Jason Bertrand scaled Purple's e-commerce operation past $450 million by betting on simplicity over technology complexity. He breaks down the headless commerce tradeoffs, how the brand holds premium pricing in a discount-heavy category, and the AI-driven SEO playbook built for how people actually search now — in full sentences, not keywords.
Behind the Expert
Jason Bertrand is VP of E-Commerce at Purple, the mattress and sleep-products brand known for its hyper-elastic polymer "gel flex grid." He built his career almost entirely on the brand side of direct-to-consumer, starting in paid-search training and agency work at Advertising.com in Baltimore before moving in-house to Under Armour in 2007, when the company was a $300 million business — he stayed for five years as it grew toward $2 billion, watching its e-commerce channel scale from roughly $20 million to $150 million in revenue. From there he moved through other consumer brands, including Skullcandy, before joining Purple just over a year before this conversation, where he now leads a roughly 20-person team spanning UX, analytics, merchandising, Amazon, CRM, and SEO.
The Quick Hits
- Bertrand treats a brand's own site less as a transaction engine and more as its best-controlled storytelling channel — a place that has to deliver the most relevant content and best research experience regardless of whether a visitor ultimately buys there, in wholesale, or in a store.
- Purple deliberately limits its own promotional flexibility to protect brand equity: any product also sold at wholesale has to hold price and promotion parity across the entire year, communicated 90 days in advance, while web-exclusive products — like the lower-priced Purple Flex — are the one place the team can experiment with offers.
- Bertrand is candid that constant discounting in the mattress category has conditioned customers to believe paying full price makes you "kind of a fool" — and that Purple's wholesale-driven pricing discipline, while limiting in some ways, protects the brand from that same trap.
- He's skeptical of chasing "the latest and greatest" technology by default, citing Purple's post-COVID move to headless commerce as an example that added tech debt the team is still unwinding, given the brand's genuinely simple SKU catalog — a handful of mattresses, pillows, and sheets — never needed that complexity.
- Purple's current AI use is deliberately narrow and internally focused: a new tool ingests customer service chats, emails, and calls to surface recurring customer concerns in a recurring "voice of the customer" meeting, with plans to extend the same approach to reviews and Amazon data — customer-facing AI use remains limited, mostly to a fairly analog mattress-recommendation quiz that still captures first-party data through email opt-in.
- Bertrand's SEO team shifted focus from keyword-based optimization toward natural-language, question-based content — the kind of full-sentence queries people now type into ChatGPT or Perplexity — and reports "massive growth" in referral traffic from those platforms as a direct result of that shift over the prior six months.
- On first-party data and AI, Bertrand admitted candidly he doesn't have a fully worked-out answer for how to build guardrails that let Purple use AI's compute power without exposing customer data — framing AI today as a good starting point for ideas, not yet something to be "fully trusted."
From Under Armour's early digital team to Purple
Bertrand's path into e-commerce ran through paid-search training at Advertising.com in Baltimore, working just upstairs from a then-small, fast-growing Under Armour. He made the jump to the brand side in 2007, when Under Armour was a $300 million company with a roughly $20 million online business.
From Under Armour, Bertrand moved through other consumer brands, including Skullcandy, before joining Purple just over a year before this conversation, where he now leads a roughly 20-person team spanning UX, analytics, merchandising, Amazon, CRM, and SEO.
Your own site is the one channel you fully control
Asked whether the priority at Under Armour was revenue growth, profitability, or brand-building, Bertrand said it was always some combination — but that the site's role as the brand's best-controlled storytelling channel became clear as the company scaled.
Guardrails, not blanket discounting: how Purple protects brand equity across channels
Purple sells across three channels — e-commerce, roughly 55 owned retail stores, and thousands of wholesale doors — and Bertrand described a deliberate structure for managing the natural conflict between them.
Bertrand was direct about what constant discounting does to a category — and to the team running it.
Why Purple pulled back from headless commerce
Bertrand's stated philosophy on technology is to keep the stack as simple as the business genuinely requires — and he pointed to Purple's own post-COVID move to headless commerce as a case where that discipline slipped.
Repositioning around what the brand actually stands for
Purple's own consumer research surfaced a gap between recognition and positioning that Bertrand said reshaped how the team thinks about both data collection and brand messaging.
That repositioning is tied directly to obsessively understanding the customer — particularly given how infrequent the purchase cycle is.
Where AI actually lives today: internal, not customer-facing
Purple's AI rollout starts narrow and internal — mining customer service interactions for recurring themes, feeding a regular cross-functional meeting.
On the customer-facing side, Purple's real first-party data engine is still a fairly analog mattress-recommendation quiz — one Bertrand sees real potential to enhance with AI, but hasn't yet.
SEO for a world where people ask, not type keywords
Bertrand described a deliberate shift in his SEO team's focus, from keyword optimization toward the kind of natural-language questions people now put directly to AI tools.
