Levels built one of the lowest customer acquisition costs in health tech without a single hard sell. Ben Grynol explains how: building an entire category through original research instead of conversion copy, nurturing leads for 12+ months before ever asking for the sale, and writing niche health content that's now built to be found by AI agents, not just Google.
Behind the Expert
Ben Grynol is Head of Growth at Levels, the metabolic health company built around continuous glucose monitoring. Before Levels, he built companies across apparel and healthcare, including an early role at Canadian food-delivery marketplace SkipTheDishes, and hosts his own podcast. He grew up in Gladstone, Manitoba, and started his first business — odd jobs, yard work, painting, deck building — at 13, running variations of it for a decade before moving into building and scaling companies full-time.
The Quick Hits
- Levels was founded by Josh Clementi, an early SpaceX employee, after he traced his own unexplained energy crashes to glucose spikes from foods he'd assumed were healthy — quinoa, brown rice, sweet potatoes. Discovering continuous glucose monitors (CGMs) as a way to see food and exercise's effect on glucose in real time, he built a company to bring that access to more people, launching in June 2019.
- Levels' business model pairs hardware — a CGM sold through a third-party, prescription-based provider — with proprietary software that turns raw glucose data into personalized insight, meaning the commerce layer exists specifically to fund and distribute the software's real value.
- Grynol frames Levels' entire go-to-market strategy around one insight: trust, not price or performance ads, is the only thing that converts a customer to a product that's expensive, not well understood, and worn under the skin. The company has built a library well into the 600s or 700s of long-form, fact-checked, science-backed blog posts specifically to build that trust before ever asking for a sale.
- Levels' conversion timeline is unusually long and deliberately unhurried: roughly 20% of conversions land within 30 days, 40% within a longer window, and 60% happen more than a year after first contact — which is why the team refuses to judge partner or channel performance on last-click or bottom-of-funnel metrics alone.
- Levels runs almost no traditional performance marketing — minimal paid search, and throttled, retargeting-only Meta spend — because the product's low general awareness and high price point make bottom-funnel-optimized acquisition inefficient. Instead, spend concentrates on a small, curated set of high-conviction partners assessed on traffic and email capture, not immediate conversion.
- On measurement, Levels distrusts platform-reported attribution from Meta and Google entirely, relying instead on its own PostHog-based full-funnel analytics, UTM tracking, and an "area under the curve" approach that compares baseline web traffic to the one-to-two-day spike after a partner event, rather than trusting any platform's self-reported numbers.
- Levels' content strategy deliberately targets very long-tail health topics — like PCOS and ground flax — on the theory that being the only comprehensive, deeply researched source on a narrow topic makes a brand equally discoverable whether a reader is searching on Google or asking ChatGPT or Perplexity for a summary, while staying mindful that a meaningful share of its members — many women 45 to 65, some in their 70s and 80s — may not be using AI search tools yet at all.
A SpaceX engineer's own energy crashes, and the birth of a category
Levels co-founder Josh Clementi was working at SpaceX in 2017 when he began running into unexplained energy deficits — despite exercising regularly and eating what he thought were healthy foods. It took him years, and eventual access to a continuous glucose monitor, to understand why.
Levels launched in June 2019, into a market where the term "metabolic health" barely existed in public conversation. The company's education efforts, alongside a handful of other voices in the space, helped create the category that now shows up regularly in mainstream health media.
Selling hardware to fund software-driven insight
Levels' actual commerce layer is narrower than it might appear: the company doesn't manufacture the CGM hardware itself, but provides access to it through a third-party, prescription-based provider, and builds the software layer that turns raw glucose data into something usable.
Why trust, not price or ads, is the only lever that converts
Grynol laid out three specific headwinds Levels has to overcome with every prospective customer — and why none of them can be solved with a paid-media budget.
600-plus blog posts as the actual go-to-market strategy
That education had to meet a specific bar, in Grynol's telling: deeply researched, science-based, fact-checked to the standard of any serious editorial operation — and built as a library, not a sales funnel.
That approach changes what the customer is actually evaluating by the time they consider buying, Grynol argued — shifting the decision away from price sensitivity entirely.
