Reducing churn among Talabat Pro subscribers
Context
Talabat Pro is the subscription program of Talabat, a food delivery app used in 8 countries by 7.8M customers, with $19.8B in GMV. We reached 2M users on the Talabat Pro program.
Our strategy was focused on retention. But we noticed something that didn't add up: 30% of people cancelling Talabat Pro were saving more money than they paid for the subscription.
Before that, there was already a banner on the account page sharing the savings since they joined the program but it wasn't moving the needle.
Why this happens
First, value is invisible: the saving banner showed a number, but on its own, without any anchor or timeline.
Second, benefits don't stop immediately: people keep enjoying their perks until the end of the cycle, so the feeling of loss shows up too late to change their mind.
Problem statement
How might we remind users of the value they're getting, and show them they're already saving more than what they pay for the subscription?
Building a Claude skill to leverage research
As a content designer working across different teams, taking part of the research process and analysing user interviews can be hectic. It takes hours, and sometimes, I don't have this luxury.
So, how to turn research recordings into insights fast enough to actually act on?
I built a Claude skill to pull out mental models and language patterns straight from interviews, instead of spending hours listening and analysing every recording.
The interview transcripts
A short brief on what the feature is meant to do
User segment
From there, the report gathered:
User language
Mental models
Language to avoid
Information hierarchy
Findings
2 mentals models identified
The calculator
They do the math and want the exact savings number so they can subtract the subscription cost themselves.
The gut checker
They don't calculate precisely, but wants to feel like they're ahead.
That led to a hierarchy I could reuse anywhere we talked about savings:
- Show the exact savings amount first (that's the anchor)
- Compare it to the subscription cost
- Give it a time period to give context
The existing saving banner actually got the specific amount. What it was missing was the comparison to subscription cost, and a real time period instead of an abstract 'lifetime' total.
Solution 1
The segment we were targetting
Talabat Pro subscribers
- Who save more money than they paid for the subscription fee
- Who are on a monthly or yearly plan
The goal
Remind them of the value they're getting and that they are saving more than the subscription cost
Considerations
- Leading with the savings: Because we know that this attract attention and users care about it
- Currencies: we operate in different markets so we had to stress test the banner for every currency
- “So far” was a key word to introduce the idea that the cycle is not finished and they can still save more before the end of the cycle
- These banners were shared only to users who have a positive ROI+ and only a few times
Solution 2
The segment we were targetting
Talabat Pro subscribers
- Who already cancelled but still have access until end of cycle
- Who are on a monthly or yearly plan
The goal
Remind them what they're about to lose, and how much they'd actually saved
Considerations
- Leading with the savings: Because we know that this attract attention and users care about it
- Currencies: we operate in different markets so we had to stress test the banner for every currency
- “So far” was a key word to introduce the idea that the cycle is not finished and they can still save more before the end of the cycle
- These banners were shared only to users who have a positive ROI+ and only a few times