Justin Fredlender argues that growth stopped being a marketing-channel problem and became an organizational one. He walks through his framework for deciding what to build versus buy, why "internal stewardship" matters more than technical prowess, and why declining purchasing power and digitally native competition now demand real coordination across brand, ops, and consumer insights.
Behind the Expert
Justin Fredlender's career runs through consulting, hypergrowth D2C, and now independent advisory work across a portfolio of commerce brands. He started in management consulting, but left after two years chasing a role with more analytical teeth — a "growth analyst" job req vague enough it could have belonged to almost any industry, but whose day-to-day of mining large data sets for growth recommendations hooked him. That path led first to an early e-commerce standout that had a quick rise and an equally quick fall, then to MVMT, the watch and accessory brand he joined at 25 alongside a company of twenty-five-year-olds figuring it out in real time — he helped scale it to nine figures before it was acquired. From there he spent roughly five years leading the growth org and much of the direct-to-consumer business at Ritual, the supplement brand founded by Katerina "Kat" Schneider. For the past year and a half he's worked independently, advising brands across verticals and stages of scale — an aperture-widening move he says has sharpened his own thinking more than staying inside any single brand's playbook ever could.
The Quick Hits
- Growth stopped being a marketing-channel problem and became an organizational one. Fredlender now treats it as holistic — requiring coordination across brand, operations, product/R&D, and consumer insights, not just media buying.
- Two macro forces are squeezing every operator at once: the dollar's purchasing power has been eroding for roughly 50 years, and the number of digitally native e-commerce brands in the US has reportedly doubled in five years to over 100,000 — leaving consumers with less discretionary money and more competing options than ever.
- First-purchase discounts, the industry's default acquisition lever, are quietly training customers to switch. Fredlender argues brands need to underwrite that dynamic explicitly in their retention math rather than treat it as background noise.
- His build-vs-buy framework starts with two questions: do we actually have the capability to build this, and if we do, is someone already doing it better — and will they keep improving it? Most smaller brands fail the first test on AI-driven tooling specifically.
- The real cost of building isn't the build — it's the maintenance, and more overlooked, finding an internal "steward" accountable for the tool actually paying back its investment. Without a steward, he says, even a successfully shipped tool is "almost a non-starter to begin with."
- His refined heuristic is buy, then build: if a platform can't be extended or built around after purchase, that's a smell — a sign you've bought a legacy system without real APIs.
- AI is further along in analytics than in creative, in his experience. It's turning ordinary employees into "semi power users" of data platforms, while on the creative side it mostly amplifies people who already have a professional's judgment rather than replacing that judgment.
From a "job req that could have been anything" to nine-figure exits
Fredlender didn't set out to work in commerce. He started in management consulting and left after two years chasing more critical-thinking work, landing almost by accident on a "growth analyst" posting vague enough that it could have belonged to any industry. What hooked him was the day-to-day: mining large data sets and turning them into recommendations. That led to an early e-commerce brand that rose fast and fell just as fast, then to MVMT — the watch and accessory brand he joined at 25, when nearly everyone else there was 25 too, and helped scale to nine figures before its exit. Five years leading the growth org and much of the direct-to-consumer business at Ritual followed, then roughly a year and a half of independent advisory work across a wider set of brands.
The math has changed underneath every brand
Fredlender points to two data points he considers underappreciated. First, the purchasing power of the dollar has been declining for roughly 50 years, pushing more spend toward necessities like housing and food and less toward discretionary purchases — the exact category most of the products he's worked on live in. Second, the number of digitally native e-commerce brands operating in the US has, by some counts, doubled in just the last five years, putting the total above 100,000. Put together, he argues, today's consumer has more purchasing options and less spare money than at any point in his career — and correspondingly less patience for a product or experience that doesn't deliver.
Discounting yourself into someone else's retention problem
The conversation's host, Bryan Mahoney, pushed the point further: the industry's default acquisition lever — the first-purchase discount — has become a structural liability. Because switching costs are close to zero and customers are served competing offers constantly, a steep first-purchase discount doesn't just win a customer, it teaches them that switching is free and expected. Fredlender's framing is that every acquisition dollar has to be modeled against the near-certainty that a competitor will try to win that same customer back with an even better opening offer.
Fredlender's read is that this pressure compounds with how easy it now is to launch a store in the first place. Shopify and a more frictionless contract-manufacturing landscape lowered the barrier to entry, which is good for entrepreneurship — but it also means many founders underestimate the fine-grained execution required to win in a market that keeps getting more competitive for the same finite pool of customer dollars.
Growth is holistic, not a marketing department's job
Fredlender's biggest shift in thinking, working across many brands instead of one, is that growth problems are rarely where they first appear. His instinct now, when a client's repurchase rate is off, is to work backward — is it a product problem, an operations problem, or genuinely a marketing problem — rather than let the default reflex of "acquisition is down, talk to marketing" drive the diagnosis. That reframing extends to what he now considers the four pillars of growth: brand, operations, product and R&D, and consumer insights, the last of which he thinks most brands underinvest in relative to its value.