Measuring a conversion window that stretches past a year
Levels' actual conversion data makes the case for its patience: attribution windows most brands treat as generous capture only a fraction of Levels' eventual conversions.
A curated handful of partners, almost no performance marketing
Rather than spreading spend across a wide affiliate network, Levels deliberately concentrates on a small number of partners it has high conviction in — chosen on traffic and email capture rather than clean, immediate attribution.
Grynol drew a direct contrast with categories where the product's utility is instantly obvious to explain why Levels' performance-marketing approach looks so different from a typical DTC brand's.
The discipline pays off in the numbers: Grynol cited a customer acquisition cost in the $50-to-$80 range against an LTV around $285 to $288 — an LTV-to-CAC ratio above 5, achieved by prioritizing capital efficiency over growth at all costs.
Distrusting the platforms' own numbers
Levels' measurement philosophy starts from the assumption that ad platforms' self-reported attribution can't be trusted, regardless of how the attribution window is configured.
For harder-to-attribute channels like partnerships, the team instead compares baseline web traffic to the spike that follows a specific partner event.
Writing for the long tail, not the algorithm
Asked how Levels thinks about AI's effect on organic discovery, Grynol argued that health and wellness content has a structural advantage most product categories don't: there's nearly no ceiling on how narrow a useful topic can get.
He was equally careful not to let AI-search hype override what Levels actually knows about its own customer base.
Personalizing lifecycle email around what customers actually engage with
Roughly 30% of Levels' conversions come through lifecycle marketing — almost entirely educational, rather than promotional, content — and Grynol described personalizing that content around a member's actual health data.
He also framed sustained engagement itself as the best defense against AI increasingly filtering what reaches an inbox at all.
The next commerce layer: prescriptive, concierge-style guidance
Asked to close with a prediction, Grynol pointed to a pattern he sees extending well beyond health and wellness: customers increasingly willing to pay simply to be told exactly what to do.
Sound Bites
- “He was eating food he thought was healthy — quinoa, brown rice, sweet potatoes, proteins — but it wasn't helping his glucose stay at equilibrium. He found a continuous glucose monitor — a little piece of hardware on the back of your arm that could monitor his glucose in real time.”
- “The body is this black box where people don't understand a ton about it, and there are all these tools — just like you need wrenches for plumbing or soldering irons for electrical — that do different things. On the commerce side, we're selling the tools; on the software side, we're able to provide people an insight based on their personalized data.”
- “We're well into the six, maybe seven hundreds of blog posts now. We wanted people to read a blog post and think, maybe I shouldn't grab the Coca-Cola, or maybe oatmeal on its own gives me a bigger glucose spike than I realized. Educating people to make different decisions is the starting point into a twelve-month-plus consideration cycle.”
- “Within 30 days we get roughly 20% of our conversions. Within [a longer window] we get roughly 40%, and we're looking at 60% of our conversions coming after a year.”
- “Why can True Classic scale up the way they have? The utility is understood — I see a T-shirt, I know what it does. For a product that's niche and longer-tail, it's very difficult to capture that customer for a reasonable CAC.”
- “The platform we use is PostHog — when we're measuring bottom-funnel attribution, we're only using PostHog. We don't take reported data on Meta or Google.”
- “If we're creating something as long-tail as PCOS and ground flax, the probability of someone else having that article on the internet is pretty low. The more of this long-tail content we have in aggregate, that's where we win.”
The Chord take
Levels' long consideration cycle, its refusal to trust platform-reported attribution, and its long-tail content strategy are three expressions of the same discipline: refusing to measure or optimize for the wrong time horizon. A product this unfamiliar and this expensive can't be sold on a seven-day click window, so Levels built its entire growth model — content, partner selection, even its own definition of a "good" CAC — around a genuinely multi-year relationship with the customer, funded by education rather than performance ads, and validated only against its own first-party data rather than any platform's self-reported numbers. That same patience shows up in how Grynol talks about AI. Rather than treating AI search as a reason to panic-rewrite the content strategy, Levels checked the premise against its actual member base — many of whom are older and may not be the ones asking ChatGPT for health advice yet — and concluded the long-tail research library it had already built was the right bet regardless of which interface eventually surfaces it. The company's own definition of success was never "win the SEO game this quarter"; it was "become the most trusted source on the internet for this category," a bet that pays off the same way whether the reader arrives via Google, Perplexity, or a friend's recommendation a year later.
