How do you improve your TSM score?

What is the qualitative TSM benchmark?

The qualitative TSM benchmark is a sector-wide reference dataset of 135,000 free-text tenant comments collected from 18 anonymised UK housing associations, categorised by Wordnerds across TSM areas, the 26 HHSRS hazards, vulnerability indicators, fixtures and fittings, and a "financial" sixth bonus category that mirrors what tenants actually talk about (rent, service charges) but the TSM survey does not directly ask. Built jointly by Wordnerds and Housemark in 2026 and presented to a 173-strong UK housing audience.

The benchmark surfaces three KPIs per category and sub-category: proportion-of-volume (what percentage of tenant comments mention this issue), sentiment score (how positively or negatively tenants are feeling about it), and the gap between each contributing housing association's score and the sector average. Sizes range from small (under 5,000 homes) to extra-large (over 25,000 homes), so associations can compare against peers their own size as well as the sector overall. No individual association or tenant is identifiable in the published benchmark; the contributing 18 are anonymous to each other and to readers.

Does "fairly satisfied" actually mean satisfied?

No. Across 7,500 free-text comments from tenants who selected "fairly satisfied" for their TSM overall-satisfaction score, the average sentiment score is negative 40 — closer emotionally to tenants who scored fairly dissatisfied than to those who scored very satisfied. The comments themselves talk about long waits, unclear communication, and chasing updates: the language of dissatisfaction, not of satisfaction.

The methodological problem is that the TSM percentage-satisfied score combines "fairly satisfied" and "very satisfied" into one figure. A landlord reporting an 80% satisfaction score may in reality have 35% very satisfied and 45% fairly satisfied — two groups that behave very differently in operational and trust terms. Chris frames the "fairly satisfied" group as sitting in the "zone of indifference": fragile loyalty, low confidence, and quietly accumulating friction that surfaces later as an escalation or formal complaint. Treating this group as a win misses the early warning signal; treating it as the most actionable group is where the real TSM uplift comes from, because the comments name the friction points directly.

What pushes a TSM score up the most?

Three things push a TSM score up most, all of them communication signals rather than big structural-investment issues. When a tenant receives a quick response — the response, not the repair itself — the TSM score they give goes up by more than 41 points. When a tenant feels their housing association keeps its promises (saying what it will do, then doing it), the TSM score goes up again by close to 41 points. When a tenant is shown empathy by housing association staff, the TSM score goes up by 38 to 39 points.

These were the top three drivers across all 272 different aspects of the tenant experience analysed in the benchmark. Their unifying feature is communication: even "keeping promises" is fundamentally a communication signal, not a delivery signal. The implication for housing associations is sharp — the highest-leverage TSM improvements are not the big practical investments. They are the changeable, communication-led behaviours that stick most reliably in a tenant's head when they come to fill out a TSM survey.

What pushes a TSM score down the most?

Three things push a TSM score down most, two of them more nuanced than the obvious operational issues. The single biggest negative driver is a tenant feeling insecure in their home — when a tenant is worried about their security, the TSM score drops by around 37 points. When a tenant feels the housing association has caused or exacerbated a health issue, the TSM score drops by around 31 points — less than security, which Steve flags as itself a finding. When a tenant feels ignored or not listened to, the TSM score drops by 26 to 27 points.

Combining the up-drivers and the down-drivers gives a clear operational mapping. Communication (quick response, kept promises, empathy, being listened to) is the dominant lever in both directions — present in five of the top six positive and negative score drivers. Structural risks (security, health) are heavyweight individual issues but narrower in incidence. The benchmark lets housing associations identify which of these drivers their own free-text comments lean on, and how their proportion-of-volume and sentiment scores compare to the sector average per driver.

How does customer effort affect tenant satisfaction?

One in six tenants are expending unnecessary effort to get basic problems resolved — whether that's waiting, chasing, having to explain, having to interpret, or physically having to intervene. Customer effort is one of the strongest predictors of overall satisfaction; the easier the experience feels, the more loyalty it generates, and the harder it feels, the faster trust erodes.

The benchmark applies the Clark and Bryan customer-effort framework to tenant comments, breaking effort into four types. Time-based effort dominates at 55% of effort-related comments (long waits, long calls, chasing progress). Cognitive effort accounts for 27% (interpreting unclear letters, understanding updates, navigating complex processes). Emotional effort accounts for 26% (stress, anxiety, feeling ignored or dismissed). Physical effort, although a smaller group at 4%, captures cases where tenants have had to take physical action themselves — going to offices in person to get things sorted. The categories are not standalone; time-based and emotional effort in particular interweave (waiting on hold drives the stress as well as the time loss). Effort is one of the fastest things to improve and rarely requires a big budget — better clarity, better communication, more consistency. Reducing waiting and chasing through proactive updates can halve the effort score for many tenants.

How should housing associations act on this benchmark?

Three uses unlock once a housing association has its own data sorted by volume-percentage and sentiment against the benchmark. First, prioritise more accurately: knowing whether a 3% no-shows rate is brilliant or catastrophic compared to the sector lets leadership and frontline teams agree on which issues actually deserve attention this quarter, instead of choosing in a vacuum. Second, highlight best practice: when a customer experience team is performing above the sector average on a driver, the people doing that work deserve to know it — and the organisation can systematise what is working.

