How This Builder Used AI to Grow a “Sleepy” Business to $12M

Liam:

Jonathan, excited for this episode. We just recorded the last episode in Palm Springs, and we’re also in Palm Springs. What a place to be — much warmer than SFO and New York, right?

Jonathan:

It’s lovely out here. I just came in today. I actually took a picture from the hotel of the pond outside for my four-year-old and he just loved it. He’s like, Papa, when can we do videos? We’re going to do a video after this.

Liam:

It’s a beautiful place and excited for so many people that are in our realm to be here this week. So many people have seen your journey, or maybe haven’t seen your journey and you’re quietly growing behind the scenes — which I feel like is the person you love to be. Just grow businesses as fast as they possibly can. But for people who don’t know you, why don’t you give a quick introduction?

Jonathan:

Yeah, so I’m Jonathan Shroyer. I would classify myself as a builder-scaler. As I was coming over here, I was reminded of a comment that my wife made to me in 2015, right before I joined Postmates, my first startup. She said, “Hey, you’re really bored in enterprise, and you’ve been in the enterprise for like 15 years of our marriage. Could you go get on board and go into startups?” So I really like the entrepreneurial, the builder-scaler, the innovator space. I cut my teeth with Microsoft in the early days, 2001. When Monster.com was kind of a big company. I was there. Lived in India, lived in Boston, lived in London — loved multiculturalism, loved getting different people’s viewpoints. But at the end of the day, I really love to solve problems and bring 10x value to those problems. That’s what got me into entrepreneurship with a couple of exits at Oficium Labs and Quimby, and why I get so passionate about helping entrepreneurs grow and building my businesses.

Liam:

I feel like joining Postmates so early would have been an incredible experience. Momentum solves everything in every business, and Postmates was growing absolutely bonkers when you joined, right?

Jonathan:

Yeah, I mean, it was insane. My wife actually said, “I wanted you to join a startup, but did you have to join one that was growing 10% to 15% per week?” Per week. It was insane in the early days. It basically meant 100-hour weeks, 120-hour weeks — which I didn’t have my four-year-old at that time, so it was doable. But you’re right, momentum does solve everything. What I learned at Postmates is when you’re growing at 10% to 15% a week, things break every day. You can’t try for perfection. You’ve got to be like, how do we move forward? What do we do differently? Okay, that just broke — how do we actually fix it, and think about what this is going to look like in four, six, eight months. You can’t work in years. You have to work in days and weeks at a company like that.

Liam:

What do you think the differences are between a company that moves slow and a company that moves fast? Can you replace just working hard?

Jonathan:

I think the big difference is mindset. Nelson Mandela said, “I never lose, I only win or learn.” In both types of companies that mindset works, because you’re figuring out what the learnings are and how you grow. In the fast company, that mindset is even more critical because you can’t let mistakes slow you down. If you’re moving 10% to 15% per week, you’ve got to say, “I learned it, it’s behind me, I’ve got to go.” With a slower company you might be able to stop, think about strategy, have a meeting about it. When you’re in a fast-growing company, you’ve got to limit the amount of meetings because it has to be about action and momentum.

Liam:

You purchased Linda’s a few years ago, which is now a fast-growing brand. When you purchased it, it had $2 million of revenue. What has that journey been like?

Jonathan:

Yeah, it was interesting. Linda’s — my sister owned it and had grown the business for the last 20 years. She said, “Hey, I’d like to retire. Do you want to come and buy this business?” And I’m in tech and gaming, had just had my recent exit at Oficium Labs at 20x, and I was like, I don’t know that much about ecommerce and the brand. So I went in and did an analysis, and I saw a $6 billion TAM opportunity sitting wide open with very little technology and very little general innovation happening — a good CAGR growth, recession-proofed, 4% to 6%, so not hockey-stick growth as an industry, but a huge opportunity where I could probably build a nice moat. We’ve grown it over the last two years from $2 million to $12 million. We’ll probably do $18 million this year. That 6x growth has been phenomenal because it’s been such a sleepy industry — we introduced things that got customers excited about coming to Linda’s.

Liam:

Walk us through what has actually changed since taking it on to drive that level of growth.