AI's real constraint isn't capability — it's trust and data guardrails
Pressed on how Purple thinks about guardrails for its first-party data as it leans further into AI, Bertrand gave an unusually candid answer for a podcast interview.
Sound Bites
- “UnderArmour.com was really the store. They had a wholesale business, visibility from Cabela's to Bass Pro to Dick's Sporting Goods — but UnderArmour.com was the flagship store. It's a brand awareness channel, a place where people go to do their research whether they buy online or not.”
- “For any product also sold at wholesale, we have to maintain parity across pricing and promotion for the entire year, and communicate a promotional plan 90 days in advance. Brands like Nectar or Casper, without as much wholesale distribution, can pull levers all day long. But we have a web-exclusive product, the Purple Flex — that's where we can do different offers.”
- “In the mattress industry, customers are very conditioned that if you buy a mattress at full price, you're kind of a fool, because the industry is constantly discounting. We don't like to play the discounting game constantly — it's a path to failure.”
- “Coming out of COVID and the growth Purple experienced, it was, we've got to get on the headless bandwagon. It overcomplicated our tech stack — there's a lot of tech debt we're still unraveling. We don't have a lot of SKUs — a handful of mattresses, some pillows, some sheets. Making it too complicated just to jump on the latest and greatest is not the move I'd make.”
- “We're just starting to scratch the surface. We're implementing a tool that tracks our chats, emails, and customer service phone calls. We have a voice-of-the-customer meeting regularly, and the whole point is understanding what customers are saying. We also sell $30 to $40 million a year on Amazon, so there's a lot of data there that can help inform those decisions too.”
- “Traditional SEO is very keyword-focused — maybe some long tail, but it's not natural language. What we really need to focus on is how do we optimize the site from an AI or natural-language perspective. We've made some phenomenal advancements in just the last six months on this, and we're seeing massive growth in traffic that comes from Perplexity, ChatGPT, and so on.”
- “I don't feel that AI is at the point where it can be completely trusted. I think of it as a starting point to get thoughts or ideas across, but it's not fully trustworthy yet. I don't have a great answer for this.”
The Chord take
Purple's wholesale-parity pricing discipline, its retreat from an over-engineered headless stack, and its narrowly-scoped, internally-first AI rollout are three versions of the same instinct: match the complexity of your infrastructure, promotions, and technology bets to what the business genuinely requires, rather than to what a louder competitor or a hyped trend suggests you need. A brand with a handful of SKUs doesn't need a headless commerce platform built for thousands; a brand whose biggest promotional risk is degrading trust in a discount-saturated category doesn't need every pricing lever available everywhere — it needs one well-chosen web-exclusive product and firm parity rules everywhere else. That same restraint shows up in Bertrand's answer about first-party data and AI guardrails — and it's more valuable for being unresolved rather than confidently wrong. Plenty of leaders will claim they've solved the problem of using AI on customer data safely; Bertrand's willingness to say "I don't have a great answer for this" is itself the discipline this episode is really about: don't let the pressure to move fast on AI talk you into pretending a hard, unsolved problem is simple.
Put it to work
- 1Before adding promotional flexibility across every channel, decide which products, if any, can be genuinely web-exclusive, and hold pricing and promotion parity everywhere else to protect brand equity in a heavily discounted category.
- 2Match tech stack complexity to actual SKU and business complexity. Audit whether infrastructure decisions — like a headless commerce migration — were made to solve a real problem, or to keep up with a trend.
- 3If your brand has strong recognition but a fuzzy point of view, invest in consumer insight work specifically to find that gap before repositioning. Don't assume awareness equals understanding.
- 4Start any AI rollout with a narrow, internally-focused use case — like mining customer service transcripts for recurring concerns — before extending it to customer-facing experiences, and make sure that insight actually reaches product development, not just support.
- 5Shift SEO content strategy toward natural-language, question-based optimization for how people phrase queries to AI answer engines, while continuing traditional keyword optimization in parallel — treat this as additive, not a replacement.
- 6Before feeding customer data into any AI tool or model, be honest about what guardrails you actually have in place. "I don't have a great answer for this yet" is a safer starting point than assuming it's fine.
Full transcriptShow ↓
Jason Bertrand
If you buy a mattress at full price, you're kind of a fool, right? Because the industry is constantly discounting.
Bryan Mahoney
Did you do anything extra special online?
Jason Bertrand
Yeah — you can get into a bad place if you go down the path of constant discounting and offers, because it'll degrade your brand.
Bryan Mahoney
We have outlawed our team from using AI because it doesn't understand our brand voice — and I was like, well, you might be missing an opportunity to train the model to create those guardrails.
Jason Bertrand
What we really need to focus on is how do we optimize the site for AI — what are the kinds of questions people are most likely to ask.
