Why itsu's app customers keep coming back with Katie Grundy, CRM & Loyalty Manager at itsu

This week Chris sits down with Martin, founder of Feedelity, in London for Hospitality Tech Expo. Before starting Feedelity, Martin worked in private equity and then at Hawaiian Poké Bowl, one of the fastest-growing chains in the Benelux with 50 sites. There he saw two people trying to handle around 2,000 guest messages a week across Google, Deliveroo, Uber Eats, Just Eat, TripAdvisor, email, contact forms and surveys. They were copying and pasting complaints into ChatGPT to reply. So he built something better. Chris and Martin talk about how Feedelity pulls every feedback channel into one AI-powered inbox and categorises each message down to the dish. They cover how the data revealed that service and friendliness come up in reviews more often than the food itself. They also get into the “Wheel of Fortune” feature that gives the silent, happy majority a reason to leave a review. In this episode: How a 50-site chain was drowning in 2,000 guest messages a week Going live in two minutes, with no credit card and 30 days of feedback analysed straight away Why measuring feedback beats reading every review Real stories: two sites ignoring the music policy, and a supplier issue that made the corn too salty AI replies that learn your brand’s tone of voice The Wheel of Fortune, which gets 8 to 12 times more five-star reviews Why asking for feedback isn’t annoying your guests Why the problem is even bigger in the UK than in Belgium Martin’s one piece of advice: every piece of external feedback should trigger an internal action Chapters: 00:00 Intro 01:09 Martin’s background: private equity to Hawaiian Poké Bowl 02:17 The problem: 2,000 messages a week, two people 04:18 From late nights to quitting the day job 06:07 What Feedelity actually does 07:23 Using competitor data to price with confidence 08:01 Why service trumps food 10:04 Measuring vs reading: the music policy story 11:43 AI replies in your own tone of voice 13:45 Why simple is the hardest thing to build 15:09 The Wheel of Fortune 18:08 Why asking for feedback isn’t annoying 19:04 Replying to reviews: 23% more repeat delivery orders 19:29 Red flags: catching the serious stuff fast 21:03 Results: 4.8 on Google and the salty corn 22:39 UK vs Belgium 24:45 Martin’s one piece of advice 26:44 How to get in touch Connect with Martin: LinkedIn, or book a call via the Feedelity website. 🎧 New episodes every week from The Stacked Podcast. Subscribe so you don’t miss one. wearestacked.io

Guests don't just book a room anymore. They book the spa, the dinner, the round of golf, the whole weekend. Most hotel tech still hasn't caught up.This week Chris Fletcher is joined by Louise Ryan (CMO) and Joe Watkins (Head of Data and AI) from Journey, the Cheltenham-based hotel tech company helping luxury hotels drive total revenue per guest.They cover why hotels need to think more like retailers, why RevPAR is giving way to TRevPAR, and how one hotel partner moved from 50/50 OTA-to-direct to 80/20 direct. They also discuss where AI is creating real commercial value right now and why trust is still the biggest barrier to adoption.Chapters00:00 Intro01:30 Meet Louise and Joe02:08 What Journey does02:33 From digital agency to all-in-one booking engine04:06 The pressure on luxury hotels right now05:08 Thinking like a retailer: monetising every square inch07:35 Where guest data actually lives (and who just resigned with it)08:42 Hotels as multiple businesses under one roof09:22 From RevPAR to TRevPAR11:44 380 nights at The Other House and still not understood12:37 The Cotswolds spa booking that went wrong14:24 Taking back control from the OTAs16:05 Personalisation and the data flywheel17:26 AI chatbots and the adoption problem19:27 Convincing luxury operators to change20:30 Why experiences are still so hard to book online21:41 One basket, one transaction22:32 Where AI is showing real ROI today24:52 MCPs and booking hotel experiences through AI assistants25:30 Is trust the barrier to AI adoption?27:00 Grantley Hall: building an events programme from guest data28:08 Best-in-breed vs all-in-one29:28 Data consolidation and the future of hotel tech31:29 What's next for luxury hotels32:53 Where to find JourneyKey takeawaysSome properties pay 23–25% commission to OTAs. Driving direct saves margin and gives hotels ownership of their guest data.Around 66% of accommodation is booked online globally, compared with just over 30% of experiences. That gap is a revenue opportunity.The quickest AI win is often learning from the questions guests ask, like "can I charge my EV?", that your site never answered.Guests now search for an experience first, then a destination, then a flight, then a hotel.Find JourneyWebsite: journey.travelLouise Ryan and Joe Watkins are both on LinkedIn.The Stacked PodcastWhere hospitality meets technology. New episodes every week.wearestacked.io

