Grainger · Housing
From “I think” to “I know”
How Grainger's leaders went from gut feel to checking their instinct against what renters say.

The Guinness Partnership: from a hunch about ivy to evidence about damp
Chris Haynes noticed a couple of tenants blaming ivy for the damp in their homes. Using Wordnerds, The Guinness Partnership's customer insight team turned that hunch into evidence, and found the kind of problem no survey question asks about.
Last updated 12 November 2026
At a glance
- The Guinness Partnership, with over 70,000 homes across England
- Told by Chris Haynes, Head of Customer Insight
- Survey and complaint comments grouped and scored
- A tenant hunch about ivy tested against damp and mould data
Guinness had thousands of tenant comments and no objective way to read them
Chris Haynes is Head of Customer Insight at The Guinness Partnership, one of England's largest housing providers. Chris joined Guinness after insight roles at Eurostar and NatWest.
Tenants had plenty to say, in Guinness's surveys and in complaints coming through its contact centre, but none of it was grouped or scored. The only option was to read it, which is worth doing every month but isn't objective, because it's easy to remember the comments that match whatever you happen to be interested in that day.
The usual shortcuts didn't help. A few comments tacked onto the end of a report is a superficial version of the voice of the customer, and a word cloud only shows which words come up most. Chris wanted a way to process every comment objectively, so the team could trend it, dig into the detail and find topics nobody had considered.
To read every comment objectively, Chris's team put them all through Wordnerds
Guinness uses Wordnerds to group, score and search what tenants write.
-
Trained its own themes
The team trains its own themes in Wordnerds and tests how well each one fits, so the same issue is grouped the same way every time.
-
Scored how tenants feel
Wordnerds scores the sentiment of each comment, so the team can see how strongly tenants feel about something as well as how often they raise it.
-
Gave the assets team tenants' view
For Guinness's asset optimisation team, Chris measured how often tenants mention parts of a building, like brickwork, windows and doors, and how they feel about them. It gave a team used to building data some customer data to set alongside it.
A couple of comments about ivy pointed to more damp and mould
Then Chris noticed a couple of comments from tenants who felt that ivy and creeping plants were blocking the ventilation on the outside of their homes, and believed that was causing damp. It was exactly the kind of small issue nobody had thought of before.
So Chris built a topic in Wordnerds to catch every comment that mentioned those plants, a fairly botanical one that covered all the ways tenants might describe them, whether or not they mentioned damp. Chris then cross-referenced it against Guinness's own damp and mould data. Homes where tenants had mentioned the plants, in a comment or some other contact with Guinness, had a higher incidence of damp and mould.
A loose hypothesis from a couple of tenants had become something data-led, that Chris could prove or disprove.
Guinness now works from what tenants write, counted and checked
Three changes in how Guinness uses tenant comments.
-
Open complaints ranked by risk
Descriptions of outstanding complaints go through Wordnerds to pick out sentiment and red-flag topics like vulnerability and health and safety, so the complaints team works from a prioritised list.
-
Comments counted, not cherry-picked
Wordnerds groups tenant comments into themes the team has trained, so the team can size issues and trend them over time.
-
Hunches tested against Guinness's own data
When tenants raise something nobody expected, the team can build a topic for it and check it against operational data, as Chris did with the ivy.
Guinness can find what tenants raise that no survey asks, and test whether it matters
The Guinness Partnership uses Wordnerds to group and score every comment from its surveys and complaints, so its customer insight team can find issues nobody thought to ask about. When tenants blamed ivy for damp, Chris Haynes built a topic for it and showed that homes where tenants mentioned the plants had more damp and mould.
Chris is moving Guinness from looking back, with a monthly report on the last four weeks, towards predicting problems. Guinness has built its own models of the likelihood of tenancy failure and of unreported damp and mould, and Chris sees Wordnerds as part of that shift.
Chris's next aim was to treat each Wordnerds theme as an on-or-off measure and test its link with satisfaction. Guinness's driver model for the Tenant Satisfaction Measures can only include what the survey asks about, and plenty of what matters to tenants only shows up in what they write.
About The Guinness Partnership
The Guinness Partnership was founded in 1890 to improve people's lives and create possibilities for them. It has over 70,000 homes across England, from Manchester to Cornwall and London to Liverpool, and provides homes and services to around 150,000 residents in 150 local authority areas.
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 housing leaders ask
How do you find problems in tenant feedback that you didn't know to look for?
Start from what tenants actually say. At The Guinness Partnership, Chris Haynes spotted a couple of comments blaming ivy for damp, built a topic in Wordnerds to find every mention of creeping plants, and cross-referenced it with damp and mould records. Homes where tenants mentioned the plants had a higher incidence of damp and mould. Under Awaab's Law, catching those signs early matters even more, as our Awaab's Law webinar explains.
How do you prioritise a backlog of open complaints?
Let what's in them set the order. The Guinness Partnership put the descriptions of outstanding complaints through Wordnerds, which picked out sentiment and red-flag topics such as vulnerability and health and safety. The complaints team got a prioritised list to work through, so nobody had to make a manual call on which case to tackle first.
Why isn't a word cloud enough to analyse tenant comments?
A word cloud only shows which words come up most. Chris Haynes, Head of Customer Insight at The Guinness Partnership, wanted to group and score every comment so the team could trend issues and find topics nobody had considered. Guinness trains its own themes in Wordnerds and tests how well each one fits.
Can tenant comments explain satisfaction the TSM survey doesn't?
The Guinness Partnership's driver model for the Tenant Satisfaction Measures only includes what the survey asks about. Chris Haynes planned to use Wordnerds themes as on-or-off measures and test each one's link with satisfaction, to find drivers outside the questionnaire that tenants only mention in their own words.