Transport for Wales: when station teams said 'all clear' but social media told a different story
Passengers don't always say what's wrong — and when feedback is scattered across social media, surveys and complaints, the signal is easy to miss. Here's how Transport for Wales found what its station teams couldn't see.
During lockdown, Transport for Wales needed to understand how safe passengers felt — but concerns were voiced inconsistently and customer data was fragmented across social media, surveys and complaints. Managing it by hand was too slow to catch emerging trends or get to root causes quickly enough to act.
Passengers don't always raise concerns openly or consistently, so the picture of how safe people felt was incomplete. The feedback that did exist was spread across separate systems — social media, surveys and complaints — with no single way to see it together.
Managing all of that manually was inefficient, and it meant emerging issues could move faster than the team could respond. TfW needed a scalable way to analyse everything at once and reach root causes quickly.
The solution
TfW brought every channel into one platform and trained custom models to surface the themes that mattered.
Every channel in one place
Social media mentions, untagged tweets, customer complaints and survey data were unified in Wordnerds for a single, comprehensive view.
Models trained on their world
Custom AI models were trained to surface common themes across all sources, so the analysis reflected how TfW's passengers actually talk.
Find what you didn't know to look for
Unsupervised clustering surfaced emerging topics, and contextual search let the team explore conversations without guessing keywords.
The impact
Finding the issue the frontline couldn't see
Analysis surfaced passenger safety concerns during the third lockdown — and a closer look at untagged social comments revealed onboard mask-compliance issues that contradicted station-team assurances. After TfW presented the finding to its board and increased enforcement messaging, mask-compliance sentiment moved into positive territory the next period.
Station teams had reported things as 'all clear', but the untagged social conversation told a different story: passengers were flagging onboard mask compliance. Because those comments weren't tagged, the signal would have been invisible to a manual process.
TfW took the evidence to its board, increased onboard enforcement messaging, and watched sentiment on the issue improve the following reporting period — a clear loop from feedback to action to measurable change.
About Transport for Wales
Transport for Wales is a Welsh Government-owned train operating company running rail services across Wales and the English borders, with a focus on passenger safety and sustainable travel.
What would Wordnerds surface from your feedback?
Book a diagnostic and we'll look at your customer feedback together — and show you the themes hiding in it.
So you're reading the footer now? Either you ❤️ Wordnerds or you're desperate for something to read. Either way, CX Corner from Wordnerds is the answer. Fortnightly Voice of Customer bombs dropped in your box. Signup 👇 or find out more.
Avanti West Coast · Transport
Tens of thousands of comments, zero manual tagging
How Avanti West Coast's Voice of Customer team analysed 1.3 million voices across surveys and social — and saved over 7,000 working days of manual tagging.
Greater Anglia · Transport
Tables, toilets, and targeted action
How Greater Anglia mapped feedback to each stage of the rail journey — and found that tables and toilet cleanliness moved satisfaction as much as punctuality.
Sainsbury's · Retail
When 2 million surveys tell you everything and nothing
How Sainsbury's CSAT & Customer Closeness team turned 4.5 million voices a year into proactive decisions — without predefined categories.
