About
I started my career writing code for data and analytics systems. After a few early stage…
Articles by Ryan
-
Farewell to Microsoft, hello to my next adventure
Farewell to Microsoft, hello to my next adventure
Last week was my final week at Microsoft. It was an incredible 10 year journey starting with the founding of VoloMetrix…
114 Comments -
8 years and 15 HBR articles
8 years and 15 HBR articles
It's officially been 8 years since I began the journey into People Analytics (which wasn't really a thing back then!)…
11 Comments -
The Paradox of Workplace Productivity
The Paradox of Workplace Productivity
Click here to go to HBR and see an article I wrote on the differences between personal productivity and organizational…
5 Comments
Activity
5K followers
-
Ryan Fuller shared thisI thought the hard part was the data. Seemor now covers more than 2 million restaurants in 71 countries. For the ones we’ve analyzed, we track 300+ things about each one, from how loud it gets to whether locals actually go there. When you look up a restaurant on Seemor, I’m pretty confident you’re getting the real story. I’m proud of that. The obvious next step was recommendations. Nobody wants to spot-check twenty restaurants. You want to ask the way you’d ask a friend and get a few great places back. Turns out that part is much harder. People misspell places. “Cambridge” exists on two continents. Sometimes the location is a neighborhood, sometimes near a park, hotel or street address. Then there’s what you mean by the restaurant itself: dinner, coffee, a bar that serves food, Northern Italian, sushi, good gluten-free options, or just “surprise me.” Then all the softer stuff: quieter, dressier, takes walk-ins, mostly locals, cozy, buzzy, good for six people, somewhere for a late bite. We have to turn all of that into something structured enough to search against our data, then rank the results by what you were most likely asking for. In a few seconds. Across 71 countries. It's one of those problems that seems like it shouldn't be that difficult until you dig in. Which is why obviously wrong answers feel so bad. The image has a few of our better failures. I originally assumed one good AI call could just figure this out. I was very wrong about that. We went from one call, to two, and as of last week five separate AI calls work together to interpret the request before we look anything up. Testing says it’s much better now. That said, users are extremely good at finding things I haven't tried yet, so we'll see... As always, feedback is very welcome.
-
Ryan Fuller shared thisI had the absolute pleasure of participating in Customer Confidential with Rob Markey last week. For anyone who has worked around customer feedback, loyalty, or NPS, Rob needs very little introduction. So it was especially fun to spend an hour with him talking through a question I spend a lot of time thinking about: What incentives are hiding behind the information we get as consumers, and how should we decide what to trust? Part of the inspiration came from two London Airbnbs my family stayed in after moving to Europe. Both were rated 4.9. Both were terrible. After we left honest reviews, both hosts separately offered us £200 to remove them. That sent me down a rabbit hole on reviews, and eventually helped inspire Seemor. In one analysis of 4,000 restaurants, 70% showed at least one unusual pattern - review bursts, repeated phrasing, thin reviewer histories. That does not mean 70% were fake. You can never be completely sure. What I’ve become interested in isn’t just whether a particular review is “fake.” It’s understanding the system around it: who benefits, what behavior gets rewarded, what signals are missing, and how much confidence we should actually have in the result. The image below shows five real restaurants rated 4.5+ on Google. Seemor sees a lot more uncertainty underneath those numbers. One gets a B. Another gets a C? (The question mark is deliberate). Rob and I got into this, the incentives behind customer feedback systems, and what trust looks like in an AI world. Really enjoyed the conversation. Link in the comments.
-
Ryan Fuller reposted thisRyan Fuller reposted thisFake reviews are fraud. Deliberate attempts to mislead in order to make more money. They actually enrage me. They mislead buyers. Worse, they punish the good companies that don't cheat. Ryan Fuller, serial entrepreneur and founder of Larrocos Labs, really pushed my thinking about fraudulent reviews. He stayed in two London Airbnbs. Each had a 4.9-star rating. Both were terrible. He left negative reviews, detailing how bad the experiences had been. And then, each host reached out, offering him £200 to change his review to five stars. Obviously, that's misleading to future guests considering staying in those properties. It reveals, however, a more systemic issue: If enough reviews of a hotel, rental property, restaurant, product, or company are fake, manipulated, or otherwise untrustworthy, then the companies that invest and work hard to deliver genuinely better experiences suffer huge penalties, too. That problem multiplies with AI. The cheaper it becomes to produce plausible reviews, the more fakes get generated. And every successful detection method gives manipulators a new target to work around. The frauds are in a race with the platforms, trying to outsmart each attempt to police the reviews. His Airbnb experience led Ryan to examine millions of restaurant reviews. The work he did made me wonder whether the answer is less about declaring reviews true or false. Maybe its more about showing the evidence, source credibility, and degree of confidence behind the ratings. Regardless, this whole situation raises a hard question for the companies that actually deliver better experiences. If reviews are becoming less and less trustworthy, how can you give potential customers confidence in the quality of your service? I wish I had a better answer. What would you do? Listen to my conversation with Ryan on the latest episode of Customer Confidential. Link in comments.