Third, separate what is movable from what is not. Some things (geography, property type, building age) are immovable; others (communication speed, kept promises, empathy, clarity of letters) are highly movable and are exactly the levers the up-driver and down-driver findings identify. Sarah's recommendation for the report-out: pair the proportion-of-volume and sentiment KPIs with a clear note on what your team can actually change in the next quarter, what you cannot, and which sector-average gap is the highest-leverage one to close first. The benchmark turns "we have a problem with X" into "we have a problem with X, and the sector does too — or doesn't" — and that contextualisation is what makes prioritisation decisions defensible to leadership and to the regulator.

Full Webinar Transcript

Sarah Wilson: Welcome everybody. We're absolutely thrilled to have you all with us, thank you so much for joining. This is our first webinar of 2026 and we are so excited about this launch and this project. Housemark and Wordnerds have done bits and pieces for the last few years together, but this is our first big joint project and we just can't wait to share it with you.

So, who are we? We're very lucky — we've got a mixture of friendly faces, people who've been to webinars before, brand new people, and customers. The difference in the pictures on the slide is really stark: there's Housemark at a beautiful conference, presenting; and there's Wordnerds mid-barbecue at what looks like a garden party.

Housemark are the UK's number one benchmarking standards body for the social housing sector. They've been the leaders for years in setting the standard across the industry, and we're delighted to have Chris from Housemark joining us today. Wordnerds are the experts in customer feedback analysis — however people speak, whether that's sarcasm, misspellings, or emojis, everything is taken into account with the model we use. We felt this was basically the dream partnership: linguistic analysis mixed with the sector standards that Housemark have got.

Sarah Wilson: Who have you got today? That's me — I'm the ironically leopard-print-loving vegan on the left. I'll be taking you through the introduction and some of the closing parts.

Then we've got the man our CEO Pete described as "absolutely reeking of experience" — which I think really says it all about Chris. We feel privileged to work with Chris. He's super experienced in the customer experience space, and one of the things I love about his approach is that he's about the doing — not just reporting the numbers and then closing the laptop. He's very much about the insights, the actions, and moving things forward.

And then we've got Steve. Steve actually doesn't need any introduction — you've probably seen him on TV already. He made his debut on BBC's Only Connect. His superpower is finding connections between utterly random topics — on Only Connect that meant 17th-century poets and types of pasta. He's the backbone of everything that happens at Wordnerds: head of the insight and innovation team, and he worked directly on this report with his team.

Sarah Wilson: A few bits of housekeeping. Yes, this is being recorded, and everything will be shared with you after the call — so don't worry about taking notes. The other thing I'd say is that Steve loves to show very intense visuals, so if you have another monitor, now would be the time to give us the big-screen treatment.

So what will you get from today? You've signed up for a nice hour-before-lunch session, and in the hour we're hoping to cover: why we built this, the reasoning behind it; then a tag team of Steve and Chris going back and forth on the framework, the data, the dashboards, some key findings from the report and what that means for moving forwards. Then I'm going to show you how you can actually use this data in your own day-to-day. And then a couple of really exciting bits to stick around for at the end — a little care package for everyone, and a custom offer just for people who've attended.

We've left loads of time for Q&A — that tends to be a really juicy part. We want as much interaction as possible.

Sarah Wilson: Where are we starting from? We're having conversations all day, every day with housing associations, and what I'm hearing is that there are two schools of thought. You've got tons and tons of tenant comments coming at you from all angles, all different places, and you really need to be on top of every single source. You might handle that in one of two ways. Typically you're either old school, looking at keywords, taxonomies, setting up custom tags in Excel — which we all know is a bit of a nightmare. Or you've made the investment in some expensive AI-powered solution that promises insights, and probably just tells you that repairs and communication are your top categories. That only gets you so far.

You go away, you work really hard, you build reports, you do dashboards, you give insights, you put your name against these recommendations — and everyone seems quite happy. But then a director might pop up a hand on Teams and say, "Is 3% of no-shows actually a problem? What is everyone else seeing?" And suddenly your mind goes blank. You're making really big decisions ultimately in a vacuum, and you don't know if that 3% is brilliant or catastrophic. It's an educated guess, but it's still definitely a guess. The home is the one place we want to feel safe — the world shouldn't be able to get you there; it should be a sanctuary. These decisions are showing whether the sanctuary holds up or falls apart.

Sarah Wilson: The satisfaction gap is something we've been talking about more and more at Wordnerds. The problem is that your tenants aren't just comparing you to other housing associations — they're comparing you to things like Amazon, telling them their parcel is two stops away, showing them where the driver is, updating them when it's delivered. That is the satisfaction gap, and it's really tricky.

The most recent Housemark report had a statistic in it that terrified me. There is now a 20-point gap between tenant satisfaction with repairs and overall trust in the landlord. You can do all the work, fix the leak in 24 hours, tick the box, smash that KPI — but you're still failing the tenant. Why? Because nobody told them when the van was coming. They've waited in all day, used a day of annual leave, potentially lost money — or the contractor did the job but didn't introduce themselves properly, didn't bring shoe covers and trampled through the house. Which, yeah, is a personal bugbear of mine. The follow-up was non-existent. You fixed the property but you traded the relationship.

In recent reports, 100% of C-graded landlords got slammed — not for slow repairs, but because tenants didn't feel heard. The regulator isn't just asking for a score anymore. Walking into a meeting and saying "we've got our TP01 at 68%" doesn't cut it. You need to map what they're measuring — which is speed — against empathy. And this benchmark finally lets you say that. You can go in confidently and say, "We complete repairs faster than 80% of the sector, but we know that our communication is terrible, and that's what we've got to work on." That's really powerful.