Put it to work
- 1Before judging a channel or partner's ROI on last-click or 30-day attribution, map your product's actual conversion timeline — a high-consideration, low-awareness product may convert mostly outside any standard attribution window.
- 2If your product requires real education before a customer will trust it, invest in deep, fact-checked, long-form content as the actual acquisition engine, not a supporting asset bolted onto paid acquisition.
- 3Choose a small number of high-conviction partners assessed on traffic and email capture over a wide network of affiliates measured only on immediate conversion, especially if your consideration cycle is long.
- 4Build your own measurement layer rather than trusting platform-reported attribution numbers — they come from platforms with a different incentive than yours.
- 5Target long-tail, highly specific content topics where being the only deeply-researched source on the internet makes you equally discoverable through traditional search and AI answer engines — but check that assumption against your actual customer base's tool usage, not just industry hype.
- 6Personalize lifecycle and email content around what a customer has actually engaged with or shared (onboarding data, purchase history, etc.), since consistent engagement is likely what keeps a brand's messages surfaced as AI increasingly filters inboxes.
Full transcriptShow ↓
Bryan Mahoney
Thanks for tuning in to the Brilliant Commerce Podcast, where I get a chance to sit down with the brightest minds in commerce and unpack the secrets behind some of the industry's most iconic, or as I like to say, soon-to-be-iconic brands. Normally I'm joined by brand operators, but once again today we're going to change things up. I'm thrilled to be joined by Ben Grynol, Head of Growth at Levels, and also a podcaster in his own right — so double-dipping a little today. I'm hoping to get some really good insights out of Ben, but also learn how to perhaps be a better podcast host.
Bryan Mahoney
Ben isn't a commerce founder or operator, at least not in the traditional sense — he's built companies in apparel and healthcare, and now leads growth at one of the most compelling health tech startups of the decade, Levels. We'll get to the Levels story in a couple minutes. Through it all, what I've learned from Ben is that he's developed a deep understanding of what drives people to trust, engage, and stay loyal to a brand, commerce or not. When we're thinking about acquiring customers or building communities, there's definitely a through line there. So in this episode we'll talk about community through content, education, without focusing too much on actually selling.
Bryan Mahoney
And I know commerce brands tuning into this can learn a lot from what it takes at any company that's using data the right way to turn insight into action. So with that, Ben, welcome to the pod.
Ben Grynol
Yeah, thanks for having me on — super excited to do a pod here with a fellow Canadian, dig into everything commerce, marketing, and growth related.
Bryan Mahoney
There you go, you buried the lede, I was going to get to that. We are going to do our best to not out-Canadian each other, but it might be hard. So maybe kick us off — give me the three-minute background on you and what brought you to Levels. We now know you're Canadian, but there's far more to the story than that.
Ben Grynol
Yeah, so I grew up in a small town called Gladstone, Manitoba, a couple hours from Winnipeg — a town of 300 people, lived there till I was three, then moved to the big city. Ever since I was a young kid, I was sort of the lemonade-stand kid, thinking about buying and selling things, and that's what got me interested in the idea of commerce. When I was 13 I started my first business — odd jobs, yard work, painting, deck building, you name it.
Ben Grynol
I did that for about ten years, kept launching different micro-projects and businesses, and fast forward to now I've done a suite of things across commerce — one of the more Canadian things was, before Levels, I was part of an on-demand food delivery company we built out of Canada called SkipTheDishes. Really fun to go head-to-head in the marketplace space and watch that space grow a number of years ago — pretty wild, exponential growth trajectory to be on. A lot of fun.
Bryan Mahoney
That's awesome — amazing to have a front-row seat to something that grows like that. Someone unexpectedly being Canadian, obviously familiar with it — I also started my own business when I was 13, so we share the Canadian roots and perhaps the entrepreneurial ones too. I have to ask, we're now April 1st, fast approaching the end of the hockey season — you're obviously a Montreal Canadiens fan, being born in Winnipeg.
Ben Grynol
I don't know, there are a number there, but I think I have to be loyal to the Jets.
