Avanti West Coast · Transport
From Excel sheets to feedback the frontline reads
How Avanti West Coast's Voice of Customer team stopped sorting passenger comments by hand, and now shares what passengers loved, and who they named, with staff each period.

Transport for Wales: the safety worry passengers weren’t reporting
Graeme Baldock’s team at Transport for Wales could track the issues passengers faced, but when few people complained, some stakeholders couldn’t see why an issue mattered. Wordnerds brought social media, surveys and complaints together, and found a safety worry passengers had never raised with TfW directly.
Last updated 13 November 2026
At a glance
- Transport for Wales, the not-for-profit transport company wholly owned by the Welsh Government
- Onboard mask messaging and enforcement stepped up after a finding went to the board
- Social media, surveys and complaints analysed together
- Told by Graeme Baldock, Resource Planning and Analysis Manager
When hardly anyone complained, how could TfW show a problem was real?
Graeme Baldock is Resource Planning and Analysis Manager at Transport for Wales, whose trains run across Wales and on to Manchester, Chester, Shrewsbury and Hereford. TfW already had a fairly mature data model for tracking the issues passengers faced, but some stakeholders still couldn’t see what those numbers meant for a passenger’s journey, especially when only a few people had complained.
Lockdown made that harder. Passenger safety sits at the heart of TfW’s brand, and the team wanted to know whether there was more it could do to help people making essential journeys feel safe. But passengers don’t always complain where a train company can see it, or in the same way, and that’s especially true when the worry is about feeling safe.
What passengers did say was spread across social media, surveys and complaints. Managing all of it by hand took too long to spot new issues and get to their causes while they still mattered.
To hear what passengers weren’t reporting, TfW put every channel in one place
TfW set Wordnerds up to read what passengers wrote, wherever they wrote it.
-
Tracked the tweets that never tagged TfW
Wordnerds tracked social media mentions of TfW and public tweets that didn’t tag it, and TfW added its complaints and survey data, so the team could see all of it together.
-
Trained its own themes in Wordnerds
TfW’s team trained and fine-tuned its own themes, which is less technical than it sounds, so Wordnerds could pick out the subjects that came up across every source and show which ones were pulling sentiment down.
-
Let unexpected topics come to the surface
Wordnerds grouped comments into topics nobody had thought to look for, and the team could search conversations by their context without guessing every word a passenger might use.
In tweets they never sent to TfW, passengers were worried about mask-wearing on board
An early finding was that during the third lockdown, how passengers felt about their safety fell short of the standards TfW sets itself. So TfW asked Wordnerds for a deeper analysis of the public tweets that didn’t tag TfW, where passengers talked about problems they weren’t raising with it directly.
Wordnerds’ report found passengers worried that masks weren’t being worn on board, even though station teams had said they were enforcing the rules. Passengers hadn’t raised it with TfW directly, so it was exactly the kind of problem a low complaint count hides.
TfW’s Insight and Innovation Manager took the finding to the weekly board meeting and asked for action. TfW alerted its onboard teams, and stepped up its messaging and enforcement on mask-wearing.
TfW’s stakeholders can see the passenger behind each number
The mask finding changed how Graeme’s team used passenger feedback across TfW.
-
Added depth to the numbers they report
Graeme’s team used Wordnerds to show stakeholders what each issue meant for passengers, so the figures carried weight even when the complaint count was small.
-
Took the evidence to every stakeholder
The team could point to the issues that mattered most to passengers and present how strongly passengers felt about each one to everyone who needed to act.
-
Kept watch on social media in real time
Wordnerds went on to give TfW real-time insight from social conversations about performance problems and key stations, to find where passengers’ journeys could improve.
Mask sentiment moved into positive territory the next reporting period
Transport for Wales found a safety worry passengers weren’t reporting to it, in public tweets Wordnerds analysed. After the finding went to the board, TfW stepped up onboard messaging and enforcement, and sentiment about mask-wearing moved into positive territory in the next reporting period, helping passengers feel safer while restrictions were in place.
For Graeme, the mask finding showed what a problem with few complaints could look like once passengers’ own words were behind it, and it wasn’t the only one. Several major changes to customer journeys followed from the stories Wordnerds helped TfW tell.
About Transport for Wales
Transport for Wales is a not-for-profit organisation, wholly owned by the Welsh Government. It connects people and communities by train, bus, walking, wheeling and cycling, and its trains run across Wales and on to Manchester, Chester, Shrewsbury and Hereford.
About Wordnerds
Wordnerds turns what customers say into what organisations do. We ingest feedback from surveys, complaints, reviews, calls and social; apply transparent, explainable AI to surface themes and drivers; and serve the insight two ways from one semantic model: full-detail Power BI for analysts, plain-language AI-chat answers for everyone else. Built for UK housing associations, transport operators and regulated sectors that need auditable evidence, not a black box.
Questions CX leaders ask
How do you prove a customer problem is real when hardly anyone complains about it?
Show stakeholders what customers actually wrote about it. Transport for Wales tracked issues in a mature data model, but some stakeholders couldn’t see the impact when complaint numbers were low. Graeme Baldock’s team used Wordnerds to add depth to those numbers and present each issue to every stakeholder, and several major changes to customer journeys followed.
How do you find customer problems that don’t show up in complaints?
Look at what customers say to each other as well as to you. Transport for Wales had Wordnerds track public tweets that didn’t tag TfW alongside its complaints and surveys. A deeper analysis of those tweets found passengers worried about mask-wearing on board, which they hadn’t raised with TfW directly.
How can you check what frontline teams report against what customers experience?
Compare it with what customers write unprompted. At Transport for Wales, station teams said they were enforcing mask rules, but public tweets analysed by Wordnerds showed passengers were worried about mask-wearing on board. TfW took the finding to its board, stepped up onboard messaging and enforcement, and sentiment moved into positive territory the next reporting period. We use it as the worked example in our five-step loop for acting on feedback.