-
Ryan Fuller shared thisThousands of people use Seemor every month to figure out where to eat. That's super cool. Unexpectedly, the biggest consumers (by far) of Seemor are AI systems. I built a restaurant app for humans. I didn't anticipate that machines would be my biggest users. The major AI assistants started finding Seemor on their own. No outreach, no deals. ChatGPT, Claude, Perplexity, and a handful of others just started pulling from it. A lot. I'm talking hundreds of thousands of hits per month. And I can see the loop closing: people now land on Seemor directly from recommendations in ChatGPT and Gemini. When you ask an AI chatbot where to eat, there's a decent chance the answer already ran through our data. It makes sense. We've now analyzed more than 250,000 restaurants across 60+ countries, with hundreds of data points on each one. As far as I can tell, it's become the richest structured dataset about restaurant anywhere. Like us humans, these AI systems need structured judgment about restaurants, not just more information. Is the place actually good? Who's it for? What's it like inside? Is it authentic or a tourist trap? A 4.5 star rating doesn't tell them (or us) any of that. That's what we've spent the last year building, one restaurant at a time. Turns out the machines have exceptional taste in data :-)
-
Ryan Fuller shared thisI set out to answer what I thought would be a simple question: Which countries have the most restaurants per person? I counted every restaurant, bar and cafe across 22 countries using a single methodology. The results looked straightforward. South Korea had over 4x as many venues per person as the UK. Portugal nearly 3x. Greece beat Italy. The obvious explanation was that people in those countries simply eat out more. Nope, not that simple. Greece's density turned out to be largely a tourism artifact. Mykonos has 24 restaurants per 1,000 people. Thessaloniki, where Greeks actually live, has 1.5. I thought I'd found the answer. Then Spain broke the model. Barcelona and Valladolid have nearly identical restaurant density despite a 6x difference in tourism. Whatever is driving Spain's density, it isn't visitors. Affordability explained Japan and seemed promising, until Germany broke that explanation. Then Las Vegas broke things in a completely different way. I came away wondering whether restaurant density is simply an outcome of local culture, or whether it also helps shape it. It might be both. I wrote up the full investigation with charts for each finding: https://lnkd.in/eSiMSAzdI Counted Every Restaurant in 22 Countries. It Raised More Questions Than Answers.I Counted Every Restaurant in 22 Countries. It Raised More Questions Than Answers.
-
Ryan Fuller shared thisChris Brahm and I started VoloMetrix together 15 years ago. Now he's writing about the problem I'm living every day in 2026. Within weeks of launching Seemor.ai, every major AI platform found our restaurant analysis data and started consuming it at high volume. No one asked. No one paid. I'm a solo founder and LLM bots now hit the site millions of times per month. Chris's new piece lays out the framework for protecting proprietary knowledge in the AI era. Seemor is one of his examples. The distinction he draws between raw inputs (e.g., menus, reviews) and the analytical model built on top of them is exactly right. That model is the thing worth protecting. The interesting question I'm focused on now is whether any of these AI companies would ever be willing to pay for it. TBD on that. Worth reading if you're building anything with proprietary data.Ryan Fuller shared thisAgents done right define, deepen, and automate your competitive advantage. But without the right technical and business architecture in place, the very agent you're building to strengthen that advantage could give it away. CEO and former lead of agentic systems at AWS Aaron Rallo and I write about where AI might be undermining your business and how to defend yourself. Timely given recent news that shared Claude Chats and Artifacts can actually show up in public searches. Read about it here.Is your AI Agent Really A Trojan Horse? Three steps to protect your proprietary knowledgeIs your AI Agent Really A Trojan Horse? Three steps to protect your proprietary knowledge
-
Ryan Fuller shared thisWorld Cup final today. Argentina vs Spain. Jen and I lived in Buenos Aires in 2010 during the World Cup. We weren't football fans. That changed fast. Every time Argentina played, the entire city stopped. Buses stopped. Banks closed. Everyone crowded into the nearest bar and watched together. Even if the game was at 6am. The passion was unlike anything we've experienced before or since. When they lost in the quarterfinals, we were briefly worried. But the city just picked itself back up and kept going. We left as Argentina fans for life. Thinking about Argentina vs Spain today made me wonder what these two countries actually have in common beyond language. And since I can't stop posting about restaurant data lately, I figured I'd look at this through the lens of food. The answer: almost nothing. Argentina pretty much rejected its Spanish heritage when it came to food. Seemor looked at 7,600 restaurants in Buenos Aires. Over 1,400 serve pizza or Italian food. 42 serve Spanish food. Forty-two. Between 1880 and 1950, five million Italians moved to Argentina. They brought the food, the hand gestures, and most of the accent. Argentine Spanish sounds pretty Italian-ish. Today half the country has Italian ancestry. Several of Argentina's current World Cup players play for Spanish clubs, but they all grew up eating pasta (and, of course, lots and lots of steak). Vamos Argentina. p.s. we look exactly the same as we did 16 years ago, RIGHT?