Bryan Mahoney
All right, I think they're at the top of the NHL power rankings as we're recording this, and I'm very tempted to go down the path of who I'm drafting in my fantasy playoff pool, but there are more important subjects we can get to — specifically commerce. I kicked us off saying this was non-traditional for me, in that you're not a brand operator, there isn't a commerce component. But what I learned in the warm-up is that Levels isn't just a health tech app — there's a commerce component to what you're doing. I'm really interested in the whole story — can you tell us about Levels?
Bryan Mahoney
What I've seen from the outside is a company that's grown very quickly, backed by fantastic investors, going after a problem space I find really interesting. Tell us about Levels, and then weave us into the commerce component under the hood.
Ben Grynol
Cool. There's a lot of media out there about Levels and how we got started, but I'll give the shortest version. Back in 2017, Josh Clementi, founder of Levels, one of the co-founders, was working at SpaceX at the time, and was running into these challenges of not understanding why he had what felt like an energy deficit — in his own words, he felt severely sick, like he thought he was truly sick. Eventually he figured out he was having challenges with his glucose from fueling his body incorrectly. He was exercising regularly and eating what he thought was healthy — quinoa, brown rice, sweet potatoes, proteins.
Ben Grynol
But what was happening is he was eating food that didn't necessarily help his glucose, his main energy source, be at equilibrium — he was getting these oscillating dips in his energy levels. Once he realized that, he started taking finger sticks, and realized there was this thing called a continuous glucose monitor — a little piece of hardware that goes on the back of your arm, but monitors your glucose in real time. He thought, cool, I just need to find a way to provide access to people so they don't run into the issues I've been having.
Ben Grynol
It took him a couple years to break through that cycle, because at the time there weren't doctors willing to prescribe it — it's still a bit of a Sisyphean endeavor depending on who you talk to, though CGMs have become more prevalent in the marketplace, not as difficult to access as five-plus years ago. His insight was, if I can bring this technology to the world, that'll let us better understand our health. He finally got access through a third-party provider where we could actually do commerce — sell the hardware, but build software on top of it so people could see how different food and exercise affected their glucose level.
Ben Grynol
You're tying all these insights together, and fast forward — when Levels started, in June 2019, the word metabolic health was not something you'd hear regularly. The idea of glucose monitoring, or even getting blood work or lab panels, wasn't something people talked about. As we've educated and built the market, we created this category of metabolic health — now you hear Andrew Huberman, Peter Attia. I'm not saying we were first, but we helped create the space through a lot of the education we did — here's what this is, here's what metabolic health is, here's why you should care.
Ben Grynol
Fast forward to today, we've realized the body is this black box people don't understand a ton about, and there are all these tools — just like you have tools to build a house, wrenches for plumbing, soldering irons for electrical — different tools do different things. A DEXA scan tells you something different than a glucose monitor, than a blood test, than a VO2 max test. You need to wrap up all these tools to cover a wide metabolic surface area to help people. On the commerce side, you're selling the tools; on the software side, you're able to provide people an insight based on their personalized data. So that's—
Ben Grynol
—the story of Levels, how we've evolved, our journey and vision for where we think the world needs to go. Luckily, with everything happening with AI, I think there's a lot of opportunity for companies — not just Levels — to take that data and provide a deep foundation of insight and actionable plans for people to take their health into their own hands.
Bryan Mahoney
That's just an awesome origin story — so often we hear a founder going after a problem they were compelled to solve for themselves. We try to eat healthy, think we're putting the right fuel in our bodies, but it's interesting when you get access to the data, and sometimes it can be too much. I fell into this myself maybe ten years ago when I started tracking my sleep — I'd tell anyone who wanted to hear that I was a fantastic sleeper, wore how little sleep I could get as a badge of honor, thought I was highly functional. Then I looked at the data, and it told a very different story.
Bryan Mahoney
So how do I balance the data I'm seeing with how I'm actually feeling? I find the problem space you're going after fascinating — I have so many follow-on questions about the technology, data, privacy, and you've already said the letters AI, normally I try to save that for the end, but that's okay, we've got this. I appreciate you showing me the CGM on your body. I was at a beauty company for four and a half years, Glossier, and we talked a lot about the power of brand and how important brand trust is whenever you're creating a product you put in or on your body.
