Fidji Simo - Creating Delightful Consumer Experiences - [Invest Like the Best, CLASSICS]
What this covers
Patrick talks with Fidji Simo, who ran the Facebook App before becoming CEO of Instacart. The conversation moves between her product career and the grocery business she led. She describes Facebook's playbook of treating unintended user behavior as a signal of demand, with Marketplace as the example. She also recounts about a year of failure with Facebook Watch, when she decided the problem was her own leadership rather than the difficulty of the task, and that pressuring teams harder was the wrong response.
On grocery, she cites a roughly $1.1 trillion North American category with online penetration at 12 percent, against 25 to 30 percent in fashion, beauty, and electronics. She pushes back on the idea that online grocery has peaked or cannot be profitable. The operational detail is concrete: orders can be batched because groceries need no temperature control, shopper aisle data compounds into faster picking, and Instacart's e-commerce software runs grocer-owned sites such as Sprouts and Publix. She also covers advertising, including a claimed 15% sales lift and how Starday Foods A/B tests product names and packaging.
The most contested stretch is health. She argues that diet drives most health care costs, that doctors should be able to prescribe fresh food as easily as medication under a food as medicine model, and that insurers would save about $185 billion. She is cautious about AI nutrition advice, which she says is not yet trustworthy. She also predicts that AI will commoditize blogger recipes and push content toward personality-driven characters, such as a French chef.
Simo argues that Instacart's advantage comes from owning hard-to-replicate data, catalog, and physical-store infrastructure, and that AI makes consumer software easy to build, so durable value shifts to those hard assets and to ecosystem-wide 'grow the pie' partnerships.
- Instacart's 80,000 existing stores and proprietary catalog and shopper data are costly to replicate, while automated warehouses are capital-intensive and far from customers
- AI lowers software development barriers, so companies should rebuild products AI-native from first principles and focus on hard problems
- Instacart monetizes by platforming grocers and brands, so its success depends on partners rising together
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Without unique data that is easy to feed into models, it is very hard to benefit from new AI technologies, so Instacart made a large investment in data infrastructure.
“without unique data and data that's easy to feed into models, it's really, really hard to take advantage of these new technologies”
Instacart takes the e-commerce software it built for its own app and makes it available to grocers to run their own owned websites, such as sprouts.com and publix.com, making Instacart a technology platform rather than only a delivery service.
“We actually take the software we built for the Instacart app, and we make it available to every grocer that wants it on their owned and operated properties.”
Instacart Health uses Instacart's infrastructure to let doctors and hospitals prescribe fresh food as easily as medication, with payers reimbursing it, under a food-as-medicine model.
“making it available to the health care industry so that they can scale food as medicine”
About 80% of health care costs come from complex chronic conditions spanning multiple body systems, yet care is organized by specialties that each look at one organ, so patients bounce between specialists without a whole-body view.
“about 80% of health care costs come from complex chronic conditions that affect multiple body systems. And yet, the health care system is still organized by specialties that look at you as if you were just one organ.”
AI assistants could democratize nutritionist-level diet advice that is now a luxury, but because health mistakes are scary, the technology is not yet trustworthy enough and still requires human intervention.
“if all of that knowledge is captured by an AI assistant, you can make nutritional advice completely democratized”
Many expected grocery to return to stores after the pandemic and some think peak online penetration has been reached, but at 12 percent online penetration grocery is still vastly behind other categories, so the shift is far from finished.
“everybody was, oh, after the pandemic, everybody is going to go back to stores. And fundamentally, we've reached peak online penetration for grocery.”
AI will make software development easy and lower barriers to entry, so after a decade of value from easy consumer software, the advantage will come from building much deeper technology and tackling hard problems.
“AI is going to make software development easy. And that means that there's going to be fewer barriers to entry.”
The most common investor misperception is that online grocery models are fundamentally unprofitable, and the industry must show that these models can be very profitable at scale.
“there is still a big cloud about the industry of thinking of these models as fundamentally unprofitable”
About 87% of health care costs come from diet-related diseases, yet it remains far easier for doctors to prescribe medication than to prescribe a diet.
“about 87% of health care costs come from diet-related diseases. And yet to this day, it is much easier for a doctor to prescribe medications and to prescribe a diet.”
Recent studies suggest health insurers funding fresh food for food-insecure people and those with diet-related conditions would save about $185 billion in health care costs from avoided ER visits and medication.
“if health insurance companies were funding the cost of fresh food ... they would save $185 billion in health care costs down the line”
Grocery orders can be batched, with a shopper picking two or more customers' orders in one store trip, because groceries do not need temperature control, which makes this cost-saving approach harder for restaurant delivery.
“most of the time, you don't need the items to arrive at a certain temperature. That's much more the case in restaurant delivery.”
