Growth by Sabir: Fix the Foundation Before You Scale With AI

[00:00:00] Sabir Semerkant: When you scale garbage, you have a massive garbage dump. That's what you have. Scaling problems give you more problems. You need to fix the problems first, the fundamental problems first and your foundational problems with your business.

[00:00:15] Kailin Noivo: Quick one before we dig in: if you run an online store, we've opened up something just for Ecommerce Toolbox listeners. A free trial of Noibu's AI agents. They find what's slow, broken, confusing or leaking revenue on your store. Build the change, and nothing ships until you approve it. Grab it at noibu.com/podcast-audit. Okay, back to the episode.

[00:00:35] Kailin Noivo: Super excited for our episode today with Sabir. Sabir goes through a lot of the tactics and tricks that he used at VaynerCommerce and also how he developed his own program around this. Really excited for today's episode.

[00:00:41] Kailin Noivo: Welcome to another episode of The Ecommerce Toolbox: AI in Retail. Joining us today, we have Sabir Semerkant. And Sabir, tell us a bit about yourself. Welcome to the show.

[00:00:56] Sabir Semerkant: So, it's been 25 years, so I'm going to keep it as simple as possible. Right? Before Shopify existed, since the birth of ecomm I have been in ecomm. So, that's how long I've been in it. You know, in the middle of my career, I co-founded VaynerCommerce with Gary Vee, a great guy. He's a brother from another mother. Now, brands come to me. I run a coaching program called Rapid 2X. It's live coaching. It's not a digital course where, you know, we solve growth problems with ecommerce DTC businesses.

[00:01:31] Kailin Noivo: Very cool. So, it sounds like you've had a lot. You've also been an ecomm operator, right? I saw you used to work at a couple of different brands as well.

[00:01:38] Sabir Semerkant: Many. One of them is, I mean, just a few examples: Vitamin Shoppe. When I was there, I was brought in because vitaminshoppe.com had gone public. It went bankrupt, and then they bought the assets from a fire sale from the courts, you know, so they brought me in to put it back together. Before it went bankrupt, it was roughly around an 8- to 12-million-dollar business. You know, this is, just to kind of date it, Google was at Stanford and beta testing. I grew that at that time from, let's call it, $12 million, take the high-end number, to $52 million in four and a half years. And that gave me quite a bit of learning about what works, you know, and then I said, “Okay, let me make sure that this is not a fluke. Let me keep on repeating this.” And I kept on repeating the same principles, and that became Rapid 2X that I documented in my book. I did that with Ashley Stewart. I did it with Perry Ellis, Tommy Hilfiger, so many brands. And now I have scaled that into the Rapid 2X Protocol, where 200-plus brands have gone through the program so far, with an average of a 108% increase in revenue, profitably, in the first 12 weeks of the program. I mean, that's called, like, sharpening your tool, you know.

[00:03:01] Kailin Noivo: No, for sure. And we're going to get into the how, and mainly on AI agents, which is what I really want to speak to you about. But this is definitely not in the show notes. But Ashley Stewart's a customer. Did you work with Lori at all when you were there?

[00:03:15] Sabir Semerkant: No. So, Ashley Stewart, unfortunately, one of the other skill sets that I have is that I take bankrupt brands and revive them, you know, and Ashley Stewart, in their first bankruptcy, went bankrupt. And actually, the CFO from Vitamin Shoppe, who had seen me in action, called me and said, "Hey, I'm now with the CFO of Ashley Stewart. You know, I want to talk to you. I'm at this bankrupt brand. It just got bought out. You know, we have new owners. I would like to talk to you to see if you would be interested in joining.” Right? So, when I joined, it was less than $3 million in ecomm revenue. It was a very nascent channel for them. In two years, it went to $30 million. And then the CFO knew that I had the skill set to kind of revive brands like that.

[00:04:08] Kailin Noivo: Very cool. All right, let's get into the era of AI. Obviously, AI has become super mainstream in the last 6 to 12 months, especially with Shopify stores. So, very, very open-ended question: how are you seeing the fastest-growing DTC businesses implement AI agents tactically? What types of AI agents? Which ones are working? Which ones are not working?

[00:04:29] Sabir Semerkant: So, in our program at Rapid 2X, we do a combination. The entire program is AI-based, by the way. So, we use both Claude Code and Codex, Opus 5.5. Sonnet; it is a mix of models and also GPT-6.1 Sol because there is no one perfect model out there, you know. And if anyone just signs up for just Claude Code and thinks that it will do all the job, it will have massive amounts of holes that GPT-6.1 Sol actually picks up. So, we do a mix of models, and that's in our workflows and everything in Rapid 2X, actually, I'm the one, I author all of the workflows and I give that to my clients in the program and they implement that. And those agents keep on reiterating.

[00:05:19] Kailin Noivo: Amazing. What are some of those agents? Like, what's the first thing you go after? Is it ROAS? Is it like customer service? Like, what are those first tactical ones that you drop in?