Bryan Mahoney
Welcome back, everyone. Thanks for tuning in to the Brilliant Commerce Podcast, where I get a chance to sit down with some of the brightest minds in commerce and unpack the secrets behind some of the industry's iconic, or as I like to say, soon-to-be-iconic brands. This morning I'm thrilled to be joined by Jason Bertrand, VP of E-Commerce at Purple, the most comfortable mattress of all time. Jason, welcome to the pod.
Jason Bertrand
Yeah, thanks.
Bryan Mahoney
We covered this morning — we're both up a little early to be recording this — where am I catching you today?
Jason Bertrand
I'm in Park City, Utah.
Bryan Mahoney
Cool, how long have you been there for?
Jason Bertrand
About 14 years — moved here from out east, Baltimore area.
Bryan Mahoney
Okay, and during the warm-up we covered that we were both in Montreal for some time — so you've been out east 14 years, is that right?
Jason Bertrand
No, pretty much my whole life, really — I was born in Montreal but grew up in the States, went to college in Halifax, Nova Scotia.
Bryan Mahoney
So—
Jason Bertrand
But career-wise, it's been all in the U.S.
Bryan Mahoney
All in the U.S. What did you study at McGill?
Jason Bertrand
Well, Dalhousie, actually.
Bryan Mahoney
Yeah, I screwed that up — born in Montreal, went to Dalhousie. What was your area of focus in university?
Jason Bertrand
I was part of the first year at Dalhousie that had, in the Bachelor of Commerce program, a co-op program — that was one of the reasons I decided to go there. It was three internships throughout my college experience, so it was pretty cool, and I did all three of those in the U.S., which kind of kicked off my career that way.
Bryan Mahoney
That's awesome — I did a BCom at McGill, and I think that's why I screwed that up. We didn't have that program, and there was no notion of anything like e-commerce related at all — I did finance and entrepreneurship. Looking back at how those programs have evolved, it's pretty wild. But the internship program sounds fascinating — what were some of the companies you interned at?
Jason Bertrand
At the time I wanted to be in hotel management, so I actually worked mostly at hotels for my internships. But when I graduated, I ended up getting into software training, and that got me into Advertising.com in Baltimore, where I was training agencies on how to use very early bidding software for paid search — this is back in the Yahoo days, way back. That's basically how I got into digital marketing. Literally right upstairs from Advertising.com was the headquarters for Under Armour, when they were a little-known brand growing pretty fast, and that's when I made the shift from agency side to client side.
Bryan Mahoney
Okay, so how long did you spend on the agency side before making that switch?
Jason Bertrand
Including the software training and the Advertising.com piece, probably about five years.
Bryan Mahoney
Okay.
Jason Bertrand
And then shifted to Under Armour, and I've been with D2C brands since then — that was 2007 when I started with Under Armour.
Bryan Mahoney
2007 — how important was online to Under Armour at the time?
Jason Bertrand
It's interesting — when I started with Under Armour, it was a $300 million business. When I left, it was close to $2 billion. When I started, the online business was only $20 million, so a pretty small chunk. It went through a few redesigns and replatforms, and by the time I left it was about $150 million — pretty rapid growth, right? It was such a great brand to be with, and it went from being sort of known to being a household name by the time I was done, so it became much more important over those five years.
Jason Bertrand
When I first started it was just a small channel, but as we grew it became critical — started with maybe 20 people, by the time I left it was double that, maybe 40, 50. Now it's probably 150 people running e-comm. Pretty rapid growth for that brand and that channel especially.
Bryan Mahoney
Yeah, I think that's interesting — like you, I've spent a long time in commerce, I wrote my first e-commerce application in '97, still in university. I remember that first wave of direct-to-consumer e-commerce — very much how do we disintermediate the middleman, get a product into customers' hands faster so we have more margin. Where I really started to get excited about D2C is when it became more about creating that connection with the customer — how you have a brand like Under Armour go from little-known to a household name.
Bryan Mahoney
From your point of view at Under Armour, starting with a small portion of the business online and growing it meaningfully — was the goal to grow top-line revenue, achieve a more profitable sale, or really become that household name and expand the brand? What were some of the things you and the team did to achieve that growth in a pretty short amount of time?
Jason Bertrand
Honestly, Bryan, I think there's a little bit of all those things. When I was there, revenue was definitely the first thing, but as the brand evolved it became clear that UnderArmour.com — and eventually, while I was there, they purchased UA.com — was really the store. They were opening stores left and right, they had a wholesale business, visibility from Cabela's to Bass Pro to Dick's Sporting Goods, hundreds of thousands of wholesale partners — but UnderArmour.com was the flagship store. So it really has to be a combination of revenue growth, because of your point about profitability, higher profitability within a DTC business.
