Chaz Flexman (Starday Foods) - How He's Building The Next Food Conglomerate With Data Science

0:00This episode is brought to you by Vobin from Carta. Vobin from Carta is the easiest way to launch and run your venture investing. They offer SPVs and fund vehicles for GPs at all stages of the journey, from your first syndicate to operating a multi-million dollar venture fund. If you're interested in investing in startups, stick around after the episode where I chat with Gabriel Shin from the Vobin from Carta team, who shares his perspective and tips about how to start investing and how Vobin from Carta can get you set up. The link to Vobin from Carta's website is in the show notes. Where did the name Sarday?

0:35What was inspired by the name? Was it The Weeknd? No, it wasn't. I was not actually looking for a theme song. We had a different name that we're incorporated under. And for some reasons I won't go into, realized we needed to change that name shortly before launch. And it was kind of making sure that this name worked for trademark and as a consumer brand. But it's really around like, great, we're building a constellation of brands. Our partners are our star partners. It had some depth to it that led us to kind of like, oh, that's a great name that can encompass a whole lot of different brands and products underneath it.

1:08Hey, I'm your host, Mike Gallup, and this is the Consumer VC Podcast, where we discuss the intersection of venture capital and consumer innovation. If you're enjoying the show, also subscribe to the newsletter at theconsumervc.com, where you're going to receive all new episodes straight to your inbox, and a weekly recap of all the consumer deals that are happening. All content and episodes are for informational and entertainment purposes only and is not investment advice. This episode is brought to you by Vauban from Carta. Vobin from Carta is the easiest way to launch and run your venture investing.

1:39They offer SPVs and fund vehicles for GPs at all stages of the journey, from your first syndicate to offering a multi-million dollar venture fund. If you're interested in investing in startups, stick around after the episode where I chat with Gabriel Shin from the Vobin from Carta team, who shares his perspective and tips about how to start investing and how Vobin from Carta can get you set up. The link to Vobin from Carta's website is in the show notes. Our guest today is Chaz Flexman, who is the founder and CEO of Sarday. Sarday is a next generation food conglomerate that uses data science to predict product market fit and create new brands.

2:15We discuss why he left pattern brands to go into food, how they create and launch new products very cheaply, how they think about data and trends and different categories to enter, and also how to build relationships with retailers. Without further ado, here's Chaz.

2:35Chaz, thank you so much for joining me here today on the show. How are you? Doing well. Appreciate you having me on. Thanks for your time. So you left Pattern Brands, which was an aggregator holding company focused on brands for the home. I know you come from a software tech background. Why CPG? Why food? Yeah, so I originally kind of got into the hybrid of software and hardware when I ran a company called Wink, which is a smart home platform. Kind of Alexa, pre-Alexa, but we made a couple of our own products and definitely fell in love with the haptic feedback loop of seeing a product you made in someone else's hands.

3:12It's just like it's really kind of powerful, you know, incredible feedback loop, which I really enjoyed, but also love how the combination of software really brings an element of data and thought and, you know, kind of complexity to it that I've always really enjoyed. And so I've kind of straddled, you know, between the two for a lot of my career. after I left Pattern I joined up with the folks over at Equal Ventures they're at CTH Fund out of New York and they're a very thesis driven fund and we're writing this kind of memo around you know the future of F &B and retail and I was like sure let me like scrub it and help you I kind of joined as a venture partner and not sure if I wanted to go back into venture full time but I'm just happy to help kind of fill out this thesis and about a month or two into it I was like oh I don't want to invest in this like, I want to go build it.

4:08Like no one's going to walk in the door with a better background than me. And this feels like kind of a summation of everything I've done in my career to date. And so started incubating it with them over the course of 2020 and then off the races. I think why food, there's a couple of things. Like one, it's really trend and data driven, like attribute driven. So you can use data to really like drive a lot of your decision making. And that wasn't something that was like prevalent today. Meaning, you know, while it wasn't a consensus thing or consensus approach. It was one I thought there was, you know, a lot of opportunity to kind of create change and more predictable outcomes.

