LNER: Collecting customer feedback at scale via Facebook Messenger
The brief.
LNER (London North Eastern Railway) operates high-speed rail services on the East Coast Main Line. Like most rail operators, they collect customer satisfaction data — but the traditional channels (paper surveys, email forms, app-based feedback) all have the same problem: low completion rates from the people who are most motivated to respond, and near-zero rates from casual travellers.
The brief was to find a feedback channel that met customers where they already were, removed the friction of downloading an app or navigating a web form, and captured enough usable data to drive real operational improvements — starting with timetabling.
The strategy.
We built a Messenger-based feedback quiz. Customers initiated it via a link shared on social media, at stations, or via QR codes on trains — and completed it entirely within Facebook Messenger, a channel they were already familiar with. No redirect, no login, no download required.
The design prioritised accessibility and completion. Questions were short, conversational, and broken into natural turns rather than a long form. Automated follow-ups were sent to customers who started but didn't finish — catching a significant portion of near-completions. The result was a dataset that was both larger and more representative than LNER's previous feedback mechanisms, and directly fed into timetable review decisions.
No-friction entry
Messenger as the interface means customers don't leave their native app environment. One tap from a QR code or social link opens the conversation — no redirect, no form, no account required.
Automated recovery of incomplete responses
Follow-up messages were sent automatically to people who started the quiz and dropped off. Recovering near-completions significantly improved the usable response rate without any manual effort.
Actionable data structure
Feedback was collected in structured format from the start — not free text that had to be manually coded. Sentiment, route, time of day, and specific pain points were all captured as data points, not narratives.
What we shipped.
The build was a structured feedback flow inside Messenger, with a follow-up sequence and a data export pipeline that fed into LNER's internal reporting. We handled the conversational architecture, the follow-up logic, and the data schema design.
- ✓ Messenger feedback quiz with conversational question flow (no traditional form UI)
- ✓ Automated follow-up sequences for incomplete responses — sent at 24h and 48h intervals
- ✓ Structured data capture: route, time of day, satisfaction rating, specific service dimensions
- ✓ QR code and social link entry points for multi-channel distribution
- ✓ Data export pipeline feeding into LNER's existing analytics and reporting stack
- ✓ Response volume dashboard giving the team real-time visibility of feedback volume by route
The numbers.
The primary measure of success was data quality and operational impact — whether the feedback was structured enough, and representative enough, to make real decisions from.
Using social channels as a collection point made the feedback highly accessible — people could engage where they already were. The automated follow-ups were particularly effective at recovering responses we would otherwise have lost.
What this means for similar businesses.
The LNER pattern applies wherever organisations need high-quality, high-volume feedback but face low completion rates on traditional channels. Transport, hospitality, healthcare, local government, education — any setting where the audience is diverse and the barrier to downloading an app or navigating a web form filters out exactly the respondents you most want to hear from.
The Messenger approach is particularly powerful when paired with automated follow-ups. The single biggest driver of low survey completion isn't unwillingness — it's interruption. Someone starts, gets distracted, and never returns. Automated recovery at 24h and 48h turns an abandoned start into a completed response.
The structured data capture matters too. Free-text feedback is rich but expensive to analyse at scale. Building the structure into the conversation flow means the data is analysis-ready from the moment it's collected.
Want a case study that looks like this one?
Most engagements started with a 30-minute discovery call. Book yours.