He credits consumer insights specifically with letting him target tests and diagnose problems in ways Shopify's own dashboards never surface — the kind of signal that tells you whether a repurchase-rate dip traces back to the product itself, fulfillment, or something else entirely, before a single media dollar gets reallocated.
The build-vs-buy framework
Asked how he advises brands weighing whether to build internal tooling or buy an off-the-shelf platform, Fredlender reduces it to two questions: do we actually have the capability to build this, and — if we do — is someone out there already doing it better, with the ability to keep improving it faster than we can? He illustrated the framework with a recent, concrete case: a client brand that had outgrown ad-hoc analytics and needed a centralized data platform. His first instinct was to build, because most packaged analytics tools just ingest Meta or Shopify data into generic dashboards that don't flex to a specific brand's needs. But once the team mapped out what building would actually require — and weighed it against how fast AI is advancing inside analytics platforms themselves — the calculus flipped.
Internal stewardship matters more than technical prowess
Fredlender's sharpest point is that the real cost of a "build" decision isn't the build itself — modern AI tooling has made writing the code the easy part. It's finding someone inside the organization who owns the tool long after launch, accountable for proving the investment actually pays back. He noted that most companies have spent the last several years deliberately reducing organizational redundancy, which means a build project rarely gets a dedicated owner — it becomes something somebody does off the side of their desk, with no guarantee they have the bandwidth to maintain it once the initial excitement fades.
Mahoney sharpened this into a working heuristic for evaluating vendors: buy, then build. If a platform can't be extended or built around after you've purchased it, that's a smell — a sign you've bought a legacy system without real APIs, one that will be painful to work with the moment your needs evolve past its defaults.
AI is further along in data than in creative
Fredlender estimates most brands he works with sit somewhere between 50 and 75% AI adoption — not because AI isn't useful, but because most organizations lack a dedicated resource who can evangelize and coordinate its use across functions. Analytics is where he sees the clearest gains: LLMs are naturally suited to interrogating data sets, and prompting a model to get 75% of the way to an answer turns far more employees into what he calls "semi power users" of BI tools — a much easier ask than expecting someone to mine raw data themselves in a platform like Looker or Chord.
He draws a sharp line between that kind of organization-wide democratization and AI's role in creative work, which he sees as enabling an already-skilled professional rather than replacing judgment broadly. A media buyer's AI-generated ad still has to survive an executive's sniff test; a creative director who already has a clear vision can use the same tools to execute faster without losing that credibility. Both he and Mahoney also flagged a data-safety trap they're seeing at larger, well-resourced brands: teams defaulting to "build it ourselves" specifically to avoid exposing data to LLMs, when the real fix is building context and guardrails that let teams query safely — not funneling raw PII into a chat tool, but not over-building out of an abundance of caution either.
Sound Bites
- “What really intrigued me was the day to day of the job, which was analyzing big data sets and coming up with recommendations for growth or consumer journeys.”
- “Consumers have never had so many opportunities to make a decision on the products they want to purchase. And they're also going to be hypercritical of the products they purchase after they receive them.”
- “Especially with what I think, as an industry, we've decided to do with these first-purchase discounts — we are almost incentivizing the switch. You have to think about that in your model.”
- “Too many entrepreneurs and CEOs see customer acquisition go down and they say, "marketing, what's going on?" And there could absolutely be something going on from a marketing perspective — but it all plays together.”
- “We ran the numbers — it would be way more expensive, and we were going to build some tech debt. Is it going to become obsolete in six to 12 months? Do you have not only the ability to build it, but also to maintain it?”
- “It's not just maintaining it — it's even finding a steward to own it within the organization. If that in itself is a hard question to answer, then building it yourself, and also having the steward, and also maintaining that technology — that's almost a non-starter to begin with.”
- “The capabilities of an LLM lend themselves better to analyzing data sets than they would to creating content that a CEO or founder or whoever would feel good about putting out.”
The Chord take
Fredlender's framework reframes "build vs. buy" as an organizational-stewardship question rather than a technical one. The instinct when evaluating a new AI-era tool is to ask whether your team can build it — and increasingly, with AI-assisted coding, the answer is yes. His argument is that this is the wrong first question. The harder, more decisive ones are whether someone else is already doing it better and will keep improving it faster than you can, and whether anyone inside your organization is actually positioned to own the thing once it ships — accountable for proving it pays back the investment. Skip that second question and even a technically successful build becomes dead weight: an unmaintained system nobody is responsible for, built by someone doing it off the side of their desk in an organization that has spent years cutting the redundancy that used to make that kind of ownership possible. That reframing connects directly to his macro argument about why coordination beats channel optimization right now. With purchasing power declining for decades and the number of digitally native competitors reportedly doubling in five years, customers have more options and less money — and the discount-driven acquisition tactics most brands default to are actively training those customers to keep switching. In that environment, growth isn't a lever one team pulls harder; it's diagnostic work that has to run across product, operations, brand, and consumer insights before a dollar gets reallocated. The brands that will separate themselves aren't the ones building the most internal tooling or spending the most on acquisition — they're the ones with someone accountable for making sure the underlying customer experience, and the tools built or bought to understand it, actually deliver.
