Hungryroot

Each year, we recognize our “Hungryroot Heroes” — team members who embody our values of being positive, proactive, and transparent, and who drive outsized impact for our customers and business. I’m proud to announce this year’s honorees: Erin Hoodlebrink (Inbound Supervisor), Chris Wittrock (Senior Planning Manager), Sonam Shah (Principal Product Manager), Joshua "jag" Ginsberg (Staff Software Engineer), Lindsey Baker (Photo and Video Manager), James Chen (Senior Backend Software Engineer), Vanessa Quandt (Grocery Category Lead), Jared Heller (Senior Manager, Fulfillment Systems and Analytics), Laura Cleary (Food Safety Manager), and Connie Ho (Senior FP&A Manager). Please join me in celebrating this year’s Hungryroot Heroes. Thank you for all you do!! I’m grateful to work with you

5 years ago, we pivoted Hungryroot from a struggling packaged food company into the personalized grocery service we are today. In many ways, we are still in the midst of that pivot. While revenue has grown 20x since then and we’re now generating meaningful annual profit, we’re still evolving the business at a foundational level. The difference is that each change we make now reaches hundreds of thousands of Americans. This simultaneously speeds up learning and slows down progress (as complexity is higher). It’s maintaining this nimble, always innovating mindset that I know is critical to our success, particularly in an ever changing AI-driven future. I look forward to the next 5 years!

3 takeaways from this week’s Goldman Sachs and last week’s Citi conferences: The AI “pyramid” is inverting. You can think of value creation in AI as shaped like a pyramid. Most of the value has gone into the bottom, infrastructure layer (chips, data centers) and middle, model layer (LLMs). Much less has gone to the top, application layer. On the “AI Powered Platforms” panel at the Citi conference, Ming S. Zhao Sirisha Kadamalakalva and I spoke about how this pyramid is likely to invert in the near future (as it did with the internet and cloud). Many people — credit image to Altimeter and Apoorv A. — now believe the most exciting opportunities of the next several years will be in the application layer with companies such as Harvey, Runway, and Hungryroot. Especially in the application layer, creating value from AI requires a systemic approach that combines model design with analytics, digital experience, operations, and brand. For example, at Hungryroot, the way our models are designed impacts the inventory we sell, and vice versa. Similarly, the way our UX is designed impacts the data we collect that feeds our models. Value is derived by integrating these various functions to optimize customer satisfaction and minimize operational costs and risk. On “The Role of AI in Transforming the Consumer Experience” panel at the Goldman Sachs conference, Wes Schroll and I spoke about how much of the focus to date in AI applications has been on cost cutting and enterprise solutions (e.g. streamlining customer service or back-office tasks). This is starting to change, and much of the future opportunity will be in consumer value creation (e.g. helping people eat healthier, be more creative, etc.). AI is moving so quickly that it can feel dizzying, and yet at the same time energizing. What changes are you observing or most excited about? Let’s discuss! Thank you Elinor L. Hoover and the Citi team, and Kay Lee and the GS team for including us.

I've learned a hard truth over the years: not all teams are created equal. Some teams are quick to shoot down new ideas, while others nod along without offering real feedback. But the truly great teams are the ones that embrace diverse perspectives and work together to elevate the best ideas. The Naysayers are the most common and harmful of teams. As a functional owner (e.g. product, design, marketing, etc.), it’s easy to feel that if you didn’t conceive of the idea yourself, it will reflect poorly on you or must be flawed (you’re supposed to be the expert!). But many times the best ideas come from a fresh perspective that can be challenging for functional owners to have themselves. The Yessirs are less common than The Naysayers, but equally dangerous because when you’re not the functional owner, there are key details you’re not aware of that must be factored in. The Collaborators represent what we must strive to cultivate. These teams welcome new ideas from outside of their function, and they recognize that there’s always something to learn from new ideas. They also know that the initial presentation of an idea is likely to have flaws, and so they work to better understand and improve the idea. I’ve worked with a lot of teams over my 20+ years as an entrepreneur. I’ve never felt more alignment with our teams at Hungryroot than I do today, and it’s because they’re a team of collaborators. There’s very little ego; the focus is on how we can collectively deliver the best possible experience for our customers. It makes the work more fun and allows us to be more innovative. Have you found ways of fostering this collaborative spirit? I’d love to hear from you.

Getting our customers' groceries from the fields to their doorstep takes hard work. On a recent visit to one of our friends and partners, Riverfresh Farms, we got a behind-the-scenes peek at how it all happens—and left more grateful and inspired than ever. Take a look!

This week, Ben McKean, our Founder and CEO, was honored by Goldman Sachs as one of the Most Exceptional Entrepreneurs of 2024, at its Builders and Innovation Summit in Healdsburg, California. Ben joined other entrepreneurs as well as esteemed speakers such as David Solomon, Fred Smith, Candace Parker and more for a two-day event sharing insights on topics ranging from values leadership to the geo-political landscape. The team at Hungryroot is grateful to Goldman Sachs for the recognition. #gsinnovators

161,312. That's how many meals are being donated to local food banks through our in partnership with Mighty Spark.   This summer, we pitched in to support Mighty Spark's mission to donate 1 million meals to Feeding America each year. Every Mighty Spark purchase made through Hungryroot from May through July generated a meal donation. And we're thrilled to share that thanks to this partnership, 161,312 meals will be distributed this fall to food banks that are currently in the greatest need of funds and meals.   Thank you to Mighty Spark for including us in this amazing initiative—we're proud to be part of it!