4:40And two, again, it's just that haptic feedback loop. I mean, one of my favorite moments from this company has been when a friend of mine in Nashville went to a kind of have wine and drinks at a friend's house and they were serving gooey. It just happened. It's like, oh, it's my buddy's bread and sent me a photo of it. And just seeing that and hearing from consumers and how much they love the product is just second to none. Can you talk a little bit more about when you were at equal, what that original thesis was? Truthfully, it's pretty spot on to what we've built to date. You build a BenX great F &B conglomerate using data at the core of how you understand pent-up consumer demand and bringing brands to market that you know are going to work and predicting product market fit.

5:20I'd say the only thing that we've realized more and more over time is that we're a B2B2C company. So we partner with big retailers to help make sure that we're solving their assorban risk with our products, but also giving them category lifts, meaning similar velocity about higher what's on shelf and at a slightly higher price point that consumers are willing to pay for. I fundamentally believe that you build your bed online, you build your business in retail, and that's only kind of doubled down on that philosophy as we've kind of gone along. I really appreciate that. I know that part of also the reason why you started Sarday as well was due to some of the principles and methodologies that you saw happening in vertical SaaS and kind of applying it to food.

5:59Can you give like a breakdown of like what that means in this kind of third generation is, as you put it, of vertical SaaS? One of our investors is Fiji, the CEO of Instacart. We were having this conversation at some point and she's like, Chaz, you know, I used to run advertising at Facebook and coming to F &B, I thought everyone's going to be doing what you're doing. It's like, yeah, it's not rocket science. It's just taking all the stuff we know from consumer software and applying it to F &B. And truthfully for us, we think about like the product is the process, the output is just the brand.

6:24The data models that we built, like, you know, are kind of the core intrinsic value and the brands are what comes out of it. So analogy I like to give is Ford with the Model T. You know, the Model T, like they invented the assembly line. The output was the Model T that was, you know, more reliable, better priced, more, you know, a better car. But the true innovation was actually on the assembly line side. So we think about it from the same way. The core underlying data model lets us create a more predictable sort of outcome and predict product market fit. And that's where it is. That being said, you know, I always get the question like, well, why don't you sell your software?

6:58And truthfully, F &B is littered with companies that are trying to sell data into F &B. And the reason for that is it's a very barbelled industry. You've got like the Nessies of the world that while they might spend$100 million on Nielsen, don't like really use data at all in terms of their decision making around what products to build or how to run their business. And then you've got like a bunch of smaller brands that, yes, they might pay for spins just to see how their products are selling, but don't have the balance sheets to actually really pay for software. there's no like real strong SME market which means you know if I know a brand can can do 50 million dollars or 100 million dollars or 200 million dollars like why would I want to sell that data to someone else instead of like use it ourself to then go build those brands and capture the value.

7:40You know that makes sense in terms of in terms of why you would actually want to build those brands yourself as opposed to just being a part of maybe like someone else's stack per se as they're building up their own brands. Can you walk me through a little bit about how you are leveraging data science, as you say, to like de-risk a product market fit or to find product market fit and figure out which types of products you actually want to launch? Of course. So, you know, like what's been historically done with like spins or Nielsen or IRI, they basically just kind of tell you basket share or like what's happened in the past of like nut butters, jellies, and jams.

8:16We start with a core consumer need and build around that. So it's like that's something like anti-inflammation or gut health or top nine allergen free or sustainability. And we create these ontologies around them, which is kind of the umbrella under which the data lives. We use third party data APIs. We do web scraping. We've got some proprietary data partnerships. We kind of suck that in and then use like NLP on top of TikTok or Instagram because consumers aren't talking about products or their consumer need like they are nut butters, jellies and spread. So you've got to use a bunch of NLP to understand kind of what they're saying and how that's related.

8:53We run a bunch of models to kind of say, like, great, what are the core attributes or product categories that are then associated with this core consumer need? And let that, you know, really kind of dictate where those opportunities lie. A great example of that is low FODMAP. Our third brand, Kazumi, is a low FODMAP rice blend. There's more demand online for low FODMAP than there is for collagen. You know, it's the normal prescribed diet to fight IBS and bloating, particularly within women. And yet there's only a couple of brands that are on the market servicing it today. And so we can use that data to really make sure that we're finding like pent up consumer demand.

9:30We then create a PRD just like you would in software and, you know, work with our R &D firm to go spin up a couple thousand samples and do blinding consumer testing through a third party service to test for taste and texture and all those kind of key attributes. Because this is a repeat purchase at the end of the day and, you know, consumers have to love it. And so it's not, you know, it's not just like 10 of my friends be like, Chaz, this tastes great. Lastly, if you think about walking down the aisles of a Target or a Kroger and you see a box of chiskets that's sitting on a shelf, to us it's just an adding of that sitting on a shelf, an adding with kind of only five main variables of diet, health, ingredient, cuisine, and then brand.