Put it to work
- 1Before building an internal tool, apply Fredlender's two-part test explicitly and write the answers down: do we have the capability to build this, and is somebody already doing it better who will keep improving it faster than we can.
- 2Name an internal steward for any analytics or AI platform — built or bought — before you greenlight the investment. If no one can credibly own it, don't build it.
- 3Audit your first-purchase discount strategy against your churn and switching data. Model how much of your "acquisition" is actually discount-shopping that a competitor's better first offer will reverse.
- 4When repurchase rates dip, resist routing the problem straight to marketing. Diagnose whether it's product, operations, or fulfillment before touching the media plan.
- 5Treat "can we build around this platform after buying it" as a purchasing criterion. A platform without real APIs is, in Fredlender's words, a smell — not a minor inconvenience.
- 6Push AI adoption in analytics ahead of creative. Use it to get more people to "75% of the answer" and free up analysts for deeper questions, rather than expecting it to ship finished creative untouched.
Full transcriptShow ↓
Justin Fredlender
The purchasing power of the dollar has been going down for about 50 years. There's another data point that's also very important, which is the growth of e-commerce brands has, by some sources, doubled just in the last five years. So you need to focus first and foremost on what the actual customer experience is.
Bryan Mahoney
There's no tricking customers today. Unpack the fabric of any brand — they obsess over the relationship they have with the customer, they obsess over the product. That's what makes them great.
Justin Fredlender
Communicate across the business how they're using AI. Most companies are probably somewhere between 50 and 75% adoption.
Bryan Mahoney
Those customer insights — actually hear that maybe there is a product problem. And it's not just, 'no, it's a marketing problem.'
Justin Fredlender
What I still don't hear people talk about enough in the world of commerce is how this actually changes how consumers shop.
Bryan Mahoney
Well, thanks everyone for tuning in to another episode of the Brilliant Commerce podcast. This time, once again, in real life — this is my favorite way of doing this. On the Brilliant Commerce podcast, I get a chance to sit down with people who are working behind the scenes at iconic brands, or as I like to say, next-to-be-iconic brands. Today I'm joined by Justin Fredlender, who has spent his entire career behind the scenes at some of these iconic brands, most recently at Ritual, and is now acting as an independent consultant for a number of brands. Justin, welcome to the pod.
Justin Fredlender
Appreciate it. Thanks for having me.
Bryan Mahoney
Before we dive in, maybe give us a two-minute overview of your career, how you've come up through the ranks. We talked about it a little bit in the green room and I thought it was super fascinating.
Justin Fredlender
Yeah. So I started my career in management consulting and did that for about two years. What I found during that time was that there wasn't a whole lot of critical thinking involved in that career path for me, at least early on. So I really wanted to land on something that let me use some real critical-thinking skills — I had no idea what that was going to be, so I just started looking online for job opportunities and came across a posting for what was like a growth analyst. This growth analyst job req, I remember, could have applied to almost any industry — it could have been e-commerce, it could have been tech, it could have been a whole multitude of things. But what really intrigued me was the day to day of the job, which was analyzing big data sets and coming up with recommendations for growth or consumer journeys or whatever it might be — leveraging data, and an understanding of data, to drive decision-making.
Bryan Mahoney
Right, and how far back is that? That's 2013, 2014, or even earlier than that?
Justin Fredlender
Graduated college in 2013. That was 2015, when I wanted to make the move out of consulting — so that was about two years.
Bryan Mahoney
So we're just getting started with even brands having access to big data, for someone to come in and be a growth analyst. Is that fair to say?
Justin Fredlender
Yeah, and this was back in the day when MTAs were big, Convertro was big — and, as you know, none of those technologies really exist anymore. By luck of the draw, I got that job and loved it. The first company I joined in e-commerce was kind of a golden company in the space at the time — had a quick rise and fall. Then I joined a company called MVMT, a watch and fashion accessory brand. For listeners who aren't familiar with the company, it was one of the early, really big D2C successes.
Bryan Mahoney
Yeah.
Justin Fredlender
I was about 25 when I joined that company, and everyone there at the time was about 25 years old — we were all figuring it out, testing and learning, and honestly having a ton of fun doing it. We scaled that business up to almost nine figures and exited. From there I found myself at Ritual, where I was leading the growth org and much of the direct-to-consumer business — I was there for about five years. It was an incredible journey working with some of the smartest people in the space, and it really makes you understand that the success of some of the biggest companies in the DTC space isn't dependent on how you evaluate clicks and optimize CPCs. You're selling a product at the end of the day, and you need a team across many facets of the business that's able to execute incredibly well on specific strategic initiatives to really break through and drive that growth. So anyway, for about the last year and a half, I've been doing the independent consulting and advisory thing, and honestly having a great time. When you're working with one brand, you can develop a fairly siloed way of thinking, because a few things work and maybe they work well specifically for that brand — that's kind of where you spend a lot of your time. But going out and actually working with a multitude of brands, across many verticals and many different sizes and stages of scale, has, over the past year and a half, really accelerated my own learning and made me even better at what I do.