Jason Bertrand
But it's also a brand awareness channel — and this isn't just true for Under Armour, it's true for Purple especially, and Skullcandy, other brands I've been part of — it's where people go to do their research, whether they buy online or not. It has to be the best and most relevant content, really the best storytelling. So it was that balance between driving revenue profitably and being the best showcase of the brand from a content perspective.
Bryan Mahoney
I love storytelling — it's so important to hit on, it has to be your best channel, where the brand shows up in a way you can control. All the respect in the world for the team at Dick's Sporting Goods, but they're going to talk about Under Armour differently than you get to talk about it online with your customer. What advice would you give brands afraid of selling directly to their customers, where the term cannibalization gets thrown around all the time? How do you think about blending that, using that channel to be that brand ambassador and tell that best story?
Jason Bertrand
I think it really is alignment across the organization on the guardrails for who the brand is, and who the customer is — setting those guardrails so that regardless of what channel we're speaking to the customer through, whether it's the website, email, social, even stores or wholesale, there has to be some level of guardrails that keeps the consistency of the brand in place. There's going to be different creative that works differently across channels, but there have to be guardrails so customers understand who the brand is, and you're marketing generally in the same way, keeping it consistent.
Jason Bertrand
That alignment is critical — gaining that, call it a brand bible, so people are following those guidelines to some extent. There are going to be splits between channels, of course, in different creative, but it's really that consistency in the guardrails that helps maintain brand affinity.
Bryan Mahoney
Did the team do anything extra special online — product drops, exclusive offerings, either at Under Armour, Skullcandy, or now at Purple — is there anything that prioritizes traffic on the digital presence over what might be available in wholesale or owned and operated?
Jason Bertrand
I think there's natural channel conflict that comes from having multiple channels. I'll give you the example with Purple — we talk about three different channels, the e-comm channel, the owned retail channel, where we have about 55 Purple stores around the country, and wholesale, where we have thousands of doors. There's natural channel conflict from that. I love a web exclusive — absolutely love one. It's great to have a product you can market that doesn't fall within the boundaries created by that channel-conflict scenario. We have a lower-price-point mattress called the Purple Flex.
Jason Bertrand
It's about a thousand-dollar queen — we always talk about the queen being the middle of the road, the most commonly purchased. That product isn't sold at wholesale, so we can play around with it, do different offers. But for any product also sold at wholesale, we have to maintain, call it parity, across pricing and promotion for the entire year. Brands like Nectar or Casper, without as much wholesale distribution, can pull levers all day long, run different promotions.
Jason Bertrand
We have to communicate a promotional plan 90 days in advance, we can't just go run a crazy promotion. But we have that web exclusive, where we can do things a little differently. You can get into a bad place if you go down the path of constant discounting and offers, though, because it'll degrade your brand. So we're balancing how we think about promotions with our wholesale partners to maintain that parity, but we're also trying to be a premium brand.
Jason Bertrand
And as a premium brand, you can't go crazy with discounting like a lot of mattress brands do. It's a tricky balance, but I think it's one we're doing well with this year compared to years coming out of COVID.
Bryan Mahoney
That makes me think about a lot of experiences I've had where it's difficult when you decide to go down that discounting path — it's like Pandora's box, really hard to put back in. I think as an e-commerce industry, we've conditioned the customer that their first purchase is going to be discounted. I used to see this all the time — we'd offer that first-purchase discount, 10% off, sometimes more, especially on a product that was more evergreen or repeat — you're not going to buy a mattress every other week.
Bryan Mahoney
I'm sure you wish that happened, right? But I'd see it in other categories, the lengths customers would go to pretend they were a first-time customer just to get that discount — you can see it in searches, what's the first-time discount code for such and such brand. It became this nightmare on the back end trying to unify that data — not because I wanted to unify the customer profile for its own sake, but because we wanted to be better marketers, more efficient, and get to know the customer better. Instead we have the same customer five times because we've conditioned them to go find that first-time discount. It's problematic.
Bryan Mahoney
That balance between what you do to acquire a customer and what you do to keep up with the industry is really difficult. What you said about not having as many levers, those constraints — I actually look at that as a great way to deliver a more consistent customer experience. No one likes to buy something and then figure out, well, if I'd bought it on that channel, I could have gotten this discount and that discount — that's exhausting. So on one hand, I guess the constraint is, I wish we could pull all these different levers.
Bryan Mahoney
But on the other, I think it's nice to sit down and think more holistically about the strategy across all your channels, so you can meet your customer where they are — they know wherever they buy that Purple mattress, they're getting the best experience and the best price.
Jason Bertrand
Yeah, and in the mattress industry we always talk about how customers are, to your point, very conditioned that if you buy a mattress at full price, you're kind of a fool, because the industry is constantly discounting. The other point I'd mention is, pulling these offer levers, these discounts all the time, really isn't that much fun.
