YuppTV+18%
nomadTravel / eSIM
How Nomad turns customer signals into growth across dozens of markets
+22%
ARPU · measured against a randomised holdout
In short
Nomad sells eSIM data plans in dozens of countries, where every traveller has a different trip, device and price sensitivity. Markin now decides the next best action for each customer and launches it through Nomad's existing stack, lifting ARPU 22% on treated cohorts against a randomised holdout.
Results
Every decision carries a control group+22%
ARPU on treated cohorts
vs randomised holdout, full measurement window
+34%
Repeat purchase rate
travellers buying a second plan within 90 days
11×
Experiments live per month
compared with the pre-Markin campaign calendar
Anonymised range across Markin deployments in large B2C bases, read over a full measurement window. It is not an industry benchmark and not a forecast for your base.
At a glance
- Industry
- Travel connectivity / eSIM
- Markets
- Dozens, sold in local currency
- Base
- Millions of travellers, mostly one-trip buyers
- Motions live
- Cross-sell, win-back, top-up timing, price testing
- Stack
- Warehouse, CDP, Braze, app SDK, Stripe
- First measured lift
- Week 6
The challenge
Nomad's growth was capped by the shape of the product: most travellers buy once, for one trip, and disappear. The revenue question was never 'who do we email', it was 'which of dozens of plan, price and timing combinations is worth showing this specific traveller before their data runs out'.
A campaign calendar could not answer that. Segments were built by market and rebuilt every quarter, offers were the same for a two-day city break and a three-month relocation, and the team had no way to tell whether a top-up push had created revenue or simply harvested a purchase that was going to happen anyway.
Hiring their way out of it was not realistic either: the analytical work needed to keep dozens of markets individually optimised is a full data-science pod, permanently.
What Markin did
Data and context, intelligence, action01 / Data & context
Trip context, not just a customer record
Markin reads Nomad's warehouse, CDP and app events in place and builds a per-traveller picture: destination, device, plan size, consumption curve, roaming behaviour and the local price of alternatives. No data is copied into a new system.
02 / Intelligence
Hypotheses the team never had time to write
The intelligence layer surfaces revenue opportunities on its own, sizes them and turns them into candidate actions: who is about to run out of data, who is extending a trip, which market tolerates a larger plan, where a discount only cannibalises a purchase already coming.
03 / Action
1:1 execution inside Nomad's own stack
Approved actions launch through Braze, the app and Nomad's own checkout, each with a randomised holdout attached. Actions that fail to beat their control are retired automatically rather than living on in a calendar.
Experiments that ran
Hypothesis to measured lift| Hypothesis | Segment | Action | Channel | Measured lift |
|---|---|---|---|---|
| Heavy users hit their cap before the trip ends and churn to a local SIM | Travellers at 70% consumption with 3+ days remaining | Top-up offer sized to the remaining trip, not a fixed pack | Push + in-app | +19% top-up revenue |
| Multi-country itineraries are under-served by single-country plans | Bookings with connections in 2+ regions | Regional plan positioned at purchase and at first border crossing | Checkout + email | +12% AOV |
| Past travellers return on a seasonal cycle we are not anticipating | Lapsed buyers with a repeat-travel pattern | Win-back timed to their own travel rhythm, no discount | Email + paid audience | +34% repeat rate |
| Discounting new markets buys volume we would have won anyway | First-time buyers in high-intent markets | Withhold the promo, hold price, measure against control | Checkout | Margin preserved, no volume loss |
How it rolled out
Week 0
Connect, do not migrate
Warehouse, CDP, Braze and app events connected read-only. First opportunity feed generated from historical behaviour.
Week 2
First motion live
Top-up timing goes live in three markets with a randomised holdout on every decision.
Week 6
First readable lift
Top-up revenue reads +19% against control. Two hypotheses are retired for failing to beat their holdout.
Week 12
Scaled to the full footprint
Four motions running across every market, with the growth team reviewing decisions rather than building campaigns.
In their words
“Markin runs 1:1 lifecycle across dozens of markets. It's the growth team we could never hire.”
Growth Lead, Nomad
The stack Markin worked with
No migration, no new system of recordQuestions buyers ask about this case
Why does an eSIM business need decisioning rather than more campaigns?
Because the revenue question is per traveller, not per segment. A two-day city break and a three-month relocation need different plans, prices and timing, and no campaign calendar can hold dozens of markets in that resolution.
How was the Nomad result measured?
Every decision carries a randomised holdout. The lift quoted is the difference between the treated cohort and that control group over a full measurement window, not a before-and-after comparison.
Did they have to move their data to Markin?
No. Markin reads from the warehouse, CDP and product events already in place. Nothing is copied into a new system and nothing is locked in.
Which team ran it day to day?
The existing growth team. Markin does the analytical work, a travel team of that size would need a dedicated data-science pod to produce, and routes every action into the tools they already operate.
How long until the first measurable lift?
A first motion goes live in about 30 days and the first statistically readable result lands inside the following six weeks. Scaled impact typically reads at 90 days.