Bryan Mahoney
This is very much a product you have on your body. Levels has been around since 2019 — how important is your customers' affinity with your brand, and how much does that weigh into their decision to buy you over something else, and ultimately wear something on their body that isn't always hidden?
Ben Grynol
It's interesting — I don't want to digress too much, but let's fast forward a bit to AI, then talk about the foundation of education and trust. Right now with AI, the short answer is AI lets anyone create a ton of content really fast about these categories. The insight we recognized, rewinding back to 2019, was, how do we build that trust with people? You've got a device, a piece of hardware, that people don't understand the utility of, don't understand what it is or why they should use it.
Ben Grynol
Two, there's usually an aversion to the device — it does go below your skin, a micro-needle, but it feels like a little pinprick, it doesn't hurt. And three — sorry — there's a huge price point, it's expensive for people to use. Taking all these factors in, the only way you're ever going to convert somebody, full stop, is by trust. And that comes through a very long consideration cycle — not something you can run Meta ads against and just snap your fingers.
Ben Grynol
Somebody discovers the product, understands the utility, pays 400 bucks, and they're through the door doing this thing — you have to build that trust. The way we did it was all through education. We said, if we believe this is true, the education needs to be deeply researched, science-based, data-backed information, all fact-checked the exact way you'd do any editorial process. We're well into the six, maybe seven hundreds of blog posts now — that was our foundation of content. We said, we're going to create the widest and most highly trusted source of metabolic health information on the internet, full stop. Once that became our game, we realized we weren't in the game of selling anything to anyone.
Ben Grynol
What we wanted to create was behavior change, just having people read blog posts as a starting point — maybe, like you said with sleep, maybe I should think about this differently, maybe I shouldn't grab the Coca-Cola or the Twinkie — that's hyperbolic, but also things people think are healthy, like grapes, which are high in sugar, or oatmeal, which can give people very high glucose spikes if they're consuming it, we'll call it, naked, without enough protein, fat, or fiber with it. Educating people to make different decisions is the starting point into this twelve-month-plus consideration cycle for some of these more expensive new categories of products. We've always thought the way to build that trust is strictly through education.
Ben Grynol
If you can do that, then when people think about which brand to go to, it becomes less about price sensitivity and more about, do I understand why I trust this company. If the answer's yes, because it's transparent, and the value exchange feels weighted more in the consumer's favor — they've been getting a lot of information, getting something for free without being asked in return, now buy this thing — I think that's helped us, over almost six years now, really build that deep buy-in and engagement from the community.
Bryan Mahoney
I love that answer. I'll bring it back to a word I like to use a lot, iconic — I associate that with iconic brands, and another term I wrote down while you were talking was, I think about Levels as a high-consideration, high-AOV purchase. Those are often found at iconic brands.
Bryan Mahoney
That's where investments in trust and brand building — we don't talk about them enough, because they're difficult to quantify, like, oh, this affected my conversion rate, or lowered my CAC, because I've done all these brand campaigns and my customers trust me and will be advocates for me — especially with this high-consideration purchase twinned with the fact that you're putting it on your body, in your body with that needle. In your role as Head of Growth, I'd be curious how you think about managing spend against traditional acquisition versus brand building versus education versus community.
Bryan Mahoney
How do you rationalize that to your board, your investors, to get sign-off, and what advice would you have for a brand just getting started, maybe not in your category but with a high-consideration purchase and high AOV, thinking about the most effective way to operate as a marketer?
Ben Grynol
We invested deeply in the content side, and that's always the hardest thing to get people on board with, because there's not necessarily clear attribution — you can't say new piece of content X produced Y results. It's in aggregate over a long period, they're perennial, they keep compounding — the Lindy effect, the longer something's around, the more successful it is, exact reason Back to the Future, one of the greatest movies of all time, will never get old. But we've taken a big stance internally, and luckily everyone's aligned that this is the swing we want to take. Content expands across written, audio, video — we're constantly putting out media in very long-tail categories.