10:05So we can start running a multiple variety of tests just like you would in consumer software to figure out like, great, if I call something dairy-free versus vegan, how does that change pent-up consumer demand? or you know there's a set of attributes you might have to have like non-gmone gluten-free there's ones we call them the mba's the most valuable attributes that give you the right to win so think um halo top with 250 calories per pint have they been like the low sodium brand probably wouldn't have worked but we can test into a lot of that you know that consumer demand and understanding everything from um you know flavors to the color of the box to like those attributes for instance you'll see on kuzumi with low fodmap it's actually a secondary attribute not the top MVA attribute.

10:45That's because we can actually look at where something is in its adoption curve and say, like, great, this isn't ready for, like, the primetime attribute, but, you know, as gut health is more of that, like, hook to consumers, kind of orient around that. But all of this kind of, like, flies in the face of how most F &B brands are built, which is normally around that kind of founder narrative of, I have allergens as a kid, so I'm going to get, like, an allergen-free cookie brand, and that's, like, the qualitative insight that's going to lead me to the promised land. where, you know, I don't think my taste buds represent mass America.

11:16I don't think I've got any God given insight. I, you know, I'm actually the last person to try all of our products because I don't think my taste buds matter. And so it's really making sure that data is at the core of driving, like where and what we build and let consumers tell us. So, so you kind of start, I mean, as you say, you kind of start with like what like the macro problem is that you're looking to solve and then see like kind of evaluating what the pent up consumer demand, doing a bit of analysis in terms of what other brands are there and if there's an opportunity to introduce a new one if there is only a small number of brands that are serving that consumer.

11:48Our model is really prioritizing a couple of things. It's one, trend relevance. Is this going to be something that continues to build over time or is it just a blip? Popularity and then correlation to consumer intent. How do you define in terms of if something's going to be a trend versus just a blip? Yeah, and so we're running a bunch of LLMs and regression models to see how we think this should play out historically versus certain areas of time versus in the future. We score what we call events, which are these different independent data points, and give it a ranking and scoring system to actually tell us what's the highest probability of success versus something that might be more of a slower burn.

12:26That's helpful. I think also you're saying how traditionally we've had a number of incredible founders on the show talk about this, how usually it's, it's usually a pain point from them, or maybe one of their friends, or, or maybe a, a sibling or a child or a cell or a family member, that they're not able to kind of have this product in market. So then they decided to go build it and then maybe realize that, wow, there's actually maybe a lot big, like, like, obviously, I mean, obviously, that's why they build it. But there's a much bigger audience here of consumers that are in the need that where where you're kind of more looking on the outside externally, thinking like what is kind of like the common consumer or like mass America is kind of needing for.

13:13Exactly. I also, you know, I think, and there's all credit to those founders and anyone taking the risk to go build something. I think sometimes those end up being a little more like a random walk down Wall Street, like some work, some don't. And we're trying to use data to make it like predictable success, right? It's like once I repeat that we can do this multiple times over with a consistent ability to predict product market fit at the end of the day. Because again, I don't think my taste buds represent Mass America, but I think consumers will tell you what they want if you know how to listen.

13:41Well, how also do you define as Mass America? Because, you know, and especially a lot of the opportunities that are like, or a lot of the companies that are coming on the show and also start days, like usually these are like premium, more expensive brands. And usually you're not, you know, for, you know, maybe like the everyday American. Secondly, how do you actually think about in terms of just MassAmerica in total and the overall TAM? It's something that retailers struggle with, which is like, look, just because something is selling well in 60 doors of Gelson's, does that mean it's necessarily going to work in MassAmerica?

14:10Like, maybe, maybe not. and so we kind of think about it as like building a an archetype and persona and psychographics actually around those core you know kind of the 12 retailers that really matter that drive most of the velocity understanding really great what are the psychographics of that consumer you know are they working out for strength management and stress management but not working out for weight management which might mean that like in a certain retailer products that are diet culture aren't actually going to work. And so there's, you know, something like keto might actually not be a fit.