Sarah Wilson: So what's the solution? Two lovely pictures from the 1950s. I don't know if anybody knows the backstory, but I wanted to pop this one in because I think it's really important for what we're talking about today. It's what we call the sector's Volvo moment.

In 1959, Volvo invented the three-point seatbelt that you now use every single day, every time you get in the car. They could have kept that patent and obliterated the competition. But they didn't. They open-sourced it. They gave it to every car manufacturer on earth, because they decided that saving lives was more important than making money and taking market share.

Today, we're going to show you the work from 18 housing associations who've done just that. They've put their hand up, raised their neck above the parapet, and they've shared 135,000 tenant comments — because for them, improving outcomes for the people that matter is more important than hoarding data. So for the first time, you can actually see where you sit. There's no guessing, it's just knowing. And with that, over to the smart guy. Over to you, Steve.

Steve Erdal: Sarah, thank you so much. And thank you all so much for being here. I love that seatbelt metaphor, Sarah. We work in a lot of different industries, and we've found that this is the most collaborative sector that we work in. And we are so grateful to the 18 housing associations who've provided their data for us to start to build this first iteration of the benchmark. We've obviously made the decision not to name them — they are anonymous, but they know who they are.

That group of 18 housing associations gave us their comments from their TSM survey — just over 135,000 real comments from real customers. We got a little bit of metadata to attach to that, so we could analyse the factors involved: the TSM score, the size of the housing association (under 5,000 homes being small, all the way up to over 25,000 being extra-large) so we could show different issues affecting different sizes. It's really important just to reiterate that at no point is any remotely identifying information — either of individual tenants or of the housing association in general — going to be published. Nobody gets to see the individual comments. We take very seriously the responsibility they've given us.

Steve Erdal: Those 135,000 TSM survey comments were then passed into the Wordnerds platform, and we started the process of categorising the data into key issues with customer experience that are likely to come up in the TSM data. We put them into individual TSM categories — mirroring the questions asked in the TSM survey. We categorised them in terms of the HHSRS hazards — the 26 different hazards, several of which you are now required to categorise and uncover in your data, with the others coming soon over the next couple of years. We looked at vulnerabilities — disability through to economic vulnerability and so on. And then fixtures and fittings — anything that could potentially go wrong or need to be repaired in a home.

That process of categorising the data is based not on the actual words people use, which they could use in any context, but on what they meant. What the AI allows us to do is to base it on what they meant. So we can take this really nebulous, colloquial, confusing data set of real people communicating, expressing themselves in any way they choose, and turn it into a benchmark — something we can measure, track over time, and understand.

We passed that data into Power BI and gathered the KPIs we thought were particularly important to allow housing associations to understand where they sit relative to this benchmark group. We looked at proportion of volume — how much of the data mentions an issue. We looked at sentiment — using AI to put a sentiment score on each area. And then for each of those proportions and sentiment scores, we were able to give a high score (the highest amount any of the 18 housing associations experienced), the lowest amount, and an average.

That allows our customers, as they start to take these benchmark numbers and use them as a counterpoint to their own data, to prioritise more accurately, to understand which numbers can be moved, and to highlight best practice. If your customer experience professionals are talked about more positively than the rest of the sector, those professionals deserve to know. So it lets you see not just what you can improve, but where you're already doing really well.

Steve Erdal: I'm going to briefly show you the dashboard we used to highlight some of the things Chris and I will talk about in more depth. The first thing you'll see is all of the data united together. We can see the categories: maintenance and repairs, safety, communication and engagement, estates and neighbourhood, and so on. We've tried to mirror the key question areas of the TSM. There's actually a sixth category here — we added financial. That's not something you're asked about in the TSM survey (rent, service charges) but our customers told us it was really useful, so we've added a little bonus sixth category.

You can see we're immediately looking at it in terms of volume and sentiment: how many people are talking about each area, and how they're feeling about it. If there's an area that interests you, you can click in. For example, communication and engagement — within that, we've broken it down into all the different aspects of communication and engagement a customer might experience. Click on "kept informed" and I can get an understanding of how being kept informed is affecting customers. I can look at the proportion of data that mentions it: over time, it's usually between 4 and 5% of customers talking about issues being kept informed. I can look at what's crossing over with it — things like service charge, fair and clear pricing, visibility of landlord.

So as a housing association you can start at the top with the big numbers presented to your senior leadership team, then very quickly follow the thread down: who's talking about communication and engagement, what issues are coming up within it, how common it is — and then use the benchmark to say this level of conversation is normal, or is over what we'd expect, or is less than what we'd expect.

We can also present this in the form of a journey. This is our repair journey framework. When something's linear like a repair, we can start to look at the journey from the customer's experience — starting with making an appointment, all the way through to the follow-up. As the line goes down, people are getting more negative; as the line goes up, they're getting more positive. And we've broken it down by size of housing association. So we can see, for example, that operative conduct tends to be slightly more negative with smaller housing associations — but they're actually ahead of the curve when it comes to the follow-up and aftercare. That ability to slice it allows you to make better decisions, prioritise more effectively, and ultimately find the things that will improve your customers' lives. With that I'll pass back to Sarah.

Sarah Wilson: Thanks so much for that openness, Steve. We can see the possibilities are endless with the different categories. We've had a few questions I think would be good to address. People are asking how we've categorised the comments — is it just AI-powered? And there are questions around the accuracy and the sentiment, because people have tried AI before and the sentiment is very unreliable. All the things we hear every day. Steve, would you mind talking a little bit to the methodology behind the project?