Ben Grynol
This is maybe digressing into the SEO play, since we get the question a lot — how do you win the SEO space when AI is just giving summaries? If we're creating something as long-tail as PCOS, polycystic ovarian syndrome, and ground flax, the probability of someone else having that article on the internet is pretty low, because it's so long-tail. The probability of surfacing in one of the links in the recommended summaries — let's say on Google specifically — is pretty high, or it doesn't matter, you're using OpenAI, doesn't matter which platform. But the more of this long-tail content we have in aggregate, as opposed to the middle ground of what is metabolic health — and yes, we have to have that too—
Ben Grynol
—that's where we win. So we invest very deeply in that from a spend perspective, and we also invest deeply in partners — but our strategy differs in that instead of having a wide set of partners, let's say 200-plus affiliates, we have a limited set, and we make sure we're investing only in the top partners, where attribution isn't always going to be clear, but where we've got conviction based on input metrics — how much website traffic that person drives, what the email capture looks like. We're still capturing some data, but we're not anchoring on bottom of funnel to measure success of spend, because we know within 30 days we get roughly 20% of our conversions.
Ben Grynol
Within [a longer window] we get roughly 40%, and we're looking at 60% of our conversions coming after a year. You have to find ways of nurturing people over the course of wherever that first touchpoint was — if we measured attribution based solely on partnership spend, it's, oh, that's a failure, the CAC's too high. So we've got a number of metrics we're looking at for partner assessment. Once we find those top partners, we double down on those instead of going longer tail with other partners. Those are the two main buckets we invest in — we don't do a ton on performance marketing, hardly at all, for very deep reasons, happy to get into it.
Bryan Mahoney
Let's shift gears and talk about measurability and the tech stack a bit. Levels is obviously collecting a ton of data — I imagine that's a homegrown or proprietary data-capture platform you've built and maintained, probably continuing to iterate on. On the marketing or performance-marketing side, have you built something proprietary, or take us through what the martech stack looks like — bonus points if you can tell me how it's evolved since you've been there.
Ben Grynol
We don't run Meta ads as an example — across performance marketing, other than a bit of search, modest. When we do run Meta ads, we throttle it — we're conscious of what we're doing and why. Mostly, no secret sauce, it's around retargeting based on web traffic from a very specific partner, and our campaign has to do with retargeting people based on recall of seeing that partner, wherever they saw them. High recall in the moment, retargeting based on an event like that — and then our strategy is email capture.
Ben Grynol
We're not trying to go for bottom of funnel, because we know it'll be a lot easier and less expensive to capture those email leads and nurture them with education over the course of a year, if it takes a year, as opposed to the typical strategy — retarget, go for the kill, go for the conversion, or if you're prospecting, prospect for email capture or conversion. We know, because of the nature of our product, that's not the strategy that works for us. That's how we throttle Meta. The reason we don't invest deeply in it, and how things have evolved — the consideration cycle is so long, with so much education involved, we don't have a product people understand the clear utility of the same way.
Ben Grynol
Why can True Classic scale up the way they have? The utility is understood — I see a T-shirt, I know what it does, I'm looking at the brand, I go, do I trust it'll fit well, and there you go. For a product that's niche and longer-tail in how well people understand it, it's very difficult to assume you can capture that customer for a reasonable CAC. That's how we think about performance marketing and how it's evolved — we've always had this strategy, we've just gotten more efficient with it, as capital-efficient as we can. Our customer acquisition cost is pretty low — we hit 50 to high end, like $80, which is pretty great, low for the space.
Bryan Mahoney
Yeah.
Ben Grynol
LTV sits around $285, $288 depending on any given month, so we're hitting really good LTV-to-CAC ratios, 5-plus, low CAC. We're very much focused on efficiency versus growth at all costs, because it's a very long game to get traction in the space.
Bryan Mahoney
What about the tools you're using to calculate that — do you have a dashboard you're going to daily, weekly, monthly? What does that look like for attribution too? I'd be curious what you're relying on — it sounds like some multi-touch or blended model — and what gives you confidence it's a $50 CAC and not a $150 CAC?
Ben Grynol
The platform we use is PostHog — that's where we build our full-funnel insights, and when we're measuring bottom-funnel attribution, we're only using PostHog. We don't take reported data on Meta or Google.
Bryan Mahoney
Cool — you don't trust the platforms, they have a different incentive structure.