14:44And so if you actually focus on like those kind of core retailers that matter, you can develop psychographics around it and who those audiences are, you know, across mass America. And so, yeah, you know, it's also one of those things that like if you can sell something for, you know,$20 in Air One, that's wonderful, but that doesn't necessarily mean an Indianapolis mom of two is going to buy it. And that's where you get to real mass and scale. And so we do a bunch of testing on price points where we, you know, run a bunch of traffic to our site. We'll A-B test different price points and see like where those conversion percentages change um to you know say like the great that's your market clearing price of what consumers are willing to pay uh we also spend up fake brands all the time to try and test them with that uh that core kind of mass market audience versus again just uh saying like hey this is what's worked on on the coast because that doesn't necessarily translate to you know middle america does this at all um when you think about launching new brands and thinking about if it's right for mass america i mean as you said just because it maybe works a product works and you know 20 galson stores doesn't mean it's going to work in kroger does that mean that when you think about um that you actually have a preference in terms of which which distribution channel you actually want to start in we think about starting with big box kind of from day one you know our first brand gooey rolled out in every kroger in america um you know it's going into target and hyvee and um yes we're getting into natural on the east coast just because that's got a really strong presence there but it's not saying we're starting with that local then regional, then specialty, and hopefully you're going to get to big box.

16:06But big box is what ultimately drives scale, right? That's where people do most of their grocery shopping. And that's where you get real velocity and real revenue from. And that's, you know, quarter like, you know, where we want to be building for is that kind of mass America. How much does it cost your initial run to actually like create a new product? Yeah, I mean, we can basically build a brand in less than six months for less than 200K. And so we think of that like R &D cost is like our CAC. And so, you know, if we're going and signing up a couple thousand doors, you know, those initial loaded NPO is only payback with gross margin because we can get to a gross margin profile at scale, you know, quite well.

16:41So we're not kind of slowly rolling out through different kind of stages of scaling up production. We're able to kind of go into mass production, you know, kind of right away. That means we've got to do a lot more work on the front end for on the R &D and development to making sure, you know, our formulations work at scale. but also means like we get to, you know, revenue and margins, you know, a great market profile faster. So do you do your own, your own manufacturing? No, we partner with various co-men and then a third-party R &D firm, but also have R &D in-house as well. Now you have four brands, is that right?

17:14Correct. From doing those trials, you say, you know, once you maybe produce, you'll produce, once you maybe roll up a brand or start a brand, you'll introduce a brand and do market research maybe on like 2 ,000 consumers and see if this brand has legs. How many brands have you maybe started and you're like, actually, this is not going to work? Or has it been done so far? The testing that we do is actually on the product itself, right? To get like taste and lecture and those key attributes. But what we've actually found, I thought probably at the beginning, we're going to have to kill a bunch of brands.

17:48We've actually found is because we have a stage gate model, just like it would in software, we ended up like iterating and going back. So like Kabea is a great example of this, where we got it into consumers' hands, and the scoring and feedback just wasn't strong enough. It wasn't bad, but it just wasn't overwhelmingly positive. So we went back to the model and said, great, let's go stack some attributes on this to see if we can find more of a hook. And we found this chickpea protein because we found consumers wanted more protein in their diet, but wanted non-soy protein in their diet. So we found this chickpea extrusion to make that part of the crunch of Kabea, but now offering us 15 grams of protein that's not soy-based.

18:23and we ran it back through the testing again. It's like now one of our highest scoring NPS products that we've made. And so it's more kind of iterative feedback loops where the cost of failure is a lot lower. There's also plenty of stuff that comes out of the data model that's just not feasible, right? Something might say adaption soups. And Lena, my co-founder, who's a three-star Michelin chef, food scientist, she's created the first gluten-free flour brand. So you can look at something and be like, hey, adaption soups, for instance, adaption's really bitter. You're not gonna be able to cover that flavor.

18:52So that's not going to work. So let's kind of throw it away. So it's really kind of a humans in the loop sort of thesis where, you know, you've got to understand like feasibility and viability, which go hand in hand. Are there macro trends that you might, or rather a problem that maybe is very attractive in terms of like, there's a lot of interest in this problem within food that you actually don't think is really like, is really more like the fad side than a trend side? I guess let me tweak your question a little bit. So we've done a lot around sustainability. And most of the brands that have come out are very much like nail on head, like alt-meat.