Steve Erdal: Of course. We 100% feel your pain on that front. We understand what a bad rep AI has at the moment. I actually heard that it's started in the playground — kids are starting to refer to AI as "lies": they'll say "oh that's AI", meaning "oh, that's not true." That's the vision people have of AI right now, particularly around generative AI.

What we've done is unite that generative-AI-style interface with robust machine learning in the background. So when we're talking about AI, we're not talking about ChatGPT. We're talking about a way of structuring data that interacts with the AI to tell it what to do. That's a really important distinction, because with AI generally, it will give you a different answer on Monday and Tuesday with the same data. That is the worst thing for insights professionals, because they need to be able to point to the issues. As Mark said in the chat: garbage in, garbage out — that's precisely what we're looking to avoid.

That's not to say there are never mistakes. You can disagree with a fellow human being about whether something should be in a theme. There's no perfect way of doing this — including getting a human being to read through and tag. They'll also make mistakes; they'll likely be less consistent than AI. There's always a trade-off. But what we've tried to create is a way of using AI to supercharge what a human being does, while also making it scalable, transparent — so you can see what it's done and why — and more robust than you'd expect with usual AI.

Sarah Wilson: Really helpful, thank you, Steve. We'll delve into this in more detail in the Q&A. Now over to you, Chris. What Chris is going to do is follow up on what Steve said and take us through some of the findings from the report — some of the interesting things he's picked out that he thinks are valuable to share.

Chris Elliott: Thank you. Good morning everybody. I'm delighted to have been invited, and delighted to have collaborated with Wordnerds on this report. Just to introduce me: I head up the Customer Experience Consultancy Pillar at Housemark. I work with loads of clients on their TSMs and their transactional surveys, among other things. I've worked in customer experience for over 15 years now, all around the world, in and out of the wonderful sector we operate within — just helping organisations improve their customer satisfaction and loyalty, and all of the benefits that brings.

Why is this collaboration so important and so exciting? People who know me well enough know that I spend most of my working day talking about the importance of taking TSMs beyond just compliance. I talk a lot about the importance of doing stuff, taking action on the back of your surveys. And I talk a lot about how comments are an amazing tool to help you identify what you need to do to improve. So the opportunity to work with the team at Wordnerds on something genuinely unique was too good to miss. I'm really proud of what we've pulled together. A big shout-out to my TSM clients — you know who you are — who have collaborated with me on this project. Without your help, we wouldn't be in the position we're in today.

Chris Elliott: Onto the first theme. When we consider TSM surveys, and specifically the reporting of the scores, the percentage-satisfied score is a score formed by combining people who scored "fairly satisfied" and "very satisfied" together. So if you've got an overall satisfaction score of 80%, in reality that might be 35% saying they're very satisfied and 45% saying they're fairly satisfied. So let's start with a really important question: does "fairly satisfied" actually mean satisfied?

What we found — and what I've thought for a long time — is that no, not really. When we looked at the data and analysed over 7,500 comments from tenants who selected "fairly satisfied" for their overall satisfaction score, what's really striking is that those comments alone carry a negative sentiment score of 40. And that sentiment score sits much closer to tenants who scored fairly dissatisfied than to those who scored very satisfied.

As we've said, the methodology groups these two together. But the reality is that, emotionally speaking, these groups behave very differently. When you actually read what tenants are saying, the fairly satisfied group talk about long waits. They talk about unclear communication. They're chasing updates. Just exactly the sort of thing we'd normally associate with dissatisfaction.

Chris Elliott: Why does this matter? From a customer experience point of view, customers who score fairly satisfied with their service providers sit firmly in what we call the zone of indifference. It's OK at best, but in reality they're not very happy — they're just not unhappy enough to shout about it at this stage. This is the customer group where confidence is low and loyalty is inevitably fragile.

That's where the operational risk comes in. If you treat this group as satisfied just because of the score, you're missing those key early warning signs. Problems then fester quietly in the background and only bubble to the surface later, often in the form of an escalation or official complaint. The other risk is misinterpretation. If leadership teams only look at the scores, they're assuming things are better than they really are. There's a risk of allocating resources to the wrong issues, and missing precious opportunities to relieve those pressure points that tenants are repeatedly trying to tell us about.

Where does the opportunity lie? The good news is that I actually think this group is one of your biggest opportunities, because what they're giving you is not a vague dissatisfaction. They're giving you the root cause of it. Their comments are telling you exactly what needs to be fixed in order for them to move from being fairly to very satisfied.

So rather than treating them as a win, treat "fairly satisfied" as an early warning. Look at the comments, look at the sentiment, and use them to identify the friction points they're telling you about. My experience tells me that when housing providers do this really well, they often find that relatively small targeted improvements, particularly around the areas of communication and responsiveness, can convert this group pretty quickly. And that's where you're going to see the real uplift in your TSM scores, and genuine improvements in the overall tenant experience. So in short, let's banish the view that fairly satisfied means satisfied. What it actually means is that we're nearly there, but we're not quite. And the qualitative feedback you get is the tool that will tell you how to close that gap.

Sarah Wilson: Fascinating, isn't it, Chris? Such an interesting idea. I think when we hear "satisfied", we take it as that — don't we? But really interesting. We're going to hand back to Steve, who's going to take this into more detail and give us some of the things we can do to actually affect the TSM scores. What are the drivers, both positive and negative, that will influence the TSM scores? So over to you, Steve.