Bryan Mahoney
Yeah, it sort of widens the aperture of the types of problems you see — even though e-commerce is a quote-unquote 'solved problem.' I think once you actually get in, that's always been sort of laughable. But that's a fascinating journey. I'll bring it back to Ritual for a second, because I know you spent about five years there. As I think about what it means to be a trusted brand, or an iconic brand, there's a relationship you have with your customer — especially a brand that's selling a product you're going to put on your body, or in Ritual's case, sort of in your body. You were really focused on growth, and I'd imagine retention too, since if I remember correctly there was a big subscription component. Can you talk about how data was used within Ritual to build that solid brand foundation — why does someone see a Ritual ad and believe it's the right product for them? And how were you helping the brand and marketing teams strengthen that relationship, to keep people coming back and keep the trust that they're putting that product in their body?
Justin Fredlender
First of all, the brand positioning and overall foundation of the brand were put in place well before I ever got there. Kat, the founder, had a vision for a product based on her own personal experience, and she executed on it. As the team grew, we were able to really build out that brand positioning and perspective — that part was already powerful. So when I joined the company, the brand foundation already existed, and I think what they were missing was somebody with my background and skill set — somebody more data-driven.
Bryan Mahoney
Okay.
Justin Fredlender
I'll pull it back and give you one of the first success stories I had personally during my time at Ritual. Within the first few months of joining the brand, it became very apparent that the marketing creative was fantastic and the brand positioning was on point, but they were purely investing in the wrong areas of their marketing stack. I'll bring this back to retention in a minute, but on the customer-acquisition side, it was simply: let's spend more on the channels that are actually working, if you look at attribution the right way.
Bryan Mahoney
Yeah.
Justin Fredlender
And let's see if we could accelerate some growth. When you join a company and find a quick win, it helps you build trust within the organization.
Bryan Mahoney
Sure.
Justin Fredlender
That, to me, was the first quick win — we reoriented how we were spending our marketing dollars and found accelerated growth fairly quickly, which was, you know, sweat off my brow.
Bryan Mahoney
Yeah, it's nice to get those quick wins.
Justin Fredlender
It's always nice to get a quick win. Now, relating this back to the retention side of things — I want to separate the tactical side of retention from the 'customer love of your brand' side of retention. Because, of course, having a subscription product is going to increase your retention, or the way you segment your email list is going to increase your repurchase rate and how often customers come back. You could do all of those things well — but if people don't like the product, they're simply not going to come back and purchase. I want to point to two specific data points that have been developing over a long period of time, but have specifically been exacerbated over the last five years by the pandemic. The first is that the purchasing power of the dollar has been going down for about 50 years — so more money is being spent on needs like housing and food, and there's less money left for discretionary purchases, like an expensive multivitamin, or many of the products we sell.
Bryan Mahoney
The high-consideration purchases — the barrier to convert is higher than ever before.
Justin Fredlender
Correct. There's another important data point: the number of e-commerce brands has, by some sources, doubled just in the last five years — there are over 100,000 digitally native e-commerce brands operating in the US alone. Put those two data points together, and consumers have never had so many opportunities to decide what to buy — and they're also going to be hypercritical of what they buy once they receive it.
Bryan Mahoney
Right.
Justin Fredlender
So you need to focus first and foremost on what the actual customer experience is, and whether you're delivering an experience that exceeds expectations and is better than anyone else's in the market. That's how you're going to drive retention.
Bryan Mahoney
Yeah.
Justin Fredlender
And then comes all the tactical stuff — you layer that on afterward. But it has to start there.
Bryan Mahoney
You can't earn the right to layer on the tactical stuff if you don't have that foundation — you have to deliver on the promise first. Customers are going to be more exacting and more demanding than ever before. Switching costs are basically nothing, because you're getting served these ads constantly — 'sure, I'll try this.' And especially with what I think we've decided to do as an industry with first-purchase discounts, we're almost incentivizing the switch. You have to think about that in your model. I know I'm going to acquire a customer for a certain amount, but I really have to get to work retaining them, because I know they're going to get inundated with offers from a competing product that basically gives the first product away free — and the retention game starts all over again. So how do we make sure the hard-earned dollars we spend getting that customer into our funnel actually keep them there? That's a really tough challenge.
Justin Fredlender
The other thing, just to expand on that — I think the reduction in barriers to entry to open an online store is part of it. Kudos to Shopify for making it so easy, and kudos to all the contract manufacturers out there reducing friction in their own funnels, letting folks become entrepreneurs and open their own stores. That said, it feels like a lot of entrepreneurs focus too much on 'hey, I heard this was easy, so let me just start a brand.'