[00:05:31] Sabir Semerkant: All right, so number one, why do they come to me? So, let's talk about that. Right? Because that's the root of how they come to me, right? So, the brands come to me thinking, “Traffic is the problem.” Sabir, can we improve our ROAS? Even your question was about ROAS, right? Can we improve our ROAS? But then I look at the business because I'm very seasoned at this. And literally within the first 10 seconds, I can tell you where you're leaking. You know, your business is leaking revenue, and you're also leaking ad dollars. It's very obvious to me. Right? And so the first thing that we tackle is optimizing the site to consumer attention span, which, in 2026, if your pages are not loading or the first interaction is not within the first two and a half seconds - ideally, 1.7 seconds - you are 100% losing conversion rate. You have a very high bounce rate. And derivatively, you're spending money on Meta ads and Google Ads. You're definitely losing money on ROAS.

[00:06:33] Kailin Noivo: Okay, so you think the first area to go after is effectively CRO, like on-site content load, like…

[00:06:41] Sabir Semerkant: No. Site performance. The very first thing.

[00:06:41] Kailin Noivo: Okay. So just straight up, like speed.

[00:06:46] Sabir Semerkant: Speed is your number one enemy when it comes to… And speed, not for the desktop, speed on your mobile device, because 80% to 90% of your consumers are using iPhones if you're in the United States. If you are worldwide, you should also optimize for Samsung, Galaxy and Android devices. That's the device. And if your designer, in their Zoom meeting, shows their beautiful Mac desktop, where the desktop side looks amazing and beautiful, nobody cares. Nobody cares, right? That's just, you are literally focusing on 10% of your traffic or 20% of your traffic, best-case scenario.

[00:07:28] Kailin Noivo: Okay. No, that makes a lot of sense. And have you seen, like, landing pages? Like, has AI kind of started to bring that backwards a little bit? Like, people vibe coding, like, landing pages and stuff like that. Like, are you seeing that problem start to come back up?

[00:07:42] Sabir Semerkant: That's where you arm, you know, because AI agents like Claude Code and Codex allow you to tool them, right? You can give them a tool to use. So, if you connect that to… My preferred tool is GTmetrix, right? So, I connect GTmetrix, and they have an MCP and an API. I connect GTmetrix to a tool so that, as part of one of the test cases, it has to pass for the site's certain KPIs, like time to interact, which is as soon as the consumer can start playing with the thing that you're showing them. It needs to meet that number. The ideal place is 1.7, but if it gets a little bit higher, all the way up to 2.5, it's forgivable; you know, it will be fine. And if it doesn't meet that, then it has failed the test. So, you don't take that theme page or product detail page; you don't take it live. You stage it first in a staging theme on Shopify, for example, and then you production it after it passes those tests.

[00:08:47] Kailin Noivo: Quick break from the conversation. Almost every operator we talk to is being asked for the same three things. Grow revenue, spend less on outside help and actually get value out of AI, all with the team they already have. That's why we built Noibu's AI agents: CRO, A/B testing, performance, accessibility, ROAS, and bug resolution. They start by auditing your store and ranking the biggest opportunities. Then they build the change, wait for your approval, ship it and measure what it did. At Totally Bamboo, one person runs all of D2C across more than 600 SKUs. A collection page filter tested off their own shopper data converted 24% better. As a Toolbox listener, you get an exclusive free trial of our AI agents, and it runs until you have at least two changes live and measured. Go to noibu.com/podcast-audit. Now back to the conversation.

[00:09:39] Kailin Noivo: Interesting. So, performance is the first place you look at. Where do you go after that?

[00:09:42] Sabir Semerkant: So, once that's done, next is people, again, going back to what I said about, “Oh, I need more traffic.” Right? No, you're getting traffic. Right? You're getting 10,000 visitors a month, a week, every day, depending on the size of the brand. You know, what are you doing with that, right? So, the very next thing I look at is your customers. The customers you have already acquired, whether you have been in business for a year or you have been in business for 25 years, it doesn't matter, right? 25 years, more rich data. One year, less data. It's okay. But we do a life cycle analysis of RFM, which is recency, frequency, and monetary value. Right? How recently did they buy? How many customers recently bought in different buckets? Right? 0 to 90 days, 91 to 180 days and so on. Right? You know, frequency is how many of them are one-hit wonders? Meaning, if you remember those 90s songs, you know, even songs like Macarena, it was a hit one time or Gangnam Style, right? It was a hit one time, and then it died, right? There are customers. There are one-hit wonders. And most Shopify businesses, here's the scary stat for you: roughly on average, 90% of your customers are one-hit wonders. That makes me think you don't have a brand; you have a transactional business. Literally and symptomatically, you are buying transactions from Meta ads. That's all you're doing, right? You're not acquiring customers. So, what you think when you're telling me, "Oh, my customer acquisition cost is this.” No, that's not your customer acquisition cost; it's your order cost. That's all. Because 90% of your customers are one-time transactions, and only, you know, between the numbers ranges from 6% to 12% are more than one transaction. And then the M stands for the monetary value. We look at the average order. We look at 12-month purchase and lifetime purchase, which is lifetime value, in order to see how much we should be investing in that. But as we're going through that, like, we solve the site performance problem first, then we run campaigns for the RFM life cycle to see the people you have acquired; did they care about your brand, even, right? So, let's give them some sort of what I love to call a Hail Mary campaign. Right? You know what that is? You know, in American football terms, you know, you're down, and there's this one last throw; that's your Hail Mary.