Ben Grynol
Exactly. Even if we change the attribution model, click-through versus view-through, one day versus seven, it doesn't matter, it's not going to be clearly reported. When we're actually measuring, we're looking at PostHog, looking at the UTMs that come in — and this is where it gets difficult with partners versus something clear like email. Attribution's never perfect, but with a partner we look at how much web traffic they drive, what the actual conversions were said to be, and find a middle ground to measure what we believe the return on that partner is.
Ben Grynol
There's some area under the curve — you know your baseline web traffic in a day, the event happens, one to two days of spike in web traffic, what does email capture look like, then the actual conversions coming in — even if not perfectly attributed, that gives a pretty good window, and the probability those conversions came from that event is quite high, because there's no mixed noise or signal from two events happening the same day. We've gotten pretty tight over six years fine-tuning that, so we've got pretty high confidence in what we're measuring.
Bryan Mahoney
So it's not really a build-versus-buy question on the app side — but on marketing, it sounds like you're buying some software and relying on your own skills to slice and dice that data and drive the insights you're looking for.
Ben Grynol
Exactly, that's exactly it.
Bryan Mahoney
I was somewhat surprised to hear there's some activity on Meta, given you're operating in a regulated space — does that make your job harder in any way?
Ben Grynol
We're a health and wellness company, so — because we build software on top of the devices, we're not selling the device itself. Right now the devices we build on top of are prescription devices, so they're being sold, or distributed, by another company — we're just the conduit to get that sale through. That's how you're a health and wellness company versus one regulated in the way you'd have to operate. Now the market's changed in that there are a couple non-prescription glucose monitor devices out there, so people can buy over the counter, which changes things a little.
Ben Grynol
But the way we've had to operate, we operate like a health and wellness company, we're not worrying about some of those regulations. The biggest consideration for us is making sure we're never making claims like a glucose monitor can cure X condition — as soon as you make those claims, that's not a gray area anymore. We're sure we're saying people who've used glucose monitors have found benefit in losing weight or improving their longevity — there are ways of positioning that messaging. It's about being clear that these tools can help, but there's no correlation between using the device and curing any condition. Always a challenge.
Bryan Mahoney
Got it. Let's pivot back to AI. One other thing I wrote down — it would seem to me these investments in content, which accrue to the brand trust you've established, also help from a search point of view. I'd be curious to hear about share of traffic, how things have changed over time — traditional optimizing content for discoverability from search engines, how that's shifting to people not turning to Google, but to Perplexity or ChatGPT, asking agents or chatbots to do the search for them. Have you shifted your content strategy toward relevance instead of search engine optimization? If not, is that something you're thinking about?
Bryan Mahoney
And if you have, what lessons have you learned, what's working? I was just in Vegas at ShopTalk, and brands and retailers are genuinely worried about this — how do I keep optimizing for SEO, since it's been such a strong organic channel, while also thinking about agent traffic. It all comes down to how we show up in these results, and understanding an algorithm that's changing — not one algorithm for everyone, but context-aware algorithms making their own decisions. That's a mouthful — I see you smiling, so it sounds like you've gotten started here.
Bryan Mahoney
What can you share?
Ben Grynol
It's back to what we were talking about — when people are concerned about SEO for core utility products, like a T-shirt, there's only so many ways to talk about the T-shirt and what it does, so you're in a very competitive space. But health and wellness has so many avenues for content, it's almost endless. So the way we think about SEO is ranking on all these long-tail pieces, because we're creating good content.
Ben Grynol
Even if agents are using AI to summarize a topic, the likelihood — assume it's not possible that you've got the one piece of content on the entire internet about a very long-tail thing — you're always surfacing in those results. The more we create this, the better it helps our strategy, because people are coming to Levels for the deeper dive. If people are using agents, they might get a high-level summary, not the deeply researched dive on PCOS and flax — you can do it, but also thinking about our members, a lot of whom are older.
Ben Grynol
More than half our members are primarily women, 45 to 65, and we have members in their 70s and 80s. So even the way you think about user behavior — it's easy to get lost in the tech world of what's happening now, and it doesn't mean we shouldn't be forward-looking, but what's happening with people actively using these tools versus the tools your actual customers are using — if those differ, you have to play to that strategy too.