Steve Erdal: Thank you so much, Sarah. And thank you, Chris. I think the point you made there — about how, because of the way the TSM score is calculated, we're encouraged to group very satisfied and fairly satisfied together when in fact there's a lot of difference between those two positions — is really important.

One of the things we really looked at in the analysis of that benchmark data was what, within that data, actually impacts the TSM score. So if somebody's talking about something in their free-text comments on their TSM, what does that thing tend to do to the overall TSM score they give? We're going to start with things that push the TSM score up — when this thing is true, the TSM score is likely to be higher than you'd otherwise expect.

The top three in that area: when a customer receives a quick response, the TSM score they're likely to give goes up by over 41 points. That's a phenomenal amount, and the highest of anything we found. And to Chris's point, they were not talking about the actual repair being done quickly — it's just about the response being quick. That was the thing, of all the positive aspects of the customer experience, that was the best indicator of a score going up.

Similarly, when a customer feels their housing association says what they're going to do — when they keep their promises — the TSM score goes up again by almost 41 points. And when a customer is shown empathy by housing association staff, the TSM score goes up by 38 to 39 points.

Those are the top three things, out of the 272 different aspects of the experience, most likely to impact TSM score positively. What's interesting is that they all share one thing: they are all communication things. Even "doing what they say they're going to do" — it's not just about doing stuff well; it's about doing the stuff you said you were going to do. That's pertinent because, to Chris's point, these are not the big kind of structural, practical, massive-investment issues. These are the things more changeable by housing associations, and they're also the things most likely to stick in a customer's head when they come to fill out their survey.

Steve Erdal: Conversely, if we look at the things likely to drive the score down, it's a slightly more nuanced picture. Actually the thing most likely to push the TSM score down was security. If a customer is worried about their security — if they're not feeling secure in their home — the TSM score goes down by about 37 points.

When a customer feels like their housing association has caused or exacerbated a health issue, that was the second biggest. The TSM score goes down by about 31 points. Two things there: firstly, entirely understandable that a customer would feel strongly if they felt their housing association caused a health issue — but that actually pushed the TSM score down less than when they feared for their security. That was a really pertinent finding.

And going back to what we saw in the positives — when a customer feels ignored or not listened to, the TSM score generally goes down by 26 to 27 points. Hopefully you can see what we're doing here: we are finding the issues not just that customers are talking about — not even necessarily that they're talking about negatively — but what are the things actually impacting and influencing the ultimate score they give in their overall Question 1. These are the things that matter most in their TSM score.

With that, you can then go back to that prioritisation: finding areas where you can improve, and uniting that with what happens overall. What are the elements that most housing associations are finding affect their TSM score, and how are ours different? That allows you to make the best possible decisions when it comes to improving the customer experience. And with that, I'll pass back to Sarah.

Sarah Wilson: Wow, thanks Steve. What strikes me about this is it's all about trust, right? We've all got interactions in our day-to-day lives. I've just had one with my bank. It's about sending a message even if there's no update, or just doing what you say you're going to do. People want updates. Even the speed of response increases the score that much — fascinating.

One of the housing associations I worked with recently implemented a campaign to make sure they were shouting about things you'd suggested but they didn't do anything about. They'd go back to you and say "you said we did", "you said we didn't, and this is why" — and give reasons why they didn't do that, what the thinking was. Even that tends to increase trust, the feeling of trust in that relationship, and the perception of the landlord. It's more of an equal partnership rather than that traditional paternalistic view.

Lots to think about there, isn't there? We could break it down into more subcategories — look at things like security. Now we're going to hand back to Chris. Chris is going to take us through one of our favourite frameworks here at Wordnerds, which is the effort framework. Some of the things you can do, taking that idea of communication forward. What does communication look like? How can you break it down in more detail, and what are the key drivers to improve the overall customer experience? Back to you, Chris.

Chris Elliott: Thanks, Sarah. This one's a big one for me, and I make no apologies for consistently banging on about it over the past 15 years that I've been involved in customer experience. For me, it's absolutely central to understanding the tenant experience.

Our research has shown that one in six tenants are having to expend unnecessary effort just to get the basic problems resolved. Just think about that for a minute — one in six who feel they've got to work way harder than they should. Whether that's waiting, chasing, having to explain, having to interpret, or even physically having to intervene. It's quite a number.

Why does customer effort matter? From a customer experience perspective, customer effort is one of the strongest predictors of overall satisfaction. If you think about it in your own day-to-day life — when you have a transaction with a service provider, if it feels easy, you tend to love them for it. If it feels really hard, your frustration increases, your trust erodes — and it erodes really quickly. High effort creates real-world consequences and it impacts the bottom line: more repeat calls, more chasing, more frustration. Ultimately it places big pressures on frontline staff too.

We see a clear correlation here with declining TSM scores. When I'm reporting landlord TSM scores and they also ask tenants how easy they are to deal with — and if you don't, you really should consider asking it, because you'll get a lot from it — you'll always see a direct correlation between ease and overall satisfaction. When tenants perceive they're putting in way more effort than they should, their overall satisfaction drops with it.

Chris Elliott: So what does the report show? Customer effort is no longer the new kid on the block — there's considerable academic research devoted to the topic, and lots of models. For our research, we've adopted the Clark and Bryan framework, which breaks effort down into four types.