Content will evolve toward personality-driven AI characters, such as a French chef or a warm grandmother chatbot inside a grocery app, so creator content will be built around different AI personalities.
“what we call content is going to evolve toward being much more personality driven. The next wave of creator content by these different personalities for different chatbots.”
Because shoppers have walked store aisles hundreds of millions of times, Instacart can tell them the optimal aisle order and product pick sequence to minimize in-store time, so scale data compounds into efficiency.
“We now have shoppers that have gone down the aisles of grocery stores hundreds of millions of times. And that means we can tell them exactly how to optimize”
Commerce is an unnatural search-select-cart process, whereas people plan meals from constraints like budget, family size, and allergies, so a natural-language engine that turns that intent into a one-tap cart is a more natural interface.
“the biggest change that we're seeing with generative AI is the shift towards natural language”
Data shows the first grocer a customer buys from online is usually the one they stick with, so grocers resisting online shift risk losing share to more aggressive competitors.
“There's also a lot of data that shows that the first grocer that you end up buying online from is a grocer that you stick with.”
Instacart's hard-to-replicate asset is translating a dinner intent into groceries at the door within two hours, by connecting roughly 80,000 stores and 1,100 retailers, because the physical world remains difficult to navigate even with AI.
“we are this unique translation layer between the digital world and the physical world”
Automated grocery warehouses are capital-intensive and must be located hours from customers to aggregate orders, so they offer next-day delivery; using the 80,000 existing stores is the optimal way to deliver within one to two hours.
“it is actually relatively inexpensive to send people to a store to collect these items and deliver them to the customer compared to the capital intensiveness that it takes to recreate an entire warehouse”
When Facebook Watch failed for about a year, Simo's helplessness came not from the problem's difficulty but from her own leadership falling short, and pressuring teams harder was the wrong response.
“it wasn't that the problem was hard. It was more that I thought my own leadership wasn't up to what I wanted it to be to solve the problem.”
Content is fundamentally inspirational, and a business built on inspiration rather than utility is more valuable, which is why Instacart integrated recipes and creator-curated lists to move customers out of routine.
“when you develop a business that is based on inspiration, not just utility, you have a fundamentally more valuable business”
Instacart's internal studies show brands using Instacart ads see a 15% sales lift, which Simo describes as massive for the industry.
“by using Instacart ads, brands are able to see a 15% sales lift, which is actually really massive in that industry”
Grocery is the largest commerce category at about $1.1 trillion in North America alone, yet online penetration rose only from about 3% before the pandemic to 12% after, versus 25% to 30% in fashion, beauty, and electronics.
“Grocery is the largest category of commerce. We're talking 1.1 trillion in North America alone. ... It was 3% penetrated online before the pandemic. And then post-pandemic, we are at 12% online penetration.”
Starday Foods A/B tests product names, packaging colors, and concepts through Instacart ads, treating CPG product design like software feature optimization instead of relying on focus groups.
“they A/B test the name of the product. They A/B test all of the colors of the packaging. Instead of having focused group of five people, they put that through Instacart ads instead.”
AI will commoditize some content, such as blogger recipe content, because AI can generate entirely new recipes tailored to a family's taste, which frees creators to build more inspiring experiences.
“which content gets fully commoditized versus which content ends up being truly unique”
The Facebook product playbook was to find what users already do with a product not built for that use, treat it as a signal of demand, and build a product around it, as with Marketplace, where people were already trading goods via the core product.
“understanding what users were already doing with a product that wasn't necessarily built for that use case as a sign that there was something there that you could productize”
Hard constraints such as a requirement to be profitable and have strong unit economics raise the bar on the talent a company attracts and the execution it gets, differentiating good companies from great ones.
“constraints sometimes create just a better business and better operations”
Brands are going through the same digital transformation as retailers, and because consumers stick with a brand they find online, category leaders must act first to protect share and new entrants can gain share.
“when a consumer finds a brand that they like online, they tend to stick with that brand. And so you really want to be first to make sure that you maintain your category share.”
Companies should not bolt AI onto existing products but ask how they would rebuild their product from scratch with current tools, possibly replacing a search box with a conversational experience, mirroring how mobile-first apps beat shrunken web sites.
“having AI not bolted on, on the side to your current experience, but thinking about if you were solving the problem that your company is trying to solve, now from scratch”
Data shows an omnichannel customer shopping both online and in-store is more valuable, retentive, and profitable than online-only or in-store-only customers, so the goal is moving in-store-only shoppers onto all channels rather than shifting them online.
“there are tons of data showing that an omnichannel customer that shops with you both online and in-store is more valuable, more retentive, more profitable”
Retail media networks give small emerging brands an affordable, measurable way to reach customers, letting a brand sold in a few grocery stores use online ad spend to move products off the shelf, unlike costly TV or radio.