Bryan Mahoney
Yeah, right.
Justin Fredlender
And they don't think about the incredibly fine details it actually takes to win in a market that keeps getting more competitive. Going back to the consumer — they have hard-earned dollars, fewer of them to spend. Where am I going to spend them, and why would I spend them with your company?
Bryan Mahoney
There's nothing easy about creating a brand, despite some of the hero stories we see — the in-theory overnight successes. I think if you unpack the fabric of any brand you interact with today, that you buy, that you put on your body, that you want to be friends with — they obsess over the relationship they have with the customer, they obsess over the product. That's what makes them great. It's not their ability to launch the buzziest marketing campaign. There's no tricking customers today. You might trick them into making that first purchase, but they're not going to make a second one — and they're probably going to tell five people they felt tricked. Whereas if you get those fundamentals right — deliver an amazing service, deliver an amazing product — then you can start turning those people into advocates. I remember in 2015 through 2018, there were a lot of venture dollars flowing out there for brands, so it was easy to focus on acquisition. It was kind of like a drug, to some extent. And perhaps there was an underinvestment in how we use those excellent relationships with our best customers to help go fetch other customers. There's maybe a little more work to do there. But when I always used to look at that data, customers who came in through that funnel — through, say, Justin telling me he had an amazing experience with Ritual — my friends would be more willing to trust me. Perhaps more difficult to measure, but I thought I was getting a more durable customer if I ran them through an RFM model and understood where they'd come from. I'm like, well, I want to go create more experiences like that. But there was less immediacy, perhaps — or this is my hypothesis — than just cranking the lever on Meta or on Google. So maybe a hard segue into strategy: how have you seen that landscape change over time, going all the way back to 2015, when you had the benefit of walking in with a really solid brand platform, to what brands have to do today to cut through and find that next customer — because there are fewer venture dollars floating around allowing us to spend without a lot of thought into what's actually working. I'd imagine that's a big reason people are singling you out, like, 'yeah, we tried all these things, we were told it's easy, we're figuring out it's maybe not so easy.' So how can we stand on the shoulders of those who've been there before, but don't necessarily have a playbook — and are thoughtful, and can help get them to that next level? So take us back — how has it evolved over time, and what would you advise brands today?
Justin Fredlender
Today, when I think about the world of growth — of course it's tactical, but this even goes back to what we were talking about before. You have to lay the foundation to be able to do the tactical things, with a great product and service and an eye on the customer experience. This is like the old adage over at Amazon: customer, customer. What is the experience? What I've inherently found, working with a multitude of brands now, is that many of them got up and running easily because there are low barriers to entry. Many of them found success for one reason or another — great branding, great product-market fit, the ability to operate well — something that gave them that initial push into finding some success. But what I'm finding a lot now is that the foundational work hasn't already been done. Going back to Ritual — I think that was a very unique case. The brand work evolved over time, don't get me wrong, but that foundation was already there by the time I got there. With a lot of the brands I'm working with now, they found a little success and then kind of have to work backwards a bit. What we do together is look at the data — what are your repurchase rates? They're not where they should be — why is that? Is it because they don't love the product? Is it because of ops? You've got to peel back from there. And that's actually where I think the holistic lens of growth comes into play — too many entrepreneurs and CEOs see customer acquisition go down and say, 'marketing, what's going on?' And the fact of the matter is, there could absolutely be something going on from a marketing perspective, but it all plays together.
Bryan Mahoney
Yeah.
Justin Fredlender
So I now take the viewpoint that growth is holistic. You have to be able to work across brand. You have to be able to work across operations. You have to be able to work across product — R&D, who's actually creating the product. And a big one — which I think a lot of brands dabble into, but not to the level they should — is consumer insights.
Bryan Mahoney
Yeah.
Justin Fredlender
Consumer insights, over the last five to seven years, has let me think in a very targeted way about the tests we want to run and what needs to be improved — things you can't uncover just by looking at Shopify data.
Bryan Mahoney
Right.
Justin Fredlender
So I think that's how things have really evolved — it's less about the things that are easy for me, like going to a spreadsheet and understanding the trends in the data. It's much more holistic. You really need to work across an entire company to make sure you're driving efficient metrics.