Time-based effort dominated, accounting for 55% of the effort-related comments, with key themes including long waiting times, long calls, and chasing progress. Cognitive effort accounted for 27% of comments — where tenants were trying to interpret unclear letters, trying to understand updates, and ultimately trying to navigate difficult, complex processes. Emotional effort accounted for 26% of comments — stress, anxiety, feeling ignored or dismissed being the dominant themes. And although it's relatively small at 4%, the physical effort group captured cases where tenants had to take action themselves and physically go to offices in order to get things sorted.

For me, the takeaway here is really clear. Effort is not a marginal issue. It's hardwired into the everyday experience that tenants have — be that good or bad.

So where does the opportunity lie? All is not lost. The good news again is that effort is one of the fastest things you can improve. And it gets better — because not only can you improve it quickly, it also doesn't require a big budget. What it requires is better clarity, better communication, and ultimately more consistency.

For example, reducing waiting times and chasing through proactive updates can halve the effort score for many tenants. Improving the clarity of letters or texts removes that cognitive effort immediately. Supporting vulnerable tenants reduces emotional and physical effort, which we know has a really big impact on trust. And by tracking effort themes — not just effort scores — you can see exactly where the friction's occurring in the journey, and you can redesign your services accordingly.

Look at it this way: effort is a little bit like an invisible tax that tenants are paying for poor systems. But what our report shows is that when you remove that, you don't just make life easier for tenants, you're making life easier for your teams too. And that's a really big win-win that everybody would grab with both hands if we had the chance. Think about it: less hassle, less chasing, fewer complaints, happy tenants, and a housing service that actually feels more human. What a wonderful idea.

Sarah Wilson: Thanks, Chris. You've described utopia, haven't you? The dream. Tell us a bit more about the categories and the interaction between them. How much overlap do you see in your day-to-day dealings with housing associations on things like time and cognitive load and stress? I imagine that particularly for vulnerable people, that's a real pull.

Chris Elliott: There's an argument that says they're not separate. They very definitely all intertwine. Time-based effort and emotional effort are very, very closely aligned. If you're waiting on the phone for a long time, you can feel the emotional stress and anxiety, particularly if it's something serious you want to get resolved. So they're definitely not standalone. They're very much weaved within the overall concept of effort. It's a really good model — they're clear pillars, but they're very much combined together.

Sarah Wilson: Really fascinating. Now you've got the idea. Thank you so much, Steve and Chris — I hope you've found that really valuable. Now you've got the idea of why the benchmarking is so important, how you can compare yourself, why this is so important, and what you can do today to make sure you're on the path to improving your TSM scores.

I just wanted to touch on the process of doing this yourself and what the impact can be for your organisation. The first thing you have to do is sort by the volume percentage and then sentiment for comparisons, and have the categorisation piece in place. Once you've got that, you can do three main things. It unlocks three main pots of information.

The first is that you can prioritise much better. At Wordnerds we've been doing these reports for a long time. We've been categorising for years, haven't we, Steve? But a lot of the time we'd show the report and give examples of what we think will improve things — and people didn't really know what to prioritise, what they could actually do, what they could do today. Knowing you've got a benchmark there, and a particular problem with this issue, allows you to prioritise in a much cleverer, much more strategic way.

The second thing it helps you do is see where you're doing well. Highlighting best practice in your organisation is crucial — and the flip side is obviously seeing the areas you can improve in.

The third — and again, crucial — is that without a benchmark, you don't really know what you can move and what you can't move. You don't have that plan in place for what you can actually make a change with and what you can't. Some things are just immovable — geography, property types, sometimes you can't do anything about those. It's knowing what you can influence. As Samantha said in the chat, communication is a really good one to work on, which has a really strong impact on the perception of the landlord. Knowing that really helps you move forward and improve that customer experience.

Sarah Wilson: What will you receive? We're giving away a little care package for everybody on the call. We're going to make this a whole series, because we've got so much to share. Chris and Steve will agree they've barely scratched the surface on what they can be doing with this. They've given you a few examples of things you could look at, but anything they've shown you, you can do more of a deep dive on. So we're going to kick this off with a care package: tomorrow you're going to receive visualisations from the original report, with an executive summary with all the benchmarks. You can see straight away how you sit against the housing associations. And obviously the recording and lots of resources.

Next steps. Maybe you're looking at this thinking, "this looks brilliant — I'd love to do this. Can I get my data in here?" We are offering, for the first time ever — and Steve is the person doing all these reports, so we've begrudgingly pushed him into it, twisted his arm — that we'll give anybody who wants it their own custom report.

If you see this and you think you're interested, click the link, register your interest. That's all we ask. Just put your details in, we can have a quick chat. As you can see, we're quite a friendly bunch, so it's very much no pressure. What we're hoping to do with this piece of work is have as many housing associations opt in as possible. So Cohort 1 runs to the end of March — obviously this year of the TSMs — and then we'll launch year two at the start of April.

If you register your interest and you decide this is for you — you want to see your weaknesses, how you're sitting in the industry — you can get a full custom report just for you. You get a full dashboard like the ones Steve's shown. You get a report written and overseen by Steve himself. You get a workshop where we bring in people from your organisation and talk through some of the findings, some of the learnings. Steve even does a quiz now, which has been very popular.

That usually costs £5,775. We've decided to do a launch price — we know that £5,000 or below is fairly easy to get through procurement — so we've gone for £4,900 for this special report as a launch offer. We are super keen to expand this pot to make it as valuable as possible, to improve the averages, improve the regional breakdowns. So we'd love for you to put your hand up and join.