“it also gives a shot to totally emerging brands because if you are a really small emerging brand, you have just developed this one amazing product, good luck figuring out how to spend your money on TV or radio.”
Tech companies have been given a pass on base compensation, so Instacart treats stock-based compensation as real compensation and wants to be profitable net of it before deploying cash on M&A or growth investments.
“tech industry has very much gotten a pass in the past for that base comp. ... really wanting to be profitable net of stock-based compensation.”
Simo's management method is to identify each new team member's distinctive strength, map it to problems or opportunities, and push people to articulate their superpower so it can spread to other teams.
“trying to understand what is the thing that, that person really spikes on, the thing that makes them really unique and map problems we have to solve”
During the mobile shift, companies that shrank web experiences onto small screens lost to companies that used mobile-specific capabilities like location awareness and cameras to create mobile-first experiences, and the same split will occur with AI.
“there was a clear contrast between the companies that were embracing the shift to mobile by taking their web experience and desperately trying to shrink them on a small screen”
Sending people to existing stores to collect items is relatively inexpensive compared with the capital intensity of building new warehouses full of robots, which is why 80,000 nearby stores are the biggest advantage.
“it is actually relatively inexpensive to send people to a store to collect these items and deliver them to the customer compared to the capital intensiveness”
Simo's core value is grow the pie: the best companies make the whole ecosystem rise, creating value for partners who then benefit the company, so Instacart invests in making retailers, brands, and healthcare partners successful.
“the value I care the most about is this value called grow the pie because I fundamentally believe that the best companies are really making an entire ecosystem rise”
Simo calls herself a pragmatic technologist because she judges technology by whether it solves real human problems, and technology-for-its-own-sake debates, such as AI doom, become theoretical and distract from that anchor.
“where we go wrong is when we start thinking of technology for technology's sake and when the debate gets very theoretical”
Instacart lost money on every order until it reached roughly 100 million orders, and scale in order density was what made profitable unit economics possible.
“We used to lose money on every single order until we got to 100 million orders.”
Simo connects learning across domains, such as applying knowledge from a chronic-condition medical center she co-founded to Instacart's Medicare and Medicaid partnership work, so that lessons from one part of life are applied to others.
“I try to connect the dots between all of my different passions.”
Building a catalog covering millions of items with real-time availability across U.S. retailers, nutritional values, and ingredients takes years and is the foundation for any grocery AI experience.
“building a catalog of millions of items where you know the availability all across the U.S. ... that is a very hard thing that take years to build”
Instacart's marketplace is extremely incremental to grocers' own sites because surveys show it captures customers they would not have won through their own websites, serving those who want one membership across many categories.
“We've had many surveys showing that it is actually capturing a customer they just wouldn't capture otherwise if they only focus on their own website.”
Instacart shoppers differ from app-based delivery drivers because about 70% are women and more than half are moms, and since roughly half their time is spent picking in store, they face lower fuel costs than drivers on other platforms.
“about 70% of them are women. That's really not the case in other apps. More than 50% of them are moms.”
Instacart's culture of sweating every penny, such as celebrating a $0.01 per-delivery saving, is a key reason its unit economics are strong in a market where many competitors failed.
“that culture of sweating the details, sweating literally every penny is what has made Instacart successful”
The most common investor misperception is that grocery delivery models are fundamentally unprofitable, and Instacart is only an app, when its partner network contributes substantially to the business.
“there is still a big cloud about the industry of thinking of these models as fundamentally unprofitable”
Any technology has large potential for both good and bad, and the job is to maximize the good and minimize the bad by anchoring it on solving real human problems.
“I'm thinking of any technology as having a lot of potential for good, a lot of potential for bad. And our job is to make sure we maximize the good”
People spend far more time watching food content than preparing and eating food, which Simo cites as evidence of content's central role in consumer behavior.
“People spend way more time watching content about food than preparing and eating food.”
Focusing on free cash flow is what never serves a company poorly, because it creates options to reinvest in growth, including M&A and experience investments, when opportunities arise.
“the thing that never serves you poorly is focusing on free cash flow. And as a result, we are very focused on that because that gives us a lot of options”
Instacart already uses AI across search, replacement-item recommendations informed by hundreds of millions of replacement data points, and routing shoppers to stores so orders arrive as fast as possible.
“search is obviously powered by AI. Or replacement algorithm, if something isn't available in the store, what we recommend instead, that's powered by AI”
Brands should design for the online consumer as a proxy rather than designing for Instacart specifically, targeting a category of products consumers have demand for in a niche that is not oversaturated.
“It's more designing for the online consumer, which we have a proxy for.”
During current market conditions brands have cut unmeasurable top-of-funnel advertising and moved dollars into high-performance targeted ads placed while a shopper builds a cart, which can reach the customer within two hours.
