I got tired of $15 airport SIM cards, so I bootstrapped an eSIM store. Here's what building it actually taught me.

Every traveler knows the ritual. You land in a new country, you're jet-lagged, and the first thing you do is hunt for a SIM card kiosk — or worse, leave roaming on and pray. I'd done it dozens of times: queuing at an airport counter, handing over a passport, paying $15–30 for a SIM I'd use for five days and throw away.

eSIM technology quietly killed the need for any of that a few years ago. Almost every recent phone supports it. You can provision a data plan over the air, scan a QR code, and be online before you've cleared customs. But the experience around it still felt clunky — confusing coverage maps, surprise pricing, plans that expire silently. So I built Cellesim, a travel eSIM store covering 200+ countries, as a bootstrapped, solo-ish project. This is a write-up of the parts that were harder than I expected, because those are the parts I'd have wanted to read before starting.

The "boring" infrastructure is most of the work

The pitch is simple: pick a destination, pay, get a QR code by email in under a minute, scan it, you're connected. No SIM swap, no roaming fees. Under the hood, "simple" means stitching together carrier aggregators' APIs, a payments stack, provisioning webhooks, usage syncing, and a catalog that spans unlimited plans across 160+ countries. None of it is glamorous. All of it breaks in interesting ways.

The lesson I keep relearning: in a commodity-ish market, your moat isn't the product feature, it's the thousand small reliability and trust details. QR delivered in seconds, not "within 24 hours." Honest coverage info. A refund for unused data when a trip ends early. We even bundle a free VPN for active customers — a differentiator that costs us a bit but earns goodwill that paid ads never could.

The bug that taught me the most

Here's the war story. We sell regional bundles — say, an "Asia" plan covering 12–20 countries. Internally these map to coverage codes, and the storefront would surface a representative coverage list for the bundle. A subtle ordering bug meant that for some regional packages, a narrow sub-bundle (covering only 2–3 countries) was being shown as the headline coverage for the wide one.

The data was correct. The display wasn't. And the failure mode was the worst kind: customers heading to, say, Uzbekistan or Cambodia would buy a "Central Asia" or "Southeast Asia" plan, arrive, and find their specific country wasn't on the eSIM that got provisioned. By the time it surfaced, it was a support ticket from someone standing in a foreign airport with no data.

We fixed the root cause, but the bigger lesson was about how we responded: full refunds, re-provisioning at our own cost, and a written "this won't happen to you again." Several of those frustrated customers left positive reviews afterward — not because nothing went wrong, but because the recovery was clean. In a trust business, your refund policy is a marketing channel. I stopped treating support as a cost center the day I understood that.

Localization is a conversion problem, not a translation problem

We run in 24 languages. Early on I assumed translation was the whole job — ship the strings, done. Wrong. A locale can be perfectly translated and still convert at basically zero, because the intent lives in a different language or the payment expectations don't match. One of our markets browsed heavily in its native language but only ever purchased in English. Another abandoned checkout en masse because the price flipped to an unfamiliar currency at the last step and felt like a surprise.

Fixing that wasn't about better translations. It was currency clarity, locale-aware funnels, and meeting people where they actually buy. If you're going international, treat each locale as its own little product with its own conversion problem.

Earning authority in a crowded niche

"Travel eSIM" is a competitive search term dominated by a few well-funded players. You don't out-spend them as a bootstrapper; you out-specific them. Two things have moved the needle.

First, useful tools. We built a free eSIM compatibility checker so anyone — customer or not — can confirm their phone supports eSIM before they go down the rabbit hole. Utilities like that get linked to and shared in a way that a product page never will.

Second, original data. Instead of writing another generic "best eSIM" listicle, we published a breakdown of what a gigabyte of mobile data actually costs around the world. Concrete, sourced numbers are the kind of thing journalists and other site owners cite, and citations are worth more than any amount of keyword stuffing. Pair that with genuinely helpful destination content — our blog leans into specific, long-tail traveler questions rather than fluff — and you slowly build the kind of topical authority that compounds.

The transactional pages still matter too, of course. Something like our Japan eSIM page has to load fast, show real plans, and answer "will this work where I'm going?" without making the visitor think. But the tools and the data are what get you found and trusted in the first place.

What I'd tell a past version of me

• Reliability beats features. A boring product that always works will beat a clever one that sometimes doesn't, especially when the customer is abroad and stressed.
• Your support inbox is your roadmap. Nearly every meaningful fix we shipped started as a confused or angry message. Read all of them.
• Trust scales slower than traffic. Reviews, refunds, accurate coverage — these don't spike a graph, but they're the reason repeat customers exist.
• Differentiate on the unglamorous stuff. Unused-data refunds, instant delivery, a free VPN, multilingual support that actually converts. Individually small; together, a reason to choose you.

Where it's at, and what's next

The web store is live and the catalog keeps expanding. The current focus is a native mobile app — built so customers can manage eSIMs, watch their data usage, and reinstall a QR if a phone dies mid-trip, all without digging through email. Building it has forced a lot of clarity about what the core experience really is, stripped of the marketing surface.

If you've shipped anything in the travel, telecom, or international-payments space, I'd genuinely love to compare notes — especially on cross-border conversion and trust-building, which still feel like the hardest parts. And if you travel and have opinions about what a good eSIM experience should feel like, tell me where the current one falls short. That feedback loop is the whole reason I post here.

Thanks for reading.