Jonathan:

I would say there are three main things. The first one is awareness. Linda’s had an amazing reputation in Texas for a very long time, but it wasn’t well known outside of Texas. We leveraged AI and different marketing capabilities to drive greater awareness. When I came into Linda’s they had 20,000 customers — now Linda’s has 400,000 customers. It was about how do we get out of the noise and focus on the signal of getting our message to the right customers in the right way. The second one is product expansion. When I came into the business, they had 4,000 products. When I looked at the TAM opportunity, I thought we needed 100,000 products to become the one-stop shop that creates joy for quilters. Most people would say, I’m going to build a big warehouse, buy a bunch of products, market it, and they’ll love it. But that takes a lot of working capital that I didn’t have. So instead I built a network of craft-specific dropshippers and a backend AI technology that plugs everybody in. Now we have 50,000 products but only warehouse around 3,000. We’re bringing other manufacturers and quilt stores’ brands and products to Linda’s customers while building up local economies. Then we flipped it on its head through Shopify Collective — we let every quilt store in America that wants have free access to Linda’s products and take a 15% commission if they sell it on their website or in their store. The third piece sounds simple but is super important at scale — amazing customer service. We get back to people within the hour. I learned this at Postmates: customers got hungry at 45 minutes, angry at an hour, hangry at 75 minutes, and something worse at 90 minutes. That taught me the importance of fast, quality service for customers who love your product and want it quickly.

Liam:

What does your team look like today to actually do all of this?

Jonathan:

We have what I call T-shaped employees — people who can go broad but also go very deep in a certain discipline. We’ve got 10 of them, and eight out of the 10 are family members. I kept my sister on as the affiliate director. My nephew is the powerhouse marketing face of the company — everybody in crafting knows Chief Quilter Corey from Linda’s. He’s phenomenally extroverted where I’m introverted. My other sister runs the retail store, my nephew’s partner runs the warehouse. We’re a family business and we brought some outside talent in to help us grow.

Liam:

Why do you think sleepy verticals are so underrated? Linda’s lives within a sleepy vertical. Everyone when they go and start a business chases the shiny thing that everyone’s talking about — Meta’s impossible to advertise on, organic gets you nowhere. Whereas some of the biggest success and profit margins seems to come from sleepy businesses with massive growth opportunity.

Jonathan:

The opportunity is to look at verticals that aren’t sexy — that people don’t want to say, “Hey, I raised $50 million on a $2 billion valuation because I’m changing the world.” Those are good. We need passionate people like that. But you also need people like me who aren’t those people, who say, “Hey, there’s something over there. The data is telling me there’s something magical, and I just need to go figure out how to get that magic.” I was watching Diary of a CEO and he was interviewing someone who found a sleepy vertical in trailer parks. Who’s going to go and invest in trailer parks today? But that guy sat down and said, “I think trailer parks are sleepy and I need to go after them,” and he made millions and brought value to his customers. You just have to find those areas that people aren’t paying attention to, look at the data, and if the data tells you there could be magic, go after it.

Liam:

You take on a new business doing $2 million — sleepy vertical, business with huge opportunity. What are you doing first? Particularly in a world like today with AI and so much automation.

Jonathan:

The CEO of OpenAI came out recently and said he thinks there’s an opportunity for somebody to build a billion-dollar company as a single employee. We’re not close to that yet, but we can be. And if you have that mindset when you’re building a business, you start asking: do I need to use AI for this, or do I need human capital? That changes the mindset. But actually the most valuable part isn’t even that — it’s the mindset shift itself. I come in with a laser-light focus view on certain topics. I have autism, so sometimes I really hyperfocus on things. I come in with a perspective that’s outside of the industry, and even my family members took time to adapt to having me as a business partner.

Liam:

Do you think a lot of people are struggling to adapt to AI because of job justification and fearing losing their job, or is it something different?

Jonathan:

I think it’s trust. If somebody believed that AI could help them become a superpower employee, they would immediately adopt it because that would make them indispensable — not everybody’s going to do that. But the chasm for you and me to adopt AI is much smaller than the chasm for somebody that’s not as integrated in technology. Some people see that chasm as too far. And if there’s not a bridge builder — somebody building that bridge to show them how to do it and build that trust — they’re going to have a difficult time adopting it. At the same time, you have technology companies laying off junior engineers and saying AI is better. So it’s a matter of trust. And over time, momentum will solve it with AI. You either get on the AI train and have a business in five to ten years, or you’re reluctant and go bankrupt.