Bryan Mahoney
Yeah. And I think the Amazon 'customer, customer' idea is really important here, because it speaks to the fabric of the culture within the organization — being willing to look at those customer insights and actually hear that maybe there's a product problem. Not just, 'no, it's a marketing problem,' or 'it was this campaign's problem.' It's, 'no, we actually have a problem in the way this product is being received in the hands of customers who trusted us — they liked our message, but we've got to make that better,' or maybe it's taking too long to get the package to them. We have to be willing to hear what our customer is telling us and then do the hard work from there. And I think if you have a product — a lot of these great brands, especially the digitally native ones with a founder who had a vision — they're obsessing over a category, obsessing over a product. They have an unfair advantage versus maybe a brand that saw an arbitrage opportunity, where arbitrage is the thing they're passionate about, less so 'I built this because I couldn't find anything like it for me.' I think that's really important. It sounds like Ritual, and MVMT before that, had an awful lot of those characteristics. I'm thematically hearing you talk about where you got started and where you've come from, and now where you're going — there's a natural cycle to that, like an arc. You're sort of going back to consulting. I think about cycles in tech too — when you got started in 2015, and I'm quite a bit older than you, there was a cycle before that where, even before we called it 'digitally native,' e-commerce was shifting out of catalog commerce. Shopify wasn't really a viable option, you didn't have all the tools brands have at their disposal today. So you built a lot — you basically built your own technology. It was homegrown. I still work with brands today that have technology they built 15, 20 years ago that's still working for them. And I think the era where you really started to make your mark was an era where we shifted away from building everything to buying everything. Now it feels like we're coming full circle — a new cycle where brands are getting build-curious again because of AI. So putting your advisor hat back on, having been through this over the last 10 years — what's your advice for a brand today that's looking for consumer insights, or looking to expand beyond what the Shopify dashboards are telling them? How do they get started?
Justin Fredlender
Great question. I think it's such a company-by-company thing — the answer is different on a company-by-company basis.
Justin Fredlender
When I think about the opportunity to build versus buy, two immediate things come to mind. One: do we have the capability to build it? And even if you're talking about an AI prompt that would build you some function within an organization that would make you more efficient or solve some sort of problem, the question then really becomes: okay, if you have the capabilities, is somebody out there doing it even better, and can you find that product?
Bryan Mahoney
And will they keep making it better?
Justin Fredlender
Correct. With the smaller brands I work with, they generally don't have the capabilities on the AI side to prompt an LLM or some sort of AI model into building them a viable solution. So there are a lot of solutions popping up that take advantage of AI — and Chord comes to mind. I'll share a story of how we recently onboarded Chord with a company I'm working with, specifically because of this exact problem — we were thinking build versus buy, and it was a very apparent solution for us. So anyway, I think the answer to your initial question is that it's on a very case-by-case basis, and it depends on the capabilities of the people internally, and also their bandwidth to actually build those tools.
Bryan Mahoney
Yeah.
Justin Fredlender
Now, the example I want to give, where we went through build-versus-buy and actually ended up onboarding — Chord is a company I work with that recently surpassed the size where it made sense for us to build a centralized database. Kudos to them, it was a long time coming. We'd been working together for over a year, and it was one of those proud moments where it's like, all right, we're—
Bryan Mahoney
We've graduated, we're ready, we're growing up.
Justin Fredlender
We're growing up.
Bryan Mahoney
Yeah.
Justin Fredlender
So as we started evaluating analytics options — one thing that was really important to me, coming from somebody who's inherently very analytical, was that we had the ability to take our data and be flexible. There are a lot of solutions out there that just ingest data from Meta or Shopify, but they're not very flexible — they give you out-of-the-box dashboards, and the value they provide isn't all that great. So my initial thought was: okay, we need to build this, because we need the flexibility.
Bryan Mahoney
You don't want to be put in a box.
Justin Fredlender
Correct. We need to build a solution that works for us — not something that works 75% for everyone. We need something that works 100% well for us.
Bryan Mahoney
Yeah.
Justin Fredlender
And as we started to explore options, it became very apparent to me that, with the advancement of AI and how it's being integrated specifically into analytics platforms, we were just about to build a bunch of tech debt.
Bryan Mahoney
Yeah.
Justin Fredlender
Because we were going to build this internal database — and even now, you get prompts, your own data sets — imagine what it's going to look like in six months, 12 months, 18, 24 months. And so, serendipitously, I think you reached out at that exact time.
Bryan Mahoney
Yeah, everything happens for a reason.
Justin Fredlender
And here we are on a podcast together, having gone through this exact solution — this is prototypical build versus buy. Can we build this ourselves? We ran the numbers — it would be way more expensive, and we were going to build some tech debt. So those are all the considerations you need to put in place when you're thinking build versus buy: is it going to become obsolete in six to 12 months? Do you have not only the ability to build it, but also to maintain it?
Bryan Mahoney
This is the thing no one talks about — the hidden cost. Can you build it? Building is fun — you're prompting, code is being written, you're thinking, 'holy moly, I'm an engineer.' But then you've got to find a place to deploy it, you've got to maintain it, you've got to make it better. I love people being excited about building things, I think it's really cool. But there's a little bit of, yep, I've seen this act before — if you build it, you're going to have to maintain it, and now you run the risk of: is there something better out there? Is there something getting better out there that we need to look at? Sorry, I didn't mean to cut you off — I just love that. Damn, man.
Justin Fredlender
It's like it takes some experience to know that it's not always as easy as it seems.
Bryan Mahoney
Definitely.
