How do you turn hotel reviews into board-ready evidence?

What hurts a hotel's star rating the most?

Across nearly 400,000 reviews, the single biggest driver of a falling star rating is condescending staff — guests being talked down to or made to feel small. When it happens, reviews drop by around 2.5 stars, and it holds at every tier from budget to luxury. That drop is bigger than for staff who are merely rude, unhelpful, or absent. The language guests use shifts by tier: budget-hotel guests are more likely to call staff "arrogant", mid-range guests "dismissive", and luxury guests "condescending" — but the underlying experience, and its impact, is the same. The practical takeaway for training and recruitment is that tone matters more than almost anything else: a guest who feels talked down to will punish the rating harder than one who simply couldn't find a staff member. It's the highest-leverage area to fix in frontline customer experience.

How much is a hotel review actually worth in revenue?

A lot more than most CX teams can currently prove. Science Direct research links a one-star increase on TripAdvisor to roughly 3% growth in revenue. Extrapolated to business scale, the numbers get large fast: in a global hotel market worth around £1.1 trillion, for a mid-level chain like Accor turning over roughly £250 million a year, the difference between a three-star and a five-star reputation is worth about £15 million. That is the case for treating star ratings as a financial metric, not a vanity score — and the reason Wordnerds frames review analysis around a key metric the senior leadership team already cares about (often the one their bonus depends on). Once you can attach a nominal revenue value to each point of improvement, customer-experience work stops looking like a cost centre and starts reading as revenue protection and growth.

What's the biggest difference between budget and mid-range hotels?

Dining. Across the seven-stage guest journey Wordnerds maps — booking and pre-arrival, arrival and check-in, rooms, staff, amenities, bar and dining, and post-stay — budget and mid-range hotels score remarkably similarly on six of the seven stages, and budget hotels sometimes edge ahead. The one consistent, significant gap is bar, dining and room service, where mid-range hotels pull clearly ahead of budget ones. It comes down to expectation: budget-hotel guests often expect more from the food than they get, and breakfast is a particular pain point, where missing items or lack of choice produce a sharp drop in sentiment. For mid-range chains, dining is therefore an area of genuine competitive strength to protect and leverage; for budget chains, closing the breakfast-and-food-quality expectation gap is one of the clearer routes to a better rating.

How can budget hotels increase their star ratings?

By creating memorable experiences — the lever that lifts budget-hotel star ratings by about 1.3 stars when a guest says a stay was genuinely memorable. Budget hotels have the smallest resources for guest satisfaction, so they have to spend carefully, and the data shows where memorable experiences actually come from for them. In around a fifth of cases it's first impressions: how guests are greeted, what check-in feels like. Budget chains are also winning by focusing on families travelling with children — small, inexpensive keepsakes around holidays like Christmas and Easter — which generates a noticeable rise in guests describing the stay as meaningful or memorable, which in turn lifts star rating, which in turn lifts revenue. The lesson is that emotional impact, not capital spend on décor or price cuts, is the most cost-effective way for a budget hotel to move its rating.

How can mid-range hotels improve value-for-money perception?

By adding staff at breakfast — worth roughly a 0.8-star lift. Mid-range hotels carry a specific bind: guest expectations push towards luxury while budgets sit closer to the budget tier, and it shows most in value-for-money and price sentiment, which is lower for mid-range than for either budget or luxury hotels. Of the obvious fixes, more breakfast staff beats investing in comfier beds or extra room-cleaning attention on the value-for-money measure. Two reasons show up in the data: breakfast is a moment when guests need help quickly and often don't know where to go, so visible staff are genuinely appreciated; and the optics matter — a sparsely staffed breakfast room signals that little is being spent on the guest. Concentrating staffing on those couple of breakfast hours is a low-cost, high-impact move for the tier that struggles most to feel worth the money.

How do you present customer-experience insight so leaders act on it?

Make them guess first. Alongside four analysis techniques — measuring impact on a key metric the leadership already cares about, mapping the customer journey, diving into a key issue until you reach a concrete action, and benchmarking against competitors — Wordnerds' best customers use a fifth, presentation technique: before revealing a finding to the senior leadership team, ask them what they expect the answer to be. It's the pub-quiz format this webinar itself used. Making leaders commit to a prediction surfaces their own biases and lets them be genuinely surprised, which is far more powerful than telling them a fact they can later claim they'd have guessed. The point of all five techniques is the same: get past top-level reporting to a prioritised, defensible action — and present it so the people who control the budget are ready to act on it.

Full Webinar Transcript

Helen Precious: Good morning. Hopefully you can hear us, you can see us, you've got your coffees ready? Welcome to our Wordnerds webinar, where we're talking about turning 390,000 reviews into board-ready evidence. We've written a report and we'll be talking you through the insights from it today. Let us know where you're dialling in from, what the weather's like, what coffee or tea you've got.