Camino did something most operators would never dream of: it changed nearly every platform in its stack in a single year, while doubling its estate. The result? Staff turnover has never been lower.This week, Chris Fletcher sits down with Tyrone Delaney, Finance Director at Camino, the Basque-inspired restaurant and bar group behind Camino and Pinchito, and Jackson Versatano, Senior Director UK & Europe at Lightspeed.Tyrone explains how an acquisition became the moment to overhaul the whole stack, and why he wonders why they didn't do it sooner. Automated sales syncing into Xero removed manual weekly business sheets that used to take around 20% of the finance team's time. Card tipping added roughly 10% to tronc pots across every site, meaning better pay and people staying longer. The group also uses Tenzo to plot dishes by profitability and popularity, which turned their executive head chef into a tech enthusiast.Jackson explains why "we have an integration" isn't the same as a good integration. He covers how Lightspeed decides what to build itself and what to leave to best-in-class partners, and where AI is taking multi-site reporting. Tyrone also shares a very sensible take on using Claude: how do you mandate something when you don't yet know where its ceiling is?In this episode:00:00 Intro01:06 Meet Tyrone and Jackson01:23 Camino, Pinchito and authentic Basque cuisine02:17 A finance director who ends up owning payroll, people and tech04:25 The state of hospitality: tough, but resilient and optimistic06:54 Giving operators the freedom to choose their tools07:42 Tracking campaign ROI and menu engineering with Tenzo10:38 Integration vs good integration11:04 Why an acquisition was the moment to switch everything12:28 Fear of change, and the lowest staff turnover yet14:24 Adoption, training and the Sunday app15:51 The biggest ROI: 20% of finance time back and 10% more in the tronc17:56 AI in Lightspeed and asking questions of your data20:00 Using Claude, and how to mandate something you're still learning21:35 Operators building their own tools22:27 What's next for Camino23:09 Industry outlook and alternative funding with Shuffle24:36 Lightspeed's roadmap: Tempo, Tasks and AI25:18 Where to find Tyrone and Jackson🎧 New episodes every week. Subscribe so you don't miss one.wearestacked.io

The Stacked Podcast — where hospitality meets technology. Powered by Lightspeed.

Chris is joined by Ally Fraser, SVP and GM for EMEA at Fourth, a business that has been building back of house systems for hospitality for over 25 years. Ally spent 15 years on the frontline before joining Fourth, where she started out implementing purchase to pay and inventory at operators like YO! Sushi and now runs the region.The conversation lands on one uncomfortable number. Research Fourth ran with Peter Martin and the team at Peach found an average 11% margin gap between the top performing and bottom performing site inside the same brand. Same brand. Same systems. Same menu. Ally's argument is that the gap is behavioural, not technical, and that adoption data is the thing that closes it.They also get into the UK legislation squeeze, including operators seeing sick pay costs rise 40 to 50% off the back of the day one change, why forecasting took hospitality 20 years longer than it took the supermarkets, how automated scheduling actually works, and why nobody wants a machine placing their food order just yet.Chapters01:10 Intro and Ally's route from frontline to Fourth02:54 Fifteen years, six roles, one company03:34 What Fourth actually does, and why the demand forecast sits at the centre04:55 Why hospitality trailed retail on forecasting by two decades06:21 QSR forecasting at SKU level in 15 minute increments07:36 The UK legislation squeeze: day one sick pay, VAT, the Fair Work Agency09:09 Why compliance failure gets blamed on the tech supplier11:38 Where AI actually shows up in the product12:50 How automated scheduling works, step by step14:20 The bit nobody talks about: humans have to trust the schedule15:26 Sweat your tech, before you sign another contract17:15 What best in class operators do differently18:04 Adoption by sector: QSR strong, pubs and casual dining around half, hotels lowest20:24 Benchmarking every click, and what good actually looks like21:47 The 11% margin gap between top and bottom site in the same brand23:12 Vibe coded dashboards, and why operators should not become developers26:08 Suggested ordering, safety stock, and keeping a human in the loop28:08 Quickfire: competitive socialising, the reMarkable, and the one thing she'd fix overnightKey takeaways1. The gap is inside the brand, not between brands. An 11% average margin gap between best and worst site in the same estate means the playbook already exists somewhere in the business. It just isn't travelling.2. Adoption is measurable now. Fourth is capturing every click and benchmarking behaviours. If shift swaps at one site run at 20 a week and another runs at five, that variance can be correlated to site performance and flagged before it shows up in the P&L, not after.3. Legislation is the cost story of 2026. Day one sick pay has pushed some operators' sick pay costs up 40 to 50%. Ally's view is that operators should not have to become legislation experts. That is the supplier's job.4. Rollout of automation stalls on trust, not capability. Fourth could auto send the order. Operators don't want it sent. Same with the schedule that says a 10 to 3 rather than a 10 to 4. The tech is ahead of the confidence.5. Adoption maturity varies by sector. QSR is strongest because the model is repeatable. Pubs and casual dining are sitting around half of what they could be using. Hotels and complex operations are furthest back.6. The best operators resource adoption properly. Change managers, productivity managers, whatever the title. Someone owns it, gathers feedback and iterates.