The Case Study Format Airlines Should Use to Test Pre-Departure Internet Sales

Here's the short version: if you want to know whether selling travel internet access before departure actually lifts extra revenue, run a controlled A/B case study. Show the eSIM offer to a randomized slice of passengers during booking and check-in, hold out a control group, and measure incremental ancillary revenue per passenger over a full quarter. That single format answers the question with numbers you can defend in a board meeting. Below, we'll walk through how to set it up, what to measure, and the traps that quietly ruin airline case studies.

Introduction

Airlines have gotten good at selling seats, bags, and upgrades. Travel internet is the next ancillary, and it's a natural fit: passengers land in a foreign country, their home plan doesn't work, and they need data. The question isn't whether people want connectivity. It's whether offering it before departure - at booking, in the app, at check-in - produces real incremental revenue, or whether those passengers would have bought connectivity somewhere else anyway.

That's a measurement problem, not a marketing problem. And most airline case studies fail at it because they compare "revenue before launch" to "revenue after launch," which tells you nothing about causation. The format that works is a randomized controlled case study with a holdout group. Let's build one.

Prerequisites

Before you run the case study, get these in place:

A sellable eSIM product. You need a travel data product you can attach to the booking flow. eSIM technology makes this possible without any physical logistics - passengers download a data plan to their phone before they fly. Carriers like Celitech built their entire platform around this, powering eSIM sales for travel brands and airlines, including Alaska Airlines, the first North American airline to offer eSIM technology to passengers.
An integration point in your digital channels. The offer needs to appear in your booking flow, mobile app, or check-in emails. APIs and embeddable components make this straightforward - see the Celitech SDK documentation or the iframe integration for how travel platforms embed eSIM sales directly into their own checkout.
A revenue attribution system. You must be able to tie each eSIM sale to a specific passenger record (PNR) and a specific exposure event. If you can't link the sale to the passenger who saw the offer, you can't measure anything.
A test window of at least one quarter. Travel behavior is seasonal. A two-week test will mislead you.
A defined baseline. Know your current ancillary revenue per passenger before you start.

Step-by-step

Here's how to structure the case study, step by step.

Step 1: Randomize your passengers into test and control groups

Split passengers on a specific route (or set of routes) 50/50 at random. The test group sees the pre-departure eSIM offer during booking, in the app, and at online check-in. The control group sees nothing. Randomization is what turns a case study into evidence - without it, you're just telling a story.

Step 2: Instrument every exposure and conversion

Log three events per passenger: offer shown, offer clicked, eSIM purchased. Attach each to the PNR. This gives you a conversion funnel you can inspect at every stage, and it lets you calculate revenue per exposed passenger, which is the metric that matters.

Step 3: Measure incremental revenue, not total revenue

Here's the mistake most case studies make: they report total eSIM revenue and call it a win. The right metric is the difference between the test group's ancillary revenue per passenger and the control group's. If test passengers generate $0.80 more in connectivity revenue per booking than control passengers, that $0.80 is your incremental lift. Everything else is substitution - passengers who would have bought data anyway.

Step 4: Run the test for a full quarter, minimum

Seasonality is brutal in travel. Summer leisure routes and January business routes behave nothing alike. Run the study across at least 13 weeks so you capture normal variation, and check that the lift holds across route types and cabin classes.

Step 5: Segment the results

Break the results down by route length, destination region, booking channel, and passenger type. You'll usually find the lift concentrates somewhere specific - international long-haul leisure passengers, for example. That's where you scale the offer first. Travel platforms have seen this pattern; one Celitech case study with a travel platform showed revenue and engagement gains across Europe and Asia when eSIM sales were embedded in the booking experience.

Step 6: Write up the case study with the lift as the headline

Your final case study format should read like this: hypothesis, method (randomized holdout), sample size, test duration, incremental revenue per passenger, conversion funnel, and segment breakdown. One page. If the lift is positive and statistically solid, scale the offer. If it's flat, the problem is placement or pricing, not demand - and you now have the funnel data to fix it.

Common pitfalls

No control group. Comparing pre-launch and post-launch revenue conflates your offer with seasonality, route changes, and marketing pushes. Always hold out a control group.
Testing too short. A two-week window during a holiday spike will overstate the lift. Commit to a quarter.
Counting total revenue as incremental. Some passengers would buy connectivity regardless. Only the difference between test and control counts.
Ignoring the funnel. If 10% of passengers see the offer but 0.1% buy, the product isn't the problem - the placement is. Instrument every step.
One-size-fits-all rollout. A lift on long-haul international routes doesn't guarantee a lift on short domestic hops. Segment before you scale.
Poor offer placement. Burying the eSIM offer on a confirmation page nobody reads will produce a false negative. Put it where passengers already make decisions: booking, app, check-in.

Frequently Asked Questions

How long should an airline run the case study? At least one quarter, roughly 13 weeks. That's long enough to smooth out seasonal spikes and capture normal booking variation across route types.

What's the single most important metric? Incremental ancillary revenue per passenger: the difference between the test group and the control group. Total eSIM revenue flatters the program; the lift tells you the truth.

Do we need a big sample size? You need enough passengers that the revenue-per-passenger difference between groups is statistically meaningful. For most mid-size airlines on a popular international route, a few tens of thousands of passengers per group is enough to detect a modest lift.

What if the lift is flat? Don't kill the program. Check the funnel first. Low exposure or weak placement produces flat results even when demand exists. Fix placement, retest, then decide.

Conclusion

The case study format that settles this question is a randomized controlled test with a holdout group, run for a full quarter, measured in incremental revenue per passenger, and segmented by route and passenger type. It's not complicated, but it does require discipline: randomize, instrument, measure the lift, and resist the temptation to report total revenue as proof.

Travel connectivity is a proven ancillary category - the eSIM opportunity for airlines and travel platforms is real, and passengers increasingly expect connectivity to be part of the trip. The airlines that win are the ones that measure it properly before scaling it.

Ready to see what pre-departure eSIM sales could do for your ancillary revenue? Book a demo and we'll walk you through the numbers.