I'm delighted to introduce Steve, who you may recognise from some of our other webinars. Steve is our Insight and Innovation Manager and also Chief Linguist here at Wordnerds — Chief Nerd. Super excited. Today's an exciting day.

Steve Erdal: I'm good, thank you, Ellen. It's an exciting day, as you say. I don't like to talk about it, but I'm back on telly tonight on Only Connect, the BBC Two show. This is only the second most nerve-wracking thing I have to do today, so this is a good dry run for later this evening.

Helen Precious: Thank you all for joining us this morning. Let me introduce myself: my name is Helen, I'm Head of Account Management here at Wordnerds, and as I mentioned, Steve Erdal is our Chief Scientific Officer.

Today we'll be talking about the reviews we've analysed, but in the wider context of what we're seeing in the CX space — everybody's dealing with more data, and with that comes a lot of challenges. We'll talk about the most dangerous guest complaint in the context of what it can cost you as a business. We've analysed the data through different market segments — budget, mid-market and luxury hotel chains — so we'll look at the specific challenges affecting your segment and some of the quickest wins. All of this is underpinned by our methodology for automating the analysis of reviews, and we'll share the four techniques we used to prioritise that insight for action. And ultimately, possibly the most important bit, the message on how you as a CX or voice-of-customer team can articulate ROI to the finance people in your organisation.

To give you some background on what we're seeing: more data coming into your teams, and you're being asked to do a lot more with it for a lot less. At the last insight event we went to, most people had their teams reduced to maybe two or three people — much smaller than three or four years ago. A lot of people are still in that manual-coding space, busy working through spreadsheets, and all power to you, because it's an incredibly hard job. There's also a lot more pressure to adopt AI — using LLMs, Copilot, ChatGPT — to analyse feedback. And a lot of you are already using a VoC tool like Qualtrics or Medallia, which are fantastic for scale, scheduling and getting data at various points in the journey, but maybe not so good at giving you a level of insight that's actionable.

All this data overwhelm leads to time-consuming reporting that's six weeks out of date by the time it reaches the board, insight that's top-level and unprioritised, and teams that don't feel confident that what they're recommending is being actioned or moving the needle on their KPIs. In the worst case, when budgets are cut, there isn't enough space for the CX spend you need — and that becomes a vicious cycle. The best CX teams break that cycle by linking their feedback to ROI and metrics, proving to the business that they're not a cost centre but a strategic asset.

And every interaction matters more than you think. Science Direct research shows a one-star increase on TripAdvisor can equate to 3% growth in revenue — which is huge once you extrapolate it. In a £1.1 trillion global market, for a mid-level chain like Accor turning over £250 million a year, the difference between a three-star and a five-star review is about £15 million. So if you can increase your average star rating on TripAdvisor or Google, there's a financial implication there in black and white.

We're dealing with review data today, but more broadly, voice of customer covers all sorts of conversations — your complaints, surveys, phone calls, staff conversations. You don't need to create new surveys; customers are already telling you where they are in their journey and what they want. Wordnerds is a customer-feedback analytics platform — a specialist text-analytics platform — that helps you automate customer feedback from lots of different channels, so your teams feel confident in the actions they're recommending, look like a pro in front of colleagues, and can articulate the ROI of your CX activity. We'll talk through the analysis of the 400,000 reviews — luxury, mid-range and budget, sourced from publicly available sources like TripAdvisor, Google and booking.com — and there'll be a Q&A at the end, so drop your questions in the chat. Come on in, Steve.

Steve Erdal: Thank you so much, and thank you all for being here. As Helen said, my name's Steve, I'm a recovering linguist and one of the co-founders at Wordnerds. My particular interest is how you get the most possible value out of the insights we provide — and there's a lot of value here, a lot of potential revenue tied up in your star rating.

We've looked at just under 400,000 reviews from a range of hotels, from budget all the way up to luxury. A big shout-out to my colleagues Stella and Ruth, who did a lot of the work I'm now about to take credit for. As Helen said, I'm on a TV quiz show tonight, and I basically see the world as a series of quizzes — so, unfortunately for you, we're going to present these insights as a little pub quiz. If you're feeling brave, write your answer in the chat, and at the end we'll crown a winner. There are five key areas, each looking at a different aspect of the customer experience at a different hotel type.

First question, and we're starting with the big one. A drop in star rating can seriously impact your revenue. Which of these causes the biggest drop in star rating: confusion at the booking process, cleanliness of the rooms, or condescending staff?

[Audience votes in the chat — a mix, with most going for C.]