-
Ryan Fuller shared thisWe put the internet's favorite Chinese restaurant hack to the test. A few years ago, Freddie Wong posted a TikTok claiming the best way to find authentic Chinese food is to look for 3.5 stars on Yelp. His theory: higher-rated Chinese restaurants often earn those stars by focusing on service and atmosphere, while the food can actually be better at lower-rated places. Seventeen million views later, it's become one of the internet's favorite food hacks. So we analyzed 2,900 Chinese restaurants across 51 cities to see if he was right. The answer: half right. We found plenty of good food hiding behind mediocre ratings. But the food never actually gets better as ratings fall. There's no 3.5-star sweet spot. What Freddie got right is that food quality falls much more slowly than the rest of the experience. When Google ratings drop 29%, food quality only drops 12%. Whereas service, ambience, and food safety all drop much more. That's why the hack feels like it works. You can absolutely find good food behind a bad rating. HOWEVER... nearly 40% of sub-3.5-star Chinese restaurants earn a D or F in our system. The hack finds some good food, but it also finds a lot of very bad restaurants. The fix is simple. Spend two minutes reading reviews for food safety concerns. Once you filter out the hygiene risks, the food quality gap between a 3.8 and a 4.6 star drops to almost nothing. So, Freddie's hack kind of works, but you might get sick :( P.S. Or just use Seemor.
-
Ryan Fuller shared thisSomeone commented on my last post that she's always disappointed when she tries Italian food outside of Italy. I can certainly relate to this. Italian food is the most common restaurant cuisine in the Western world, but it's often adapted in one way or another. Seemor is able to measure cuisine authenticity based on review patterns and menu analysis. e.g., are reviewers describing traditional techniques, family ownership, regional ingredients? Or a generic pasta menu? We scored the authenticity of over 7,000 Italian restaurants across 20+ countries. The ones that are genuinely authentic average 8.1 on food quality. The adapted versions? 7.1. Almost a full point lower. At least for Italian, higher authenticity = more enthusiastic reviews of the food. Authentic Italian food is out there. It's just not evenly distributed. Only about 28% of Italian restaurants outside Italy score high on authenticity. But the odds vary wildly depending on where you are. In London, 56% of the Italian restaurants we analyzed are authentically Italian. In New York, 51%. San Francisco, 46%. In Berlin? 7%. LA? 5%. For reference, even Rome is only 55%. Tourist traps are a problem everywhere. Where have you found the real thing? And how do you find it?
Ryan Fuller replied to a comment
4w
Jason Potteiger 😀
-
Ryan Fuller liked thisRyan Fuller liked thisNice try Meta. It’s not that Muse is not a well conceived product. It’s not that the Muse charm isn’t an interesting exploration. You will not get access to my life. You worked hard on losing that right. Im sure this will be adopted widely by folks who want the “free”. Not me. #ai #personalassistant
-
Ryan Fuller liked thisRyan Fuller liked thisAn adventure and honour to visit Riyadh for my first time for their 'small' tech conference LEAP - just 225,000 attendees! A unique 36hrs to see outside of the western view of the entrepreneurial world! Wonderful to meet an incredible group of people in such a short time: H.E Abdullah Alswaha - Minister of Technology for Saudi Arabia and director of HUMAIN the world’s leading arabic-language LLM Chandra Donelson - exChief AI Officer of US Space Force and children’s author! Zain Samdani - a local Saudi founder who's co-founders include his mother and sister who is building robotics to support paralysis recovery. Imane Abdel Jelil - female founder from Mauritania who is bringing AI to Mauritania and who’s uncle is an Oscar-winner! Fahad S.r-winner! Fahad S. - who has co-founded a brilliant VC 3Ventures Group in the region with Insight Partners (HelloFresh's great backer) as their partner. Nora M. Alsarhan - the female CEO of SVC, Saudi Venture Capital, who are the key catalyst of venture growth in the country. Ambassador Stephen Hitchen - His Majesty's Ambassador to Saudi Arabia who kindly gathered all Brits attending LEAP and who was cheering to help British businesses thrive in the region. Ahmed Al Rawi - Bahrain’s most successful founder who ready-prepped meal business now does >$100m revenue. Janna Salokangas - female Finnish AI-founder who lead Finland's presence at DAVOS and is expanding MIA.ai the region Barkin Ozdemir - founder and the youngest Turk to ever visit all the continents and explore Antarctica Brad W. - a wonderful welshman who does the knockout countdowns for the Riyadh huge boxing matches! and so many epic founders from the region. Founders from Saudi, Pakistan, Turkey, UAE and more. All wonderful impactful energised AMBITIOUS people! And completely energising and inspiring for me too! Good luck to them all. GO GO GO!