Ben Grynol
So we think about how we use AI when thinking through content topics, but not from a competitive standpoint of, oh my gosh, we have to shift our entire strategy — because the vision we had, to create this large repository of research-backed information, hasn't changed, and I think that'll continue to be the place going forward. There's a world where people can do deep research and get all the links, whether using Perplexity, Anthropic, or OpenAI — doesn't matter the product — you can find that information yourself.
Ben Grynol
But sometimes people still want the company to have the brand stamp on it. That's generally how we think about it, and we use AI a lot internally — even in how we're thinking about health data and providing insights moving forward.
Bryan Mahoney
I like that answer. A follow-on — I talked about SEO as an important channel, and it sounds like email capture, so you can get prospects into a funnel and give them content to start developing a relationship with the community or the brand directly — email seems like an important channel too, am I right?
Ben Grynol
Correct. Roughly 30% of our conversions come through lifecycle — a very big channel for us, all education-based. We do a little product marketing through it, but it's mostly editorial content.
Bryan Mahoney
That's another trend I heard coming out of ShopTalk — email's been a fantastic channel for marketers who've gotten it right, establishing that connection with customers. There's a lot of excitement about AI creating more content, maybe more compelling, but also a looming fear that more AI is acting on our behalf on our devices. If you look at the recent iOS update, it's now helping summarize or filter and decide which emails are important for you, Ben, or for me, and which matter a lot less.
Bryan Mahoney
A question I get asked all the time is, how can I use AI, or machine learning, or personalization, to make sure I'm relevant when someone asks a question — but most importantly, so email continues to perform, how do I make sure my emails are relevant? Is that something you and the team are already thinking about — leveraging AI, even though we don't know the formula yet, to make sure emails effective today remain effective tomorrow?
Ben Grynol
We think about the email part very much in personalization — especially as it ties into things like lab work. Somebody comes back, maybe certain markers are out of range, what's the most relevant content for them based on those tags. We're always thinking about the personalization angle related to demographic information — capturing that during onboarding, making sure we serve up content most relevant to people. I think the way to win that is, if people are consistently engaging with your lifecycle material, the probability that — if AI comes in and says, here's what's served up to you the most—
Ben Grynol
—it's going to take those user events into account, are you opening these things, actually engaging with the content. If the answer's yes, you're going to keep moving to the top. So we're always thinking about how we create value for people based on this new health information they get, and moving along in the journey that way.
Bryan Mahoney
I think personalization is a great way to put it — I like the framing of how do we deliver value, something more personalized to what we've learned from you, the information you've shared through first-party data. Deploying personalization that way is great, and it goes back to the tenets of real brand building — it's not just personalization to compel a purchase in the moment. If I'm a brand, I want to build a long-term relationship with you, using the data you're trusting me with to personalize your journey along the way.
Bryan Mahoney
Whether that's showing up relevantly in search, or making sure email content gets through to your inbox, those are some of the tenets of building a really durable brand. It seems like you all are well on your way at Levels. We've just about run each other out of time — any last words of wisdom, things you're excited about in the coming year?
Ben Grynol
One thing I've been thinking a lot about is this idea of prescriptive advice, or prescriptions — not literally, but in the sentiment of what people are looking for. One trend I've noticed is people are willing to pay for the concierge, or high-touch, experience, whether through a service or product. As we move forward in commerce, you can see it in health and wellness with protocols — people are willing to pay for it, just tell me what to do, and I'll do it.
Ben Grynol
You can extrapolate that into a lot of avenues of commerce, or services generally — I'll pay for high-touch, whatever concierge service that might be. I think people are going to be very interested in this idea of prescriptive commerce, as a bucket or layer on top of any product or service — something that'll continue to surface as people are willing to invest in high-touch, high-experience, premium products. That's probably an area that's—
Bryan Mahoney
There we go — you heard it here first, prescriptive commerce is the new social commerce, I love it. Ben, you're a fantastic guest, I always love having fellow Canadians on the pod. Go Habs, go Jets. Excited for the season ahead, not just at Levels but for the hockey team each year. Thanks for joining me, have a great week, hope to talk to you soon. Take care.
Ben Grynol
Thanks.
Bryan Mahoney
Bye-bye.