Prediction: In 2025, you’ll notice this change in how you engage with AI: Today, interactions with AI are primarily human-driven. While AI may answer your search queries, the interaction is initiated by a human. While virtual assistants can respond to voice commands, it’s still the human who initiates the conversation. While AI may recommend products to buy, it’s still the human who adds the items to their cart. This may feel like the desired behavior — do you really want AI incessantly bugging you? — and there are legitimate ethical and privacy concerns with too much AI autonomy. But there are several use cases where we will want AI to just take care of something for us, and we’re nearing that being more commonly the case. Take automated task management, for example. An AI-powered task manager would not only suggest tasks based on your priorities and deadlines, but also proactively schedule and complete tasks on your behalf. For example, it might book meetings, send emails, or even pay bills based on your preferences and habits. Or take what we are building at Hungryroot. Our AI proactively fills your shopping cart based on your predicted needs and delivers them to you without any action required (in fact, 20% of our customers don’t even view SmartCart's selections before purchasing them). As AI continues to evolve, we can expect to see a shift from human-driven to AI-driven interactions. Are you ready for a future in which AI takes the lead? 

What if your idea of healthy eating is different than what your body actually craves? SmartCart, the new AI system we introduced last week at Hungryroot, knows that people’s tastes are more complex than they think. You may believe you only like meat and potatoes, but if you enjoyed that Loaded Carnitas Sweet Potato dish that SmartCart introduced you to, it can recommend other dishes you’re likely to enjoy that you’d be unlikely to choose for yourself. That’s part of the reason Hungryroot customers triple the variety in their diet after joining the service. This also means SmartCart picks up on surprising insights about American eating patterns. For example, customers who tell us they prefer low sodium foods end up buying higher sodium foods than everyone else. When customers explicitly tell us they prefer low sodium foods, as you would expect, SmartCart prioritizes low sodium foods. But these customers tend to remove those recommendations and replace them with higher sodium foods. Why? One potential reason is that people who crave high sodium foods are more likely to have been told that they should eat less sodium. So when they’re asked if they prefer low sodium foods, they say yes, but when they select for themselves, they gravitate toward high sodium foods. This is just one example of the opportunity that SmartCart presents in helping guide people to the foods that will help them feel their best. For these types of customers, SmartCart can identify the low sodium foods that are most attractive to people who crave high sodium meals. Or it can slowly adjust the sodium levels of your food down over time, helping you ease into a new way of eating. There’s a ton of opportunity for SmartCart to help people feel better each day by making it easier to eat healthy. We truly believe we are just getting started. Does your idea of healthy eating differ from what you crave? I’d love to hear from you in the comments. #ai #healthtech #personalizednutrition 

This week marks a major milestone for AI’s ability to improve human health and wellness. Here’s a sneak peek at SmartCart™ 🎉 Food is the foundation of health, and since 2019, Hungryroot has been at the forefront of AI innovation in the food industry. Our original algorithm, known internally as "box fill logic” (or BFL), leveraged an operations research algorithm to automatically fill customers' grocery carts based on their preferences. While two thirds of what our customers bought was chosen by BFL — meaning it underpinned our entire value proposition — it was limited in its ability to dynamically scale and evolve over time. About a year ago, we started developing a new system entirely from scratch by our incredibly talented and experienced in-house team of data scientists, ML and OR engineers. The result is SmartCart™, a first-of-its-kind AI system. It’s comprised of ten machine learning models that integrate into an operations research algorithm and analyzes millions of data points to recommend groceries, recipes, and supplements to help you and your family live healthier, more joyful lives. It’s like having your own personal assistant for healthy living. Each model serves a unique and specific purpose. For example, one of the models, which is built using an approach similar to ChatGPT, analyzes your recent orders, cross-references data from other users, and applies machine learning to suggest the ideal characteristics of your next order — low sodium, quick-to-prepare meals, or high protein, on-the-go snacks, for example. Given the factors that matter most vary by person and evolve over time, there’s a model that applies a weight to each of the other nine models, all designed to optimize customer satisfaction. This flexible, customer-centric approach replaces the rigidity of traditional rule-based systems and is cutting-edge in the field of AI. (Hungryroot has already been granted several patents on the system.) SmartCart™ is already showing results. Customers who use SmartCart™ order twice as often on Hungryroot as customers who shop on their own, demonstrating its ability to help people achieve their health objectives, save time, and discover new foods that bring them joy. 90% of SmartCart™ customers report progress in their health goals since joining the service, and on average, they save 3 hours each week on meal planning, shopping, and cooking. They also discover 3 times more food than shopping on their own. SmartCart™ is a paradigm shift that will require multiple iterations, but I’m so incredibly proud of the team, and I can’t wait to see what our customers think.