The biggest drop in star ratings, of all the aspects we looked at, was condescending staff — and that's true at every level from budget to luxury. As a linguist, I found it interesting that the language differs by hotel type: budget customers were more likely to describe staff as "arrogant", mid-range as "dismissive", and luxury as "condescending". But whenever a staff member makes a customer feel small or talks down to them, you're likely to see a drop of around 2.5 stars. That's bigger than if the staff member is rude, bigger than if they're unhelpful, bigger than if they're not there at all. So when you're building training to improve customer experience, that area matters more than anything else.

Next: we put the customer experience into a journey — a standard seven-point journey from booking and pre-arrival through arrival and check-in, rooms, staff, amenities, bar and dining, and post-stay and retention. For six of those seven areas, mid-range and budget hotels have very similar satisfaction. There's one touchpoint where mid-range considerably outperforms budget. Which one?

[Audience votes — staff, rooms, amenities, bar/dining.]

The biggest difference is dining. Sentiment is on the Y axis, the bubble size is how many people talk about each stage. The budget and mid-range lines intertwine for the most part — budget sometimes scores higher — but there's one considerable gap, around bar, dining and room service. It's about expectation: budget hotels have a higher expectation of food quality than they often get, and we see a lack of choice, particularly at breakfast, where missing items cause a big drop-off. That's much more likely in a budget hotel than a mid-range one.

Now, budget hotels specifically: they have smaller resources for customer satisfaction, so they have to attribute them carefully. Which of these increases star ratings the most for budget hotels — update the décor, lower prices, or create memorable experiences?

[Audience votes — mostly 3.]

Very well done if you said creating memorable experiences. That aligns to about a 1.3-star increase when a customer says they had a really memorable experience. This is true of all hotels, but for luxury the very act of going is memorable; on a budget, how do you provide that? We looked at where budget-hotel customers said things like "this will stay with me for a long time", and what overlaps with it. In about a fifth of cases the memorable experience revolves around first impressions — how they were greeted, what check-in was like. We're also seeing budget hotels focus on families travelling with children, because it can be relatively inexpensive to create those memories — little keepsakes for children, particularly around Christmas and Easter. That leads to more people saying the experience meant a lot, which increases star rating, which increases revenue.

Fourth question, mid-range. Mid-range hotels have a big challenge: expectations can be pushed up towards luxury, but budgets sit closer to the budget end. Nowhere is that clearer than value for money and price, which are lower for mid-range than for either luxury or budget hotels. So if you ran a mid-range chain and wanted to improve value-for-money perception, would you put on extra staff at breakfast, invest in comfier beds and pillows, or employ more staff for room-cleanliness attention to detail?

[Audience votes — a real mixture.]

The one with the biggest impact on positivity around value for money is more staff at breakfast — a bump of about 0.8 of a star. A couple of reasons: breakfast is a point where customers often don't know where to go or need things quickly, so more staff is appreciated; and the optics matter — a breakfast room without much staff feels like not a lot's being spent. So zooming in on those couple of breakfast hours and ensuring sufficient team members makes a real difference to that value-for-money perception.

Final question — luxury. We're zooming into two specific hotels, two big beasts of luxury in this country: the Ritz and the Savoy. Which performs best specifically for menu variety — the amount of choice on the menu?

[Audience votes — split between the two.]

The hotel that performs best for menu variety is the Savoy. We normalise our score to 0–100, so the Savoy's 96 is pretty amazing. That's the sort of granular thing you can do when the data is categorised this way. The other point: if you only looked at the Ritz's data, a sentiment score of 87 looks really good — but in the context of luxury hotels, 87 is middle of the pack. So having the ability to scan the market and understand the context behind a score matters; this feels like it should be a strength of the Ritz, but against other luxury hotels it's an area to improve.

Before I hand back to Helen, let me quickly go through the techniques we used. First, impact on a key metric — we took star rating and looked at what aspects of the experience affect it most. A bit of advice from one of our best customers: find out what your senior leadership team get their bonus on, and make that your key metric. Second, the user journey — map all the points from first contact to post-care and show how you're doing at each. Third, diving into key issues — in this case value for money — until you hit an action you can take and measure; until you reach action, your job isn't done. Fourth, competitor benchmarks — using your industry as a control group to find your areas of real distinction. And the secret fifth technique was the quiz you all just did: our best customers, when presenting to leadership, ask them to predict the answer first. It surfaces their biases and lets them be surprised, which is far more powerful than telling them a fact they'll claim they'd have guessed. Thank you for playing — I'll hand back to Helen.

Helen Precious: Thank you, Steve. Super helpful, and great insights. We'll make the report available — the handout should be in the chat now, and we'll follow up by email. We're about to come to a Q&A, so pop your questions in the chat.