Bryan Mahoney
Yeah.
Jason Bertrand
Selfishly, I'd much rather build a brand and think about the content and storytelling than constantly think of new promotions to run — it's just not as fun. Being part of a premium brand, where it's not all about discounting, is so much more fun, and leaning into that storytelling is a more fun experience for me, and I know I can speak for my team too — we don't like to play the discounting game constantly. It's a path to failure.
Bryan Mahoney
It absolutely is. What I used to say is, anyone can sell $10 for five — what can you do as a great marketer and storyteller to get someone in the door, get them interested, educate them? A mattress purchase is a serious purchase, a real investment — you talked about that entry product still being a thousand-dollar purchase, something I'm going to spend significant time on. It's not this impulse purchase where I load up Amazon and think, I need a container for something in my kitchen, I'm not even going to think twice, whatever the algorithm tells me to buy, that's what I'll buy.
Bryan Mahoney
If I'm shopping for a mattress, I've got five browser tabs open, being really thoughtful about it, maybe even going in store because I need to lie down on it, with my partner too, since this is a decision we're making together. I've always lost those battles — like, I really like this mattress, but that's not the one we're going to do.
Jason Bertrand
That's not worth fighting over.
Bryan Mahoney
No, not worth fighting over — though sleep is probably worth fighting over, I'd be curious how much that makes its way into—
Jason Bertrand
You spend a third of your life in bed, so it's worthy of making a decent investment.
Bryan Mahoney
Yeah.
Jason Bertrand
At least that's the argument we try to make — having slept on regular mattresses and now a Purple mattress, I definitely see the difference. I'm biased, obviously, but I think it's worth putting some good money into your sleep.
Bryan Mahoney
I think it's great that you're biased — I've always loved, wherever I've worked in-house at a brand, or if I'm a brand choosing to work with Chord, I'm so pro them, very biased, I love their story, I want to get to know their products and their brand. You should be biased, proud of where you work — that's going to come through in how you tell those stories and lead your teams. So tell me in two minutes, what makes the Purple mattress the most comfortable mattress of all time?
Jason Bertrand
It's an interesting origin story — the founders started in the medical field, building this — we call it our 'squishy,' a small version of what's inside the mattress. It's hyper-elastic polymer, scientific name, we call it the gel flex grid. When you look back at mattress innovation, the biggest innovation in our lifetime was memory foam — but memory foam traps heat, doesn't let you sleep cool, traps a lot of stuff you don't want. This, you can see through it.
Jason Bertrand
It allows you to sleep cool, and what we always say is it's soft where you want it and firm where you need it — your hip, your shoulder will go a little deeper into the grid, but stay soft on top. Most people think of mattresses as, I either want firm or soft — and this, admittedly a difficult story to tell, is both. There are different varieties of foam level and grid thickness that we sell. Having moved to a Purple mattress myself, it was a game changer.
Jason Bertrand
My wife has some health issues, and we joke that we just don't want to get out of bed because it's so comfortable. So, I'm sold, and to your point, it's fun working for a brand with a great brand, a great product, and a great team — I always say if you can check those three boxes, that's the passion trifecta. That's where I've found my passion with different brands I've worked with, and why I like working with great brands.
Bryan Mahoney
I love that — we spend a third of our life in bed, probably another third working, so it's important to work with a team at a company you can be passionate about. Let's use that as a segue — tell me about the team today. You've been at Purple a little over a year?
Jason Bertrand
Actually a year yesterday — I just did my one year yesterday.
Bryan Mahoney
Congrats.
Jason Bertrand
Thank you.
Bryan Mahoney
So what does your team look like today?
Jason Bertrand
We have about 20 of us — a combination of UI, UX, analytics, merchandising, an Amazon team, a CRM team, and SEO also reports to me. So about 20, and a great team, super high-performing — I think I can speak for them in saying they're all pretty passionate about the brand and the product, and it's just fun.
Bryan Mahoney
Cool. I want to take a trip down memory lane, as it relates to team or tech stack — thinking back to 2007, what did the team look like, what were the tools like, and how has that evolved over your career? Has it changed significantly in the year you've been at Purple, in terms of how you've organized the team, the tools, the tech stack? I mean, that's a long career.
Jason Bertrand
I guess I'd say it the same way regardless of the brand — this one's a $450, $500 million brand, I've consulted with brands that were $30 million. I think the tech stack needs to be as simple as it possibly can be. I've seen an interest from a lot of people in my past to just grab the shiniest, newest, greatest thing — I'll use headless commerce as an example. Coming out of COVID and the growth Purple experienced, it was, oh, we've got to get on the headless bandwagon.