“brands have tended to cut down on unmeasurable top-of-the-funnel advertising and put more dollars behind very high-performance targeted advertising”
Simo joined Instacart because building a technology platform that helps the entire grocery industry digitally transform was more interesting than running only an online grocery delivery service.
“if it had just been an online grocery delivery service, which is what Instacart is known for, that would have been less interesting to me”
Grocers want direct relationships and remarketing capability with their customers, and Instacart powers their own-property sites with the same technology it built for its app, letting grocers collect first-party data.
“they're collecting a lot of data from their customers there, where they can build a completely direct relationship with their customers”
A fairly sizable part of Instacart's GMV already comes from transactions on grocers' own sites, monetized through a mix of fulfillment and advertising, making the platform business already very meaningful.
“Actually, a fairly sizable part of our GMV comes from transactions that happen on our grocers' sites.”
Simo learned quickly in her CEO role by openly stopping meetings to ask basic questions and asking experts to explain topics like capital allocation and HR from the ground up, which she credits for ramping up fast.
“I am never ashamed of stopping a meeting, hey, sorry, dumb question, I do not understand this. I want to learn.”
Media exposure, such as a documentary about HEC business school seen at age nine or 10 and later depictions of the U.S., expanded Simo's sense of what was possible despite her small-town upbringing.
“I was, I think, 9 or 10 when I saw a documentary about our HEC business school”
Facebook taught Simo to combine long-term vision with short-term excellent execution, since a long-term mission is never reached without short-term excellence.
“it really taught me how to combine this long-term vision with short-term excellent execution”
The thing that matters most to end customers is whether they trust the brand and get groceries quickly, which existing stores provide better than automated facilities.
“the thing that matters most to the end customer is, are you a brand that I trust and are you going to get my groceries delivered on time and quickly”
If starting from scratch, the ideal grocery system would co-locate fully automated facilities for high-velocity items with physical stores where staff manually pick specialty items.
“we would likely build a mix of fully automated facilities and the grocery stores, as we know it, usually co-located”
Fully automated grocery competitors are usually many hours from the customer and so offer next-day delivery, which has a market but is not the majority of demand.
“they're usually many hours away from the end customer. And therefore, they offer very often next-day delivery, which there is a market for, but that's not by any stretch the majority of the market.”
The health care system has not been optimized to serve patients because of misaligned incentives and oversimplification of the complexity of the human body.
“we have a health care system that for many reasons, hasn't been optimized to really serve patients. And I think a lot of it has to do with sometimes misalignment of incentives”
Future AI could let users photograph their fridge or a cookbook and automatically reorder the missing ingredients, automating the cumbersome step of turning recipes into orders.
“you can take a picture of your fridge, and you can get an immediate answer on well what's left in your fridge.”
Grocery was not technologically enabled because large players resisted online transformation until it became necessary, and now virtually every grocery retailer recognizes online will become part of its business.
“A lot of the large players resisted online transformation until it was absolutely needed. And now the big change that we are seeing is that there isn't really a grocery retailer who isn't embracing online”
Instacart's incentives are aligned with grocers because moving customers online increases their value to the grocer, so Instacart's success does not come at the grocers' expense.
“our incentives are completely aligned with our grocers' incentives because as they move these customers online, these customers become more valuable to them”
Knowledge Simo gained from a weekend medical and research center for complex chronic conditions helped Instacart pursue partnerships with Medicare and Medicaid so its service could become a reimbursable benefit.
“this is knowledge that I was able to bring to Instacart when we were trying to figure out, okay, how do we crack getting into partnership with Medicare and Medicaid programs”
Many shoppers see themselves as customer service workers who resolve out-of-stock problems with customers in chat, and this service orientation is why shoppers attract customers who value that interaction.
“a lot of our shoppers really love the customer service aspect of Instacart”
Instacart has released a ChatGPT plugin that lets users describe a need in natural language, such as snacks for a 15-kid party, receive product suggestions, and order them in one tap.
“we've already released a plug-in for ChatGPT, which allows you to go and basically start talking about anything you want to do”
Stripe co-founder John Collison reportedly said that after believing they would disrupt payments, they found payments really is a scale business, a parallel Patrick drew to Instacart's unit economics.
“oh, it turns out payments really is a scale business.”
Facebook Watch, a video destination Simo pitched, was not working for about a year and needed many pivots before becoming a successful product responsible for a large chunk of Facebook's time spent.
“for about a year, it really wasn't working at all. No one was going there, and it took many pivots”
Patrick compares customers expecting a favorite brand on Instacart to fans expecting a favorite band on Spotify, implying brand availability drives satisfaction.
“If my favorite band is not there, I get upset.”
Patrick argues that during the era of high-margin software, companies lost sight of the fact that software exists to serve customers' end goals.
“we lost sight of the fact that we create software for a reason, to serve customers' end goal”