Justin Fredlender
And I'll go one step further — it's not just maintaining it, it's even finding a steward to own it within the organization. And yeah, even with onboarding Chord, as an independent consultant it was even hard to figure out in my own head: okay, who is the internal steward making sure the investment in this platform actually pays back? Making sure we're utilizing the data to make decisions, that the investment is worthwhile. If that in itself is a hard question to answer, then building it yourself, and also having the steward, and also maintaining that technology — that's almost a non-starter to begin with.
Bryan Mahoney
Yeah, yeah. It's just like — over the years I've tried to develop this analogy: how can you get something that's 'out of the box'? Because the promise of out-of-the-box is really tempting. But I don't want to be put into the box — am I willing to take something that works really well, but I only get 75% of what I want? So I try to think the right way of looking at it is: can I get something out of the box, where that box doesn't box me in? Where I can get like 80% really quickly, but then I can extend it with the capabilities I have in the organization — the same capabilities that maybe could have built it from 0 to 80, but definitely have the capabilities to take it from 80 to 100.
Justin Fredlender
Yep.
Bryan Mahoney
And I think the challenge for companies like mine should be: how do I go from 80 to 90? I don't ever want to pretend I can do 100% of what all the different brands might want to use the platform for. But I'd love to get to a place where we can do 90, where we make the last 10% really fun and easy, where people feel like they're building on top of it. I tend to think that's where we're going. I've been trying to answer the build-versus-buy question as: actually, it's buy, then build. And if you can't build around it, to me that's a smell — that means you're buying a legacy platform that doesn't have APIs, and that's going to be a pain in the ass, frankly, to work with. If you can buy something that you can build around, that starts to feel modern to me. I want to encourage more people to move in that direction — be curious, build with AI, but be really careful about the decision to take it all on your shoulders just because you can build it and launch it. To your point, do we have a steward within the organization to champion it and maintain it? Because there are costs there too.
Justin Fredlender
Definitely. And the last thing I'll say on the topic is: if you're tasking somebody internally — unless you have a bunch of redundancy at the organization, which I think, over the last three to four years, most companies have tried to reduce as much as possible — that means if you're going to do a build, there either needs to be a dedicated resource, or somebody is going to be doing it off the side of their desk. And if they're doing it off the side of their desk, you have to ask yourself: do they actually have the bandwidth to maintain the system, or whatever it is they've built, or is it going to be forgotten and just become a waste of time?
Bryan Mahoney
Yeah.
Bryan Mahoney
I bring it back to: why do brands get created in the first place? Is it to build commerce infrastructure, or can I take that side-of-desk time and invest it in customer-facing features that let me dive into insights, or react to those insights, and bring me closer to the customer? So I think there's a tension there. I'll try to steer us — I can't believe I'm going to do this — but maybe more in the direction of AI, because I think there's an awful lot that's brand new, and an awful lot we can learn. There's obviously a ton of power there, but I think there's still an awful lot of uncertainty, or even misunderstanding, about how we might interface with an LLM — even for someone with an analytical background and proximity to data stacks. I'll share an anecdote. I was on a discovery call yesterday with a data leader at a massive organization — digitally native, now well into ten figures of revenue. We were talking about build versus buy, and this company clearly has the capabilities to build everything internally if it wants to — very profitable. Their decision, and they're still evaluating it, to build an AI solution to get more out of their data is because they don't want to give the data away, they don't want to give it to the LLMs. So theoretically, in principle, I think they're right. But to me there's sort of a misunderstanding — as a technologist, how we might want to leverage AI, where I do want to keep our data safe. I don't want members of my team downloading PII and uploading it to ChatGPT and asking it to be summarized — that's dangerous, you shouldn't do that. But there are ways to build context, to build frameworks, to ask the LLMs to help write better SQL so I can get answers faster without giving data away. I think there's an awful lot of opportunity for us as tech leaders in this space to evangelize an approach to keeping customer data safe, without forcing yourself into perhaps the wrong path of building it all yourself in the name of data safety. So I'm just curious — across all the different brands you work with, how are you seeing them use AI? Are you seeing efficiencies being unlocked? Are you seeing a lot of promise where they end up stubbing their toes and putting it in the top drawer for now? I'm just curious where you think most of these brands are on that AI adoption curve.
Justin Fredlender
I think, for the most part, everyone is adopting AI — that's a no-doubter. And I think pretty much everyone is, unless you're at a much larger brand with an internal resource — whether a dedicated AI resource, or just someone really passionate about AI who can communicate across the business how they're using it. Most companies are probably somewhere between 50 and 75% — they're all using it, but are they using it to the full 100%? No. Are they not using it at all? No. It's generally somewhere in between. In some functions it's more evolved than in others. Data is a great example, because I think data is something LLMs just inherently do well. They analyze things and come back with their own point of view, and at least help get you 75% of the way to the answer you're looking for.
Bryan Mahoney
Is it much faster than without it?