Jason Bertrand
And honestly, I think it overcomplicated our tech stack — there's a lot of tech debt we're still unraveling as a result, because it wasn't simplified. We don't have a lot of SKUs — a handful of mattresses, some pillows, some sheets, nothing like Under Armour or Skullcandy with the number of headphones and earbuds we were dealing with. It's a fairly limited SKU stack, so making it too complicated just to jump on the latest and greatest is not the move I'd make. I've been through several replatforms and redesigns, and I'd say the same thing.
Jason Bertrand
Obviously it needs to be scalable, because everyone's interested in growing, but you just need to be realistic and keep it as simple as possible. I know that's a pretty general statement, but that's been my philosophy — don't overcomplicate it, or you'll run into tech debt and regrets.
Bryan Mahoney
I think it's easy to chase that shiny object — I have an apology to make here, I was one of the people who perpetuated the headless story, had a headless commerce stack for a long time. But unless you have an engineering team, you're absolutely right, it's overcomplicated. There was a time and place for it, when you couldn't make certain front-end changes or performance was a primary concern — but that's largely a solved problem now. So I'm with you on simplification. If you don't have the experience, there's this fear of missing out — everyone's using that tool, I need to use that tool.
Bryan Mahoney
And before you know it, you've got 20 different tools largely doing the same thing in your stack, and you ask yourself, this is inefficient, expensive, you don't have a single source of truth, and you end up spending more time trying to understand why the numbers are different across platforms than actually looking at numbers and making decisions.
Jason Bertrand
So do you apply that same—
Bryan Mahoney
Yeah, exactly — just curious if you apply that same simple-is-better philosophy to how you think about data collection and data usage at Purple, and places you've been before.
Jason Bertrand
Yeah, absolutely. You can easily run down the path of collecting everything, but if you're only actioning on a portion of it, you're wasting your time. It's really focusing — right now with Purple, we're very obsessed with figuring out who our customer is.
Bryan Mahoney
Okay.
Jason Bertrand
And, like I said earlier, putting guardrails around the brand from a repositioning perspective. We found through consumer insights that we've got phenomenal brand recognition — almost everyone knows who Purple is — but they don't really know what we stand for. They know we're an innovative mattress company, but not what we stand for. You'll see the treatment we recently implemented on our logo: less pain, better sleep. That's meant to show, in a very quick and dirty way, this is what we stand for — less pain and better sleep. Very simple. Part of that concept of putting guardrails around the brand and knowing who we are.
Jason Bertrand
That very closely relates to knowing who our customer is, obsessing over that, dialing it in. You made the point earlier that you're not making a mattress purchase every couple of weeks, you're making it every eight to ten years. So we have a lot of new customer acquisition and not a ton of retention — we retarget people to buy a pillow when they've bought a mattress and that kind of thing, but for the most part it's new customer acquisition, so we have to be very clear and consistent in our messaging about what we stand for.
Jason Bertrand
That makes a big difference as we're going through this repositioning effort — not just the logo treatment, but across all our content, targeting, and audience collection.
Bryan Mahoney
How important is that — you work super hard to tell that brand story, acquire that customer, it's a pretty big first purchase I'd imagine — reviews and referrals must be really important.
Jason Bertrand
Yeah, reviews for sure. I think sentiment — and I was hoping we could get into some AI conversation, it's hard not to — but from a big-data perspective, we're really understanding what our customer sentiment is, and using AI to sort through that and understand what people aren't clear about from a mattress-differentiation standpoint, what they're asking most regularly, and how we can update the website to reflect that.
Jason Bertrand
Even as far as product development — how can we align with the product team to literally design products that reflect what customers are interested in, what they're asking about. So I guess I'm derailing your question a little, but from a big-data perspective, it's really about focusing on the right stuff, not going overboard on too much data. The most important thing is knowing your customer.
Bryan Mahoney
I love that. Let's jump into AI — that's all I've been working on with my team the last couple months, and we're excited about what we're doing with Chord AI. Being as hands-on as I've been has made me realize how much work we need to do with the data to extract any meaningful value from AI. Two things you said are worth unpacking — first, customer sentiment, understanding that feedback. I think that's a wonderful use case for AI, or natural language processing, or even machine learning.
Bryan Mahoney
Your team is doing that today, if I understand correctly — using AI to find those nuggets within the reviews data across your channels, is that fair to say?
Jason Bertrand
Admittedly, we're just starting to scratch the surface. We're implementing a tool that tracks our chats, emails, and customer service phone calls — right now it's very customer-service-focused, but we'll eventually bring in reviews and everything. What it's doing is letting us say, what are the most common concerns customers have? We're on a call this week, a voice-of-the-customer meeting we have regularly — the whole point is to understand what customers are saying.
Bryan Mahoney
Right, yeah.