Liam:

I compare the shift in AI to the creative shift when Canva started. Everyone feared Canva — every creative agency thought it was going to take their job. The best creative agencies today use Canva in their business. That same shift is coming with AI. Majority of agencies today fear it, they’re losing professional services work, they don’t embrace it, they’re losing clients. However, the best agencies that do will ultimately win.

Jonathan:

I think you’re right. But the more powerful thing everyone misses isn’t the human capital side — it’s actually the data. One of the things we learned at Linda’s is we needed to change the signal-to-noise ratio. We wanted to focus on the signal of the data so we could target the right customers and get them into our store to buy our products. We started using AI with data analytics first — understanding demography, targeting, identity, and how the business can best monetize while bringing value to the customer. That’s where I think AI is most powerful, and a lot of people are not using it that way today.

Liam:

How do you think businesses should be using AI at the simplest level?

Jonathan:

From my perspective, if you’re an ecommerce brand, at the simplest level you should be using AI for identity — that way you know 60 to 70% of your traffic on your site versus 10%, and you can monetize them 6 to 10x. You should be using AI for ads — you shouldn’t need a paid media person to tell you where to spend your money across Meta, Bing, and Google. AI should completely optimize that. And you should be using AI for retargeting, whether through email, SMS, or connected TV. If you want to come in and use AI, you can tap into your Shopify data and within less than two weeks, increase your revenue.

Liam:

What are you doing in email and SMS today?

Jonathan:

In email and SMS I do identity matching. There’s an AI brain that does the identity piece, and then there’s the flows and retargeting that sit below it. We plug into different platforms — some that help with identity, some with flows, some with CTV, and paid media platforms as well. The brain sits on top and the agentic bots sit below it, and that drives customer retention. We created some of our own IP because the solution we needed wasn’t on the market. I’m a buy versus build guy, but if it’s not there, you’ve got to build it.

Liam:

Why is retention marketing even important?

Jonathan:

You have to understand your LTV to CAC ratio. My vertical needs it to be 9. Mine currently is 25, but it needs to be 9 — margins are super slim in the quilting space, EBITDA of 10 to 15%. But the LTV to CAC is super important, and this is why retention matters — you need to figure out how many orders a customer needs to make to become your customer for life. For Linda’s, it’s 4.8 orders. Once they have 4.8 orders, they’re my customer for life. Every business with AI analytics can know what that number is. And once they hit that secret sauce, they can scale to first-order profitability. First-order profitability should be important for every ecommerce business. If it’s a subscription business there’s a little wiggle room, but for most consumable or single-sell product businesses, you should be focusing on first-order profitability. That way you can compete against the bigger brands who have the war chest of cash that you don’t have.

Liam:

So email and SMS is a huge part of your strategy for bringing shoppers back. Instant is a big part of that — maybe just explain to people how you’re using it.

Jonathan:

The way I do it today is I have the identities, and I use that identity for ads. Ads bring in new customers, but the identity also lets me identify organic customers that didn’t purchase as well as paid customers that didn’t purchase. Then I retarget them with Instant AI flows and with SMS to attract that customer and talk in their own language. What I love about AI and Instant is that it creates a unique flow for every customer. It’s the exact flow that Barbara in Wisconsin needed. It’s the exact flow that Linda in Texas needed to convert. But the last piece of our profitability strategy is connected TV. In America there are 550 TV streaming apps. That identity I have — I can now match to the TV channels they’re watching, and if they didn’t convert, send them a high-fidelity, low-cost TV commercial that reminds them why they love Linda’s. The ROAS on connected TV for me is 30x. They came to me via paid or organic, I retargeted them with SMS and email, maybe it didn’t work the first time, then I get them on TV. I call it the trifecta — the brain at the top for identities, and then the trifecta of targeting them. Most brands need a customer to see them 6 to 7 times before they purchase. For Linda’s it’s 30 times. So it’s really important that the trifecta is set up to ensure I’m always on the customer’s mind.

Liam:

Why don’t you fear experimentation? A lot of brands and founders buy what everyone else is using, they’re almost scared to be first. Whereas for Linda’s and throughout your career you seem obsessed with being first — if there’s no solution on the market, you’ll build it. Why are you so obsessed with experimentation?