19:30And consumers, while there is a core part of the population that definitely wants that, to activate mass consumers, that might not always be the fit for them. So it's about how do you ingrain sustainability in ways that increases the actual audience size of consumers who are willing to pay for that. So that could be things like GUI with palm oil-free. That's ways to introduce sustainability while still being a core mass market product. or habea you know being like people want more protein their diet they want um you know plant based offerings but they don't want soy and so it's you know it's an offering that like gets people into the category and expands it versus just kind of nail on head of like oh we need to be alt meat or alt protein got it so just kind of not as like um like a direct kind of like in your face of like this is you know like a plant burger for example versus like a normal burger per se exactly uh and so it's kind of activating that consumer through like secondary dots which we think can have more of an impact because you're getting people, you know, to actually make sustainability part of their, you know, their spend, uh, without it just being like, Oh, I just want a plant-based burger of nail on head.

20:30Like I said, how, I mean, you, you categorize as well, Sarday from the, from, from the get-go is kind of more like a B2B2C, um, type business where you're selling to the retailers. And then of course the consumers are buying the brands, which totally get like, are you also doing like, how do you also think about like pull marketing, initiatives to actually drive consumer demands for your product that actually then they actually go buy? Yeah. So I think the core of the data model, we're getting to the why behind the consumer purchase. And to be understanding the why, that means you can, if you think about kind of the brand architecture, that's kind of like the core of the brand architecture that you're building around.

21:04It also means we can run that same NLP on TikTok to say, great, who are influencers that are talking about these products in the same way that we know is going to resonate with audiences? Or on the shelf danglers or out of home that we're running, knowing the core hooks of like why consumers are buying for GUI, like low sugar is kind of the core hook there. And so if you get to that why behind the brand, it lets you kind of activate the marketing side and be much more kind of pointed with it. And to take advantage of tools, whether that's Instacart's, you know, kind of ad platform, because you know the related search terms or, you know, the messaging that you need to have on your landing page of aisle to drive, you know, bogos and driving people in store to try the, try all the product.

21:42And it's really kind of using that why and those hooks throughout kind of all the different marketing, you know, tactics that we do. Usually kind of like the story for, for consumer brands is, um, not always, but you kind of start online, um, uh, you get demand and then, and then you kind of go, um, on the channel, um, after you have a certain amount of sales online. Um, and it seems like in this scenario, you're actually kind of starting, um, per se in retail. Is that, is that relatively correct or no? We kind of do it concurrently. So we can, you know, show retailers a product that's like ready to go, uh, even before it's launched.

22:15You've got to remember retail is focused on like line review cycles. So like they might not be opening up a category for six, seven months until you can go pitch in. You know, they select who it is and they close it for the next 12 months. And so it kind of depends on the line review cycle. And then when we have a brand that's ready, you know, I think we use DTC largely just for data capture around a bunch of the testing that we're running. We think about Amazon or like a Thrive Market as, you know, kind of a brand awareness and kind of building that hail effect. and then getting into big box as fast as possible.

22:48But our first brand we signed before it was even live with Kroger. So we've definitely seen retailers that lean in at times and are willing to say like, great, we believe and understand the data model and we want to participate. Got it. And so it's using those channels to convince retailers that, hey, this is first of all a trend and secondly, why we are the brand that's actually solving this problem. Yep, exactly. That's awesome. I remember reading an article and you saying how you wanted to build, that you're building maybe more of the Ford versus the Unilever. What exactly do you mean by that?

23:23Yeah. Again, it's that kind of the analogy I gave around the assembly line and the Model T. Tech is their underlying data models that we're building. The output's just the brand. And so it's really kind of starting. It's not just like saying we're going to be a house of brands. We want to build the next Nestle. We want to build the next great F &B conglomerate. but we want to use with data at the core that's driving the decisions to give consumers what they love. We talked a bit about how emerging brands usually start, right? It's usually maybe their own pain point usually, or like a family member or someone they knew a pain point.

23:56How do you think about innovation from like the corporate level and what are they doing wrong in your mind? Yeah. So one of my board members, Rick, jokes that he was the problem. He's a former consultant. And so Big Food is, for the longest time, been supply chain driven, right? They've had consultants that have told them to go put$100 million in your CapEx to get 10 % better throughput of your core ingredients. Go do multi-year derivative deals on those ingredients to lock in the prices for years. And that's why most of the things that they launch are supply chain driven, not consumer driven.