Jason Bertrand
The point I've been trying to drive home with the team is this is applicable not just to customer service, it's applicable to everything — how we change our product detail page to reflect what customers need to see, product development should be very clear and aware of what customers don't like. To some extent they are, through reviews, but that's just the tip of the iceberg. With the benefits AI brings, we can really focus the conversation around very clear, consistent subjects, and even layer in our Amazon data — we sell $30, $40 million a year on Amazon.
Jason Bertrand
So there's a lot of data there that can help inform those AI decisions and the data we're trying to mine.
Bryan Mahoney
I love that, and I think what you're hinting at is there's an underpinning — all of this needs to be a unification strategy. There's a lot of platforms layering AI on top of their offering, but it becomes somewhat siloed. If you're only using AI on your reviews, you're missing the opportunity to use that same AI on top of your Amazon data, or other data you have. You want to be able to not just summarize and find insights, but monitor those over time — can I deploy an agent that, after I've asked a question and gotten the answer, monitors for changes, like sentiment shifts.
Bryan Mahoney
I think that's going to require more work than most brands realize — investing in their own data and unification strategy, so they can deploy AI holistically, instead of having a number of agents running around in different platforms that can't talk to each other. Is that how you're thinking about it at Purple?
Jason Bertrand
Oh yeah, definitely. Like I said, we're scratching the surface, but the unification of data is exactly it — this is arguably the starting point. If I were to say what AI is going to do to the e-commerce and retail business, I think people are going to start — and already are, though it's still early — asking AI, what's the best mattress for me. We need to be ahead of that trend, I don't want to jump on that bandwagon when it's too late.
Jason Bertrand
I want to know we're ahead of it, implementing things that let customers find the right answers to their problems, their pain resolution, and show them the correct product for their specific situation. It's so personalized — that's the benefit of AI, it's not just a grouping of segments, it's extremely personalized. We can say, based on this very specific question, for this very specific pain issue, here's the absolute best mattress for you.
Bryan Mahoney
Yeah.
Jason Bertrand
I like the combination of AI and personalization — I think that's a great marriage that's just going to continue to blow up.
Bryan Mahoney
This is amazing, and you sort of read my mind on the next question. I'll meet with brands whose teams are worried about AI rather than excited about it. We talked about reviews, and about the use case of, I might ask ChatGPT what's the best mattress for me, and it'll go do a lot of research. So you have a team thinking about SEO — I'd love to hear how that team is thinking about the SEO landscape evolving, not how you're using AI, but how you anticipate customers using AI to find Purple. Tip of the iceberg, we're just getting started, and most of us don't know.
Bryan Mahoney
Is your team excited about those challenges, and what have you seen so far that's been useful?
Jason Bertrand
Very excited. Traditional SEO is very keyword-focused — maybe some long tail, a few keywords, but it's just not natural language. When you look at snippets, or what Google's doing with regular search, showing these and letting people ask questions, we're shifting the mentality — of course we're still doing SEO optimization, that's not dead in any way — but what we really need to focus on is how do we optimize the site from an AI or natural-language perspective. We're focusing on the kinds of questions people are most likely to ask.
Jason Bertrand
My recommendation for brands curious about this is, shift your mentality from traditional SEO — I don't know if it's on the decline so much as it's on the incline, there are people asking, what's the best pillow for me based on this, that, and the other, and we need to answer that question and optimize the site for those types of questions, which is what we're doing. We had a great new hire earlier this year who's very in tune with this trend, and we've made phenomenal advancements just in the last six months on this topic.
Jason Bertrand
We're seeing massive growth in traffic that comes from Perplexity, ChatGPT, and so on, because of the optimizations we've made. So it's SEO — you've got to optimize for keywords still, that's not dead — but start thinking about the questions, the longer-tail keywords, I guess you could call it, but more natural language is how I'd describe it.
Bryan Mahoney
It's still storytelling, actually — how can your story be the most compelling, how can the reviews on your site be the most relevant, whether it's a human reading it or a large language model trying to act in the best interest of the person asking the question behind it. We rushed to this idea that shopping's dead, you're just going to have agents doing all the shopping for you — I'm like, hold on, I really like to shop. I might use an agent to whittle down options, but there's something—
Bryan Mahoney
Shopping's been around as long as I've known, and I still love walking into stores, feeling the product, or going online reading reviews and being part of it. I don't think that's going anywhere anytime soon. I'm happy to have an agent shop for my toilet paper, but that's about it — I'm a consumer, I love getting to know these brands.
Jason Bertrand
I love what Amazon is doing — instead of siphoning through dozens or hundreds of reviews, you just read the summary, they've had that for several months on Amazon. I love that, and I'd love to implement it on Purple.com — it's a great way to save time but still take advantage of all those reviews you have.
Bryan Mahoney
Yeah, the trust factor for me there is I'm always wondering about the incentive alignment for some of those reviews on Amazon — and trust is an interesting segue to continue the AI conversation. Trust is really important, especially if we're using AI to help answer questions internally about our own businesses, and there's also trust in your customers giving you a lot of first-party data. Maybe to end on this — you've said guardrails a lot throughout — how are you thinking about guardrails to protect that first-party data when leveraging AI?