And as Steve said, this benchmarking data is completely anonymised — not identifiable by either organisation or by any individual. Just remember that 18 other housing associations just like you have already charted these waters. They've been there, they've done this, they've got the same budgets, same resource constraints, the same infosec teams asking the same questions — we are very used to handling this. So please, don't be put off; it's not a brand new project. We've got 18 other people who've done it already. If you're interested, click the button, fill in the form by the end of February, because we're hoping to close Cohort 1 by the end of March.

Sarah Wilson: Question time — my favourite part, especially because I get to feel all the questions too. We've left seven minutes; we might run over, but if you've got a hard stop at 12, we have the recording.

So a really chunky one — I think this is a great one for Chris. It's around how other people are asking qualitative questions as part of the TSM questionnaires, and best practice around that. A lot of people shy away from asking too many questions because they don't want loads of text. So what's the best practice that Housemark sees for this? What's the compromise between putting too much in and not getting enough?

Chris Elliott: A really good question. In a perfect world, we'd love to flood questionnaires with more qualitative questions, but we're competing for tenant time when we're doing these surveys — increasingly difficult competition — so we've got to be efficient with our questionnaires.

For me — and anyone who knows me well enough knows that I talk a lot about the importance of, if nothing else, asking on the TP01 overall-satisfaction question: "Why did you give that score?" That will give you a tremendous amount of information in itself.

I like asking other questions too. I, unashamedly, am an advocate of ease — I'd always try and consider including that. And I think also a qualitative probe. The probe on this one — I always like doing it slightly differently. If you want to understand how you can become easier, ask a direct question. So instead of "why did you give that score?", ask "how can we become easier?" It's a more positive, more cooperative-style question.

So overall satisfaction is really good. Ease as a quant and a qual I'd definitely recommend. The thing for me is, with qualitative questions — or just all questions in general — I know we're bound by 12 that are written in stone, which we have to do in the prescribed manner. Any questions you ask, there are two things you can do: you can ask questions that are nice-to-know — they'll give you insight, but they don't really deliver any actionability; or you can ensure they're need-to-knows — things you are going to do something with. That subtle distinction is really important.

So overall, satisfied — why did you give that score? — and if you're wanting to learn how to improve, ask the question directly: how can we improve?

Steve Erdal: If I could just really quickly add there, Sarah. Debbie has written something really interesting in the chat during the effort portion: "Your effort should always be more than the customer's. Essentially, the customer should only have to raise an issue and receive the result. The rest is up to us." That's true of the questionnaire portion as well. And that's what technology allows you to do.

In the past, the only way of getting a computer to understand a human was to quantify it — to ask hundreds and hundreds of quantitative questions, drop-downs, pick from one to five. We're now at the point where we can say to customers, "express yourself however you want", and it's up to the computer to understand you, rather than you having to make yourself understandable to the computer. So smaller, fewer questions that allow the customer to expand as much as they want to on any issue affecting them — that's the best possible way of getting that feedback.

Sarah Wilson: Thanks for that. Debbie has asked another linked question. Does this approach work in theory across all surveys and consultations, not just TSMs? And we had a couple of linked questions around things like picking up handwriting and consultations. Steve, would you mind taking this one? We've tackled this a few times at Wordnerds.

Steve Erdal: Brilliant question — and so many different aspects of this are true. It works really well for TSMs. As I said before, we also automatically categorise the data into the HHSRS hazards, into customer vulnerabilities and so on.

The benchmarking thing specifically is maybe less pertinent to the HHSRS, because that is more about finding every possible instance of, for example, damp or mould, or personal hygiene, or any of the issues a customer might express at any point. So what AI can do there is give you a safety net. If a customer experience professional happens to miss somebody talking about an HHSRS hazard, it can be a second wall of defence — because as soon as you receive it, the clock has already started. So the quicker you can get that, the better.

For things like that, it's harder to say "our damp and mould isn't bad because it's below average for the sector". It's more about getting every single instance of that damp and mould. But the concept of categorising the data and getting human beings working with the AI to understand it — and to give you another lens to look at the data through — works with all these different aspects, and we have trained models to do that.

We've obviously just looked at the TSMs, and all the data we used was digital. Handwriting-reading products are getting better and better, so that's becoming more and more possible. But in practice, it's not just TSM data you're looking at, and customers are weird — they will tell you something positive about your staff in the middle of a complaint, and they'll tell you about a health issue you weren't aware of in the middle of a piece of TSM feedback. So the broader question — about uniting all these different data sets and then using this kind of benchmarking to get an over-arching understanding of where our key issues are — I think bringing together as many different data sources as possible naturally gives you the best possible outcome for your customer.

Sarah Wilson: Just worth adding to that — we do a lot with voice recording and transcripts now, so we can include things like consultations and door-to-door work. We've got some clever ways of handling those. There really is no limit to the text we can look at, and it's definitely not just TSM-based. We've got loads of other frameworks: transactional surveys, complaints (formal or informal), social media posts, Google reviews (sometimes relevant for housing associations) — wherever there's text, basically, we're interested.

Another question from one of my other favourite Chrises — Chris at Guinness. I'll give this to Chris at Housemark. "Your evidence around 'fairly satisfied' is overwhelming, but obviously we're stuck with grouping them in our official reporting. Do I have to create another internal measure to generally measure service improvements? What should we be tracking, Chris?"