[00:12:30] Kailin Noivo: 90% off.

[00:12:32] Sabir Semerkant: It's not 90% off. You know, so it's like buy one, get one 50% off or 40% off or 50% off, something like that. But we segment it because we are studying the segments. I don't care about the offer working or not. I care about activating and reactivating customers that I already have or even prospects. People who just signed up because you did a pop-up and they gave you their email. Is it really worth it? And down the line, it is costing you, right? It is costing you deliverability. It's costing you; Klaviyo is charging you by profile, right? So, if you have 100,000 profiles but only like 5,000 of them are really good, you're paying for the other 95,000, you know, so it's just not worth it. So, there are things like that. As we're going through this, we are also optimizing your expenses, too. We're cutting things down that your business doesn't need. You know, less is more is more true in ecommerce than anything, right? You know, you don't need 25 Shopify apps to do what? Oh yeah, I hired some guy three years ago, and he implemented this. He said it was good. Okay. Did you look at the analytics to see if it's, you know, helping you or hurting you? When we did the site performance, we found out that that was one of the culprits that was slowing the site down. But is it benefiting you? If the answer is yes, maybe we could either tune it or we need to replace it. Same thing. And in the second path, when we're doing the customer analysis and customer optimization, we're studying that part of it to see, okay, now what do we need to take back in that scenario? We have had brands that had a rude awakening that they had a garbage of a CRM, basically. Right? The customers they had acquired were literally orders. That's it. Right. And that's an extremely bad problem because then you don't have a business. Then you are heavily reliant, 100% of the time; you just keep on spending money on ads in order to keep on getting more orders. And it just becomes addictive. Right? And you're not making margin. As soon as you slow down that payment to Meta or Google, your business tanks.

[00:14:43] Kailin Noivo: Tell me a bit about, like, what's a good returning customer rate in your opinion?

[00:14:48] Sabir Semerkant: So, it's the journey. It's every part of the journey, right? Zero orders; they just signed up on your site. What is your first conversion rate from 0 to 1 frequency? Because you want to get them to buy the first time, right? Then, because we know that this 90% of your database is one-time buyers, how can we get them to move to the second, third and fourth and then, as more advanced topics come into existence, like a loyalty program? Right? You don't want to install a loyalty program like Smile.io, right? That's a very popular one. Right? You don't want to install it when you have a 90% problem with one-hit wonders. Because then you will think that, oh, we did a loyalty program; it didn't work. No, the loyalty program was a later problem. You tried fixing an advanced. It's like a sixth grader trying to take a college-level physics class. It doesn't make sense. It just doesn't make sense at all.

[00:15:51] Kailin Noivo: That's really interesting. Tell me a bit about where you're seeing companies hurt themselves with AI? Because I think AI does so many things so well. Like, where are you seeing some immediate like pitfalls that people are putting themselves in?

[00:16:06] Sabir Semerkant: When you scale garbage, you have a massive garbage dump. That's what you have. So, when you have, like, these two examples I just shared with you, right? When your site has bad problems, like with site performance, and now you're vibe coding new features, you're just scaling that problem. When you are in acquisition, and your customer life cycle has a problem, and you're scaling it by saying, “Oh wow, look at what this Opus can now do videos.” You've seen a lot of YouTube videos on that, right? Oh, let me do some videos on Opus. And then you start scaling that and go like, okay, let me put some more money on it and let me do that. Now, you're scaling your budget. That is not productive. Scaling problems gives you more problems. You need to fix the problems first, the fundamental problems first, and your foundational problems with your business before you start scaling them. Otherwise, it becomes a nightmare. And the nightmare, unfortunately, on the other side of it, at the beginning, when I was giving you my bio, I said that I've revived, unfortunately, a lot of bankrupt companies. I can tell you where that leads. It leads to bankruptcy. There are so many brands in Q4; they report amazing numbers that they are doing so well, and come January, you see bankruptcy filings go up.

[00:17:26] Kailin Noivo: So, you think the fundamentals. So, call it, like, is the site fast? Obviously. Is there any like theme issues? Like, is there stuff that's breaking? Like, does app A conflict with app B? And, like, it breaks, I don't know, some mobile journey, or it slows it down. Like, you think nailing the fundamentals is, like, the most important thing versus, like, adding on random AI features that