't finish. I designed a run of focused fixes, each aimed at one drop-off.
- Device check. A clear warning up front when the phone doesn't support eSIM.
- Free data taster. Free data bundled with the first plan, shown as its own line at checkout.
- Travel reminders. For people "just browsing": pick a travel date and get a nudge before the trip.
- Payment recovery. If someone closed the wallet payment sheet, the order no longer expires silently: they can resume checkout or discard it.
- Website landing page. A way to find LemFi eSIM outside the app (below).

The website landing page
The landing page has three jobs.
Show how simple it is. Four steps, from opening the app to activating once and topping up after.
Help people self-serve. Short video guides for installing, fixing connection problems and managing data.
Stop the wrong purchase. A device check before anyone pays, so people know their phone supports eSIM.
Results
Install was the weak step early on, as the research predicted. The fixes moved it from just over half to about two thirds, still short of the 90% target, so it remains the step to keep working on. Repeat buying became the standout.
| Measure | Early test group | A few months in |
|---|---|---|
| Orders delivered | Dozens | Over 1,600 |
| eSIMs installed | Just over half | About two thirds |
| Data used per plan | Not measured | About three quarters on average |
| Time to second order | A few days | About two days |
More than 8 in 10 customers who installed their eSIM bought another paid plan, usually within two days. That counts paid plans only, not the free taster.
What I learned
People came to buy data for a place. Search first matched how they think.
Every timeout and incompatible phone needed a way out, and manual setup had to be one tap away.
eSIM matters a few times a year. Reminders and the free taster do more for intent than extra explainer content.
A missing "resume checkout" button left orders to expire. Recovery states belong in the MVP, not the backlog.
Avoiding one-off eSIM components kept the product easy to move onto the new design system during the rebrand.