Chris Elliott: A great question from a Chris. Obvious, isn't it? To a degree, we are hand-tied from the regulator in terms of official reporting. But I do like the idea of an internal reporting mechanism that focuses on "just very satisfied". It's sort of hardwired from where I used to work — we used to talk about that as a top-box score, and sometimes it appears in conversation when I'm talking about this. I really do like the idea of that as an internal KPI: track "very satisfied".

I often do that when I'm presenting results — when I'm killing clients by PowerPoint. One of the things we do focus on is the change in the proportions of people who have scored "very satisfied", and indeed all of the scores. So it's a really good idea to track very satisfied internally. Sadly, we are hand-tied at this moment in time by the regulator on official reporting. But I think it's a really good call.

Sarah Wilson: Appreciate we're over time, guys. I'm just going to run through a couple more questions, and then we've committed to answering as many as we can in the care package as well, so you won't be missed out.

Another one for you I think, Steve: "Can things be broken down to a more causal reason?" — the golden thread through the data that I know you love to talk about.

Steve Erdal: Absolutely. Some of the things we've spoken about in this presentation are the top level of that hierarchy — those big umbrella topics, almost. The way in which we break down the themes is: we've got those top-level categories around mirroring the TSMs, and then broader issues like "causing or exacerbating health issue" would be one. And then we can break that down into individual health issues.

As you say, the key thing here is individual causes. One housing association we work with described it as the Delilah principle of going "why, why, why?". Start with the top level and say, well, why is that the case? Why is that the case? Why is that the case? We can drill into that using those themes. We also do unsupervised topics — rather than us saying "is it this, this or this", we just see what the data will bear. That's stuff our customers regularly use as well.

Taking those top-level things — for example, health issue — health issue itself could be a mental health issue, a physical health issue, and so on. Breaking those down into their constituent parts is what we obviously didn't have time to do today, but that's very much what you can expect from the kinds of reports Sarah's describing.

Sarah Wilson: Perfect, thank you. Probably another one for you, Steve — want to get our money's worth out of you. "How does the sentiment data work with the satisfaction score?" This is from M Soul Me. If a customer gives "neither satisfied nor dissatisfied" but the sentiment score is very low, how do you connect them together?

Steve Erdal: A great question — it's a useful overlap of KPIs here. The sentiment score is separate to the satisfaction score. As we said before, customers are weird — you can get people who'll give low scores and then praise for specific aspects, so it's not always useful to say there's a direct connection between someone's satisfaction score and what they said in their post. So we look at the sentiment separately from that.

It's worth remembering as well, of course, that when you're in practice bringing in other forms of data, some of those forms of data won't have a satisfaction score or any score attached to them. So what we try and do with the BI tool and with the platform is to allow you to see those numbers separately, but also see how they interact. What are the things that, when people talk about them negatively, will then cause them to give a satisfaction score? But you can also look at them separately, and a lot of our best customers use them as twin KPIs — because they can add extra data into the sentiment score and get a sense of how that wider data is affecting it, rather than just focusing on things that have a score attached. There are often things that don't align exactly with those two scores.

Sarah Wilson: Perfect, thank you. And the last question for today. Tammy's asked whether we're going to answer all the questions and whether everyone can see the responses — we'll definitely put together a Q&A list with all the questions and answers, so everyone can see them.

But last one: over to you, Chris. "Can you tell us a little bit more about the differences between the small and large housing associations? And how is the data changing in terms of what's been gathered and analysed?" We had a few questions around time-based changes and how it's different from last year's set.

Chris Elliott: Really good question, and really interesting. There are two answers to it. When we think about the main differences, the first thing is: are there any differences in their needs? And the answer is no. Broadly, core tenant needs are the same. Really important to them: safety, security, good-quality homes, timely repairs, maintenance, good communication. All of those key things that I've seen for 15 years now have been consistently what's really important to tenants.

But the experience can differ based on size. What you tend to find is that smaller providers and landlords often offer more personalised services. Staff tend to know people, sometimes even know people individually. There's less bureaucracy, quicker decision-making, stronger sense of community.

Larger landlords — it's more difficult, more complex for them. They've got more specialised services, more resource, more investment capacity, and more standardised policies.

Does that mean things change? What we're finding is that with smaller landlords, tenants are expecting more personal attention, more local responsiveness; with larger landlords, they're expecting more professionalism, better digital services, better and clearer processes, and more specialist support. So while the needs don't necessarily change, the delivery model is slightly different. But fundamentally, the needs aren't really any different.

Sarah Wilson: Well, I think that concludes the session. Apologies again for running over, but I think that was so valuable. I just want to say a massive thank you, as I'm sure you do, to our contributors. Thank you so much, Steve. Thank you, Chris, for all your time — not only today, but obviously getting this organised, doing the report, putting together the presentations.

The three of us are actually going to be out in the wild next week — they've let us out. We're going to Birmingham to the Housing Innovation Show. Catch us at stand 22. Chris and Steve are going to be presenting something similar to this on the Tenant Engagement stage at 11 o'clock. So do catch us there if you're at the show.

And like I said, if you do want to take a next step — put your hand up — then go into the link, register your interest. We'll send it around in an email. I'll also have a look at the questions that have been asked in the chat and get some answers across to you as well. So just expect lots of communications from us. We always welcome feedback. Our emails are very straightforward — Sarah at wordnerds.ai, and Steve's the same. Do drop me an email if you've got anything to say — any questions, anything you hated, you can just send that straight to Steve. I really hope you enjoyed it. Thank you so much for joining us today, and hopefully we'll hear from some of you soon.