-
Ryan Fuller liked thisRyan Fuller liked thisIs mass unemployment really the inevitable outcome of the AI revolution? Aaron Rallo and I believe a different story. We believe that job loss won’t be as dire as it’s been made out to be. We also believe that your experts will be the real winners and business leaders should treat them as a valuable, irreplaceable resource. How technological disruption transforms the job market is impossible to predict accurately. The only thing that is certain: As a CEO, you need to have a narrative your employees believe in. Don’t let your people stew in the soup of uncertainty. Paint a “point of arrival” that is inspirational for the company and your people: those that embrace AI can learn leading edge technology that is changing the world and can become massively more impactful. Read the story we think business leaders should be telling.In Your Business Story, The AI-Enabled Experts Will Be The HeroesIn Your Business Story, The AI-Enabled Experts Will Be The Heroes
-
Ryan Fuller reacted on thisRyan Fuller reacted on thisEmbracing Micro-Seasons: Just sent off 2 kids - one to Burning Man and beyond for her post-grad adventuring, and the other to his sophomore year in college. They were home for a little over a week of joyful chaos, and now I’m looking at 4 weeks of relative calm - no travel, weekly meetings, time to write. After that, I re-enter a more intense Board season. To navigate these shifts, I am now employing a time management strategy I call “Micro-Seasons," borrowing from the poetic Japanese tradition of dividing the year into 72 nature-driven seasons. For each Micro-Season, I identify my top 3-5 objectives and design accordingly. For example, exercise is a persistent priority across all my Micro- Seasons, but how I work out differs greatly when I can maintain a steady routine at home with my local gym and yoga studio vs. when I am on the road or prioritizing being present for my adult kids. I shared this strategy with an executive I advise, and she put it into action immediately. She identified that her next two weeks of travel put both a health and a professional priority on hold, and she enlisted her EA to build time into her schedule for them. I honestly didn’t think there was a time management strategy I hadn’t tried yet, but I found one, and it works! And the best part is that I no longer need to feel guilty when a routine that worked starts to fail. I now recognize that as a function of moving from one Micro-Season to the next, the natural order of things. Renée Cullinan Ashley Fieglein Johnson Casey Boyles Jean Fuller Ryan Fuller Chantrelle Nielsen
-
Ryan Fuller liked thisRyan Fuller liked thisI’ve been thinking a lot about *ambient learning* the learning that is pushed onto you by the people you work with, the people you meet, and the environments you’re in. This has been some of the most important learning that I've done in my career, things I didn't seek out but saw or was around and picked up. Things like: - How to navigate a difficult conversation - A standard I didn’t know existed before. - How to give and receive feedback graciously (still a work in progress!) - How someone makes a decision without having all the information - When to push on something and when to let it go - What excellent work actually looks like AI has made it incredibly easy to learn the things we know we want to learn. But it’s also made me appreciate this other kind of learning even more. You can’t ask a question about something you don’t yet know exists. Sometimes you need to be in an environment that teaches it to you.
Experience & Education
Publications
-
People Analytics Will Change the Way You Manage Your Business
American Management Association
-
Leveraging Big Data To Redefine Collaboration In The Enterprise
SAP Business Innovation blog
Patents
Honors & Awards
-
Gartner 2015 Cool Vendor
Gartner
-
40 under 40
Puget Sound Business Journal
-
Big Data 50
Startup50
VoloMetrix selected as one of 50 hottest big data startups of 2014
-
Top Ten Innovator
Software & Information Industry Association (SIAA)
VoloMetrix selected as one of the 2013 “Top Ten” most innovative companies transforming the software and services industry.
Recommendations received
4 people have recommended Ryan
Join now to viewOther similar profiles
Explore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More