Bryan Mahoney
How much of this data are you willing to not give away — not necessarily give away — but how are you coaching your team to make sure you're treating that data the right way while still taking advantage of what's made possible now by AI, specifically gen AI?
Jason Bertrand
That's a great question, Bryan — I don't have a great answer. I think you're right that AI isn't at the point where it can be completely trusted. I think of it as a starting point at this stage in AI's evolution — a good starting point to get thoughts or ideas across, but not fully trustworthy yet. Maybe this is a question we can skip, honestly, because I don't think I have a great answer for it.
Bryan Mahoney
No, it's tough for sure. The way I've looked at it — and maybe this isn't the right way — it feels a bit like what happened maybe ten years ago when DTC was becoming popular. A lot of venture dollars had flowed in, it was easy to acquire customers, so brands were in a hurry to give away first-party data to marketplaces like Facebook and Google. In exchange for giving them our hard-earned first-party data, they'd give us customers back, but they kind of weaponized that data against us. Now it's super expensive to acquire customers that way.
Bryan Mahoney
I think there's a real risk that if brands leverage LLMs the wrong way — without the right infrastructure or data philosophy in place — it's, take all this data, I'll give it to you as context, tell me what to do with these customers. I'm worried the same thing might happen, that access to these LLMs becomes so expensive we can't actually leverage them that way. I'm sort of a first-party-data advocate for customers — I want brands I have a relationship with to treat my data the way I'd treat their data.
Bryan Mahoney
How can I leverage the enormous power of LLMs while keeping my customer data private? There's so much pressure to move fast and use AI to unlock efficiencies, but I think brands, especially well-established ones, have a responsibility to think thoughtfully about using that customer data the right way — so I've seen these trends, or this has been anonymized this way, or this is the metadata, and now I want to leverage the compute power you have to run a query on top of my data. But my customer data never leaves the four walls of my organization or my data warehouse.
Bryan Mahoney
That's maybe more where I was going — I talk to brands who say, we've outlawed our team from using AI because it doesn't understand our brand voice, and I think, well, you might be missing an opportunity to train the model to create those guardrails. Same thing with first-party data — what are the hard and fast rules we can put in place to treat customers the right way while still leveraging what's available to us in something changing so quickly. So that's where I wanted to go with it — I didn't mean it as a sneaky question.
Bryan Mahoney
It's so hard to answer, because it's tempting to just say, I'm going to use it everywhere — but is it always the right place to use it? That's what I was trying to get at.
Jason Bertrand
I'd say we're more using it internally right now — for our own benefit, to understand our customer and sentiment, that kind of thing. We're not quite using it for customer-facing purposes, I wouldn't say. One example on first-party data that's critical for us is our mattress quiz — asking people to take this quiz, and I think there are some really interesting AI applications of it that could take it to the next level. But right now it's fairly analog — do you sleep with a partner, what price range are you interested in, are you a side sleeper, a hot sleeper.
Jason Bertrand
And it siphons down to a recommendation of mattresses, and of course we ask for your email address. So we have this first-party data, and can market to people in a fairly personalized way — emails, SMS — trying to get them to make that final decision. So from a customer-facing standpoint, we're relying on that first-party data, and from an AI standpoint, it's more internal benefit.
Bryan Mahoney
For efficiency — how can we supercharge our data powers by going through this data quickly and finding interesting insights, as opposed to the traditional feedback loop. Back in 2007, you'd have an insight, go to the data team, they build dashboards, send them to you, you look at them, think, actually I've got a follow-up question, back to the data team, get the answer — it worked, but the feedback loops were long. Now you have agents relentlessly working the data, finding insights, sending them to you.
Bryan Mahoney
Those things are so exciting, and I appreciate the honesty of saying, here's how we're using it, here's how we're getting started, here's an opportunity to evolve our SEO strategy, we're experimenting and learning. That's my favorite part of these conversations — it's not magic, there's no one five years ahead of you who went from a 20-person team to a one-person team with an army of agents running around. We're not there yet. I deeply appreciate being able to talk with leaders like you who say, I've got a great team, they're super smart, super curious, not worried about it, excited about it, and here's some of how we're learning it.
Bryan Mahoney
I think you've shared a bunch of knowledge that anyone running an e-commerce team today can take a ton away from. I promised to get you out on time, so I want to honor that. Jason, thank you so much for the conversation — I'm rooting for team Purple, excited to go find a store and try one of these mattresses.
Jason Bertrand
Yeah, recommend it.
Bryan Mahoney
Awesome. All right, have a great weekend — thanks again.
Jason Bertrand
All right, thank you.