Justin Fredlender
Not only that, but also, there are generally very few power users of BI tools within an organization. This is another struggle we were talking about, either earlier this week or last week — there's an example of a resource at one of the companies I work with who's overseeing a function that probably should be very data-proficient. It's just not where they want to spend their time. It's not what they're passionate about. So asking them to mine data in a platform like Looker, or within Chord, is a really tall ask.
Bryan Mahoney
Yeah.
Justin Fredlender
But if you just say, 'hey, prompt it, get 75% of what you're looking for, and then come to me and I'll help you get the rest of the way' — that's a much easier ask.
Bryan Mahoney
So yeah, it becomes more natural.
Justin Fredlender
A more natural workflow — you can make more users within a company a power user of a data platform, or at least a semi-power user, than ever before.
Bryan Mahoney
Yeah.
Justin Fredlender
Now I want to revert back — because that's on the analytics side.
Bryan Mahoney
I love that, though — that's a really powerful way you should be thinking about using AI today: you can create more data power users than ever before. If you think about being on the data side — and I'm still on the data side — we would spend all this time setting up these beautiful Looker dashboards, and then I'd see these questions come into Slack, and it would break my heart. There's a dashboard for that, we democratized access to data — why don't you just go look at it? I kept banging my head against the wall before finally realizing: yes, the data is there, and yes, we made it available to them, but there's something about it that's too hard — it's sort of human nature. So instead of looking at it, they just ask the question — in Slack, over email, or they shout it across the room. That's what's happened, and I think it's pulled the data teams apart from the rest of their business teams, because they're like, 'we did all this work, use it.' But to what you're talking about now — if you do all that work and you have this trustworthy data foundation, now we can start introducing more of a conversational interface, which is what they were doing anyway, just answering the question in Slack. But if that answer can come back 75% of the way there, well — now we can have a more robust conversation. Because now I'm actually asking the analysts to do the work the analyst has always wanted to do, which is the third-level-down question — I want to go find that insight, the thing that can move the business forward. That, to me, is where we're right there now, and I want to see more of that happening within organizations — willing to let go of the hard work we've done forever building beautiful dashboards and saying 'you can self-serve.' Because I do think traditional BI is broken. I think it's too hard. And I'm really excited about what AI can do to enable more conversations within organizations around the data.
Justin Fredlender
Yeah. And I think specifically within the realm of analytics and data, AI is probably further along than in some other areas. If I'm thinking about AI within an organization — specifically in our case, a commerce or e-commerce organization — yes, you have analytics, but you also have the ability to leverage AI to do something like creative.
Bryan Mahoney
Yeah.
Justin Fredlender
And that's where I actually think there's been awesome advancements. But I think AI from a creative perspective is not where AI is from a data perspective, because the capabilities of an LLM, from what I understand, lend themselves better to analyzing data sets than to creating content that a CEO, founder, or whoever would feel good about putting out.
Bryan Mahoney
Would feel good about putting out.
Justin Fredlender
That's exactly right. So where, on the analytics side, you're creating more semi-power users—
Bryan Mahoney
Yeah.
Justin Fredlender
—and I actually think that in the next six to 12 months they'll be straight-up power users as AI advances. But I'd say on the creative side, what you're actually doing now is probably enabling a team.
Bryan Mahoney
Yeah, a creative team.
Justin Fredlender
You're not — yes, of course, your media buyer can maybe create some ads, but if they put those ads in front of the CMO, or an executive who's approving the creative, they're probably going to get a lot of pushback on whatever they've produced. Whereas if you put an AI tool in the hands of a creative who actually understands, who has a vision of what they're trying to produce, that will probably pass the sniff test. So I think there are areas of AI that enable an entire organization — and in your case, democratize data — but there are also areas of AI that just enable a singular person who's a professional at a certain function, in this case creative. I want to extend to another part of commerce when it comes to AI, because I think we're talking a lot about the internal efficiencies we're driving through advancements in technology, in this case AI. But what I still don't hear people talk about enough in the world of commerce is how this actually changes how consumers shop.
Bryan Mahoney
Yeah. Oh man, I'm going to have to have you come back so we can unpack that — that's another hour-long episode, talking about how we think about content and how people are going to shop, and the human behavior in between. I still want to have a connection with brands I love, as opposed to having a number of agents running around the internet shopping for me. But you're absolutely right — strategically, that's going to change the way we, as a brand, need to think about how we're showing up in the world, and who we're showing up for. So maybe I'll get you to come back before long and we'll unpack that too, because I promised to get us out at a certain hour. I want to keep us to that promise. So what I'd say is: if you're a brand out there and you're curious about how to get more out of AI with the tools you have today, how to get people focused in the right area — it seems like you've developed a really unique point of view. You're working with a number of great brands, so I'd encourage everyone to reach out to Justin. He's a power user, and he's out on a mission to create more power users within organizations. So thanks for joining me today on the pod. Excited to keep the conversation going about changing behaviors in shopping.
Justin Fredlender
Appreciate it, Bryan.
Bryan Mahoney
Thanks so much for joining me.
Justin Fredlender
Yeah, thanks.