Jonathan:

When I was young I grew up in a family of 10. In a family of 10, if you want to stand out and get the attention of your parents, you’ve got to be innovative, different, unique. You only have so much time. And then if you want to stand out in sports — I played basketball and track, which weren’t sports played by many of my peers — I had to do things differently. I adopted Kareem Abdul’s hook shot because I knew I wasn’t fast enough to shoot against my competitors. So I think it comes down to where I came from. And then when I got into business, being a problem solver and a data junkie, I was like, the data is telling me we could do better, why is this not better? I just get obsessed about the opportunity to improve it.

Liam:

What do you think most brands are getting wrong about retention marketing and customer loyalty today?

Jonathan:

I think there are three things brands are afraid of. The first is they hyperfocus on their brand. They want to be the Apple of their vertical, but they don’t have the working capital to be the Apple of their vertical, and it slows their decision-making down and halts progress. You can build a great brand without being the most pristine, the most perfect, the prettiest pixel. The second thing is not enough brands pay attention to analytics. They trust other people to give them their answers — ad firms, consultants, advisors. That’s all good, but trust but verify. Have either yourself or somebody really good at data verifying everything you’re doing and that it’s actually driving toward your strategy and vision. The third thing is mentorship. They need to go talk to people that have walked in these puddles. I’ve had a career coach since I was 26, a therapist since I was 25, five mentors since I was 23. They tell me when things stink and when things aren’t going well. When I came into the quilting industry, I had three mentors in the quilting space to bounce ideas off because that accelerates my ability to achieve.

Liam:

I think one of the biggest things is the lack of experimentation, particularly in the world of AI. The brands that truly won the buy now pay later wave were the ones that jumped on it so quickly. In today’s world, that’s AI. Tools out there can experiment so rapidly on what converts best, how to get a shopper to repeat purchase, whether they’re getting stuck in the funnel.

Jonathan:

Well, one thing I want to mention since you mentioned the funnel — I actually believe the funnel is dead. It’s now learning loops. I did a blog post about this on LinkedIn last week. It’s really about using AI for your learning loops, because that’s what’s going to help you be successful. Forget about the funnel — learning loops, that’s where the key is.

Liam:

With the learning loops replacing the customer journey, what parts should be AI-driven today if you went into any ecommerce brand?

Jonathan:

The entire thing — except when a customer has an issue, they need to talk to a human. You can automate the entire thing except that. And then depending on your sensitivity to brand, you might want to do your own creatives. I have no sensitivity because the creatives I create with AI are better than the creatives that were on Linda’s before I joined, and probably just as good as most creatives. Linda’s is an edgy brand — we’re not the polished crafts brand, we’re the edgy brand that’s not going to do everything perfect, but we’re going to give you a great deal, great service, and great product. We use AI for almost everything. I can’t use it for logistics fully yet, but for the most part, if you’re bold and you have courage, you can use AI for 90% of it.

Liam:

One area of significant adoption a lot of brands are looking into is personalization. Everyone’s been talking about it forever, but it’s never really been possible for an ecommerce manager to send a unique email or SMS to every shopper that visits their website.

Jonathan:

Or a unique notification. We have AI notifications on our site that target Barbara, that target Linda, that target Jane or Bob. We say, “Hey, we know you came to the site and didn’t purchase” — it’s a pop-up notification they accept when they log in, and they get AI-driven notifications. It caters the message based on where they are on our website and where they are in their general journey. Companies that are innovating in that space are going to be the multi-billion dollar brands in the commerce space, because that’s a really hard problem that’s not easy to solve if you haven’t crossed the AI chasm.

Liam:

A shopper lands on Linda’s today — they might be unknown. An identity platform would identify that shopper and remember them forever. They leave, add an item to their cart, they get an email that’s hyper-personalized to the items they viewed, with offers pulled from the website automatically adjusted for things like Black Friday, and sent at the perfect moment based on their shopping behavior.

Jonathan:

That’s part of it and it’s an important part. But I want you to take that expansion and take that data — that same thing happens on connected TV, that same thing happens back into paid media. That’s the power of the trifecta. The flows and SMS are great, but first-order profitability requires that enriched data to go back to the customer through those other mechanisms where they’re already consuming content.

Liam:

Capturing that data in the first place — what is your strategy at Linda’s? Do you use pop-ups, forms?

Jonathan:

We do it two main ways. One is we plug into different identity graphs — Instant has an identity graph of about a billion records I think. We also have pop-up forms, and we have sister affiliate sites. We pull from all of those sources, then the AI validates it, marries it up, and places it into our CRM. Then we leverage that identity data from the CRM to target the customer in a way that brings them value for things they actually want. We know our customers and their demography — we know they have spring projects, summer projects, winter projects. We try to help them rather than spam them.

Liam:

And you think that’s definitely driven revenue growth for the business?

Jonathan:

It’s the reason why the business has grown. The business would be a $3 million business if I hadn’t installed the trifecta and pulled in that data — that’s the signal-to-noise ratio. Now all those customers are first-order profitable, and once they get to 4.8 orders, they’re my customers for life.

Liam:

If you broke it down, what do you think TV is driving versus retention?

Jonathan:

If you take a million dollars, TV drives about 20% of it, paid media drives about 20% of it, and the rest comes from the middle — whether it’s flows or AI campaigns. And if you combine the AI flows and the AI campaigns we’ve done in the last seven months just with Instant alone, it’s $3 million. One fourth of our revenue, 25%, just from Instant. We’re also noticing a massive increase in CTV because our customers love to watch TV from 5 PM to 10 PM local time. We’ve been running an experiment and the AI — not me, but the AI — is actually moving money from Facebook and Google into CTV because it optimizes across the platform.

Liam:

These shifts are crazy. I don’t think people saw them coming as quickly as they have. To end on this note — what do you think will be the biggest shifts over the next 12 months?

Jonathan:

I think the biggest shift for Linda’s specifically is that AI will allow us to double our customer base from 400,000 to 800,000 in 12 months. We’ll probably be $18 to $24 million next year. The network effect is going to kick in. For AI more broadly, what you’re going to find is that small businesses are going to realize that if they want to stay alive and scale, they’ve got to learn — whether they do the learning themselves or bring somebody in who already has the knowledge. Otherwise the larger corporations are going to make it impossible for them to succeed. They will eat them. There’s a phrase in Korea — Ppalli Ppalli — which means “hurry hurry, we’re already behind.” Samsung uses it a lot. These businesses have got to learn. They’ve got to hurry because they’re already behind. And the equivalent in ecommerce today is that if AI isn’t driving 20% to 40% of your revenue, you’re already behind. If you don’t catch up in the next couple of months, it’s starting to become the world where it’s almost impossible to catch up, because your competitors are just going to eat you alive.

Liam:

I think it’s pretty incredible watching the journey of Linda’s. It’s great to be working together. It’s crazy to think that two years ago you were generating $2 million a year, and in the last seven months you’ve generated $3 million just from one platform using AI.

Jonathan:

It’s been an amazing journey. I’m ever learning. I’ve got an architect roadmap of all the AI ideas I’m thinking about. And because Linda’s was so successful with the IP we created, I started thinking — I wonder if other brands are having this problem. I did a test last summer with four brands — I said, hey, you don’t have to pay me anything, just use this framework I put together. The best performing brand grew from $500 a month to $30,000. The worst performing one grew 30% over the summer. And I was like, okay, there’s a 10x value solution here. Let’s take it to market. We launched in November, we have about 50 customers now. It’s really about how do you create that value leveraging AI and helping small businesses — whether it’s a quilt shop, an electronics shop, or an ecommerce company. If you seek AI first and pay attention to data, you’re going to succeed.

They dive into:

Why “boring” industries are some of the biggest opportunities today

How AI is changing ecommerce, retention, and personalization

The shift from funnels to “learning loops”

Why most brands are already behind on AI adoption

The exact playbook Jonathan used to scale Linda’s

00:00 Intro + Jonathan’s Background

02:09 Hypergrowth at Postmates

04:10 Acquiring Linda’s + Starting at $2M

05:09 Finding a $6B Opportunity in a “Boring” Market

09:44 Why “Sleepy” Verticals Are Underrated

12:03 AI Mindset Shift 

16:32 How Brands Should Use AI 

20:40 How Linda’s Uses AI for Retention

23:04 Why Experimentation Wins (and Most Brands Don’t Do It)

27:36 The Funnel Is Dead (Learning Loops)

29:22 Personalization at Scale with AI

33:03 Why AI Is the Reason for Linda’s Growth

35:22 What Happens Next (AI + Doubling Growth)