Just to open the bonnet on our methodology and how we bridge the gap between what you need as an insight team and what AI is currently delivering: Wordnerds is not using Copilot or ChatGPT to do this analysis. We have our own proprietary AI software, and we believe it does a better job of giving you actionable, prioritised insight. First we classify the data — it can come in from any channel, and we sit under your frontline tools. Classifying by NPS promoters, detractors and passives often doesn't resonate with what colleagues do day to day; classification works best when you convert data into customer-journey points or departments that make sense for your business. Second, we prioritise — knowing someone dislikes the artwork isn't as significant as other issues, and we surface both what you do well and what you don't. The most important part is understanding how issues impact the scores you make, so you and your colleagues focus on the highest-value work. And increasingly, since many of you report in Power BI or Tableau, your customer feedback should live there too, alongside the quantitative data — so we push this into Power BI.

The two things I wanted to cover: moving from being seen as a cost centre to a strategic asset starts with understanding your baseline and building a cycle of continuous improvement — you can't just run NPS or CSAT without acting on it. The level above that, once feedback is linked to your key metrics, is protecting revenue: keeping review scores where they are and building on them, predicting churn so customers stay longer, and using VoC data to influence behaviour — encouraging upgrades, reorders, and use of dining and facilities. If you're wondering what your biggest revenue killers are, we're offering everybody our proof-of-concept programme: low-risk, relatively low-cost, we take the last year's existing feedback — reviews, surveys, social, anything you can get your hands on — and do the driver-satisfaction analysis we've just done in this report, putting quantifiable revenue numbers on the impact. Just book a call to chat about what it looks like for you. Now, to the Q&A.

We've got a question from Alina: how can your platform break down condescending staff per function — receptionist versus cleaning staff versus catering?

Steve Erdal: Great question. First, we can only give you information if your customer has said it — some people talk broadly about "a member of staff", others mention a specific role, others name the person, particularly for something positive. And we're not just here to find issues with staff; it's great to celebrate best practice and spread it around the organisation. But where the role is mentioned, that's one of the things we look into in depth — overlapping those points in the customer journey with the job role, or the individual, so you get a good sense of the particular challenges or strengths different teams bring. You can use it for improvement, and to celebrate best practice.

Helen Precious: The next question is from Clara: going back to mid-range and the dining offer, do you have any advice for where mid-range chains could improve to better compete with budget chains?

Steve Erdal: Dining is the area where mid-range chains have a significant advantage over budget chains, so it's about leveraging that. Food quality is normally a strength; often it's about getting the service up to that level. But it's very dependent on the specific hotel — there's a great deal of variation in strengths and weaknesses even between similar hotels. So the first thing we'd do is understand the key areas for that specific hotel, then contextualise it against similar hotels, often as a SWOT analysis — your strengths are what you do better than competitors, your threats are what they do well. It's tailored to the organisation, but on a general level, for mid-range hotels it's the service area they can bring up to the food quality, rather than the other way around.

Helen Precious: We've got one from Patricia: you mentioned linking revenue to star ratings — do you have any examples of how you might do this?

Helen Precious: A key question. This is a newer area we're seeing CX teams think about. There's a lot of research now linking revenue to star ratings — and the same for NPS and other CX metrics. We're seeing organisations do this practically in two ways. The first is getting agreement across your exec level on a nominal value per point of improvement — that works well with NPS, and where it's harder to get your finance teams to correlate the data directly. The second is to literally correlate revenue with star rating and look at what's happening with those two data points together. There's a lot that goes into star ratings, but if you work this as a continuous-improvement programme rather than just a metric no one acts on, these two things correlate quite closely. Steve, anything to add?

Steve Erdal: You articulated that really well. The only thing I'd add is, once you've taken an action, measure it. Say "we made the change at that date", make a hypothesis — "I expect conversations around breakfast to improve by this much" — and then when you see that improvement, you've got a wonderful data narrative: this is what we expected, this is what happened, this is the impact on our star rating, and we'd expect an increase in revenue off the back of it.

Helen Precious: Next is from Ayla: I'm just setting up our CX programme. What advice would you give a business that's quite far behind in adopting AI and CX tools?

Steve Erdal: Great question — and first, don't sweat it. It's easy to imagine all your competitors have robot butlers and aren't doing any work themselves. In our experience that's not what's going on; everyone's at a different point, ourselves included — we change how we do things all the time based on new advances. The key things: first, have a problem you need to solve rather than an AI you want to use — sit down with your team and figure out the key challenges, then think about how AI can help. Second, be mindful of your customers' data; this is something they've entrusted you with, and I know how seriously travel-hospitality takes that. And third, start small — what's the smallest amount of data and the smallest number of people I need to make an impact I can demonstrate, which then lets me do a bigger thing? The mistake we see is people trying to get every duck in a row before starting. Start small, show results, get more people on board, and it becomes a flywheel.

Helen Precious: Thank you all for your questions and your time. The report is available for download and we'll follow up by email. If you'd like to chat about your CX programme, give us a shout. Hopefully we'll see you on our next webinar in the series — we'll be doing this monthly, picking off some of the big VoC and CX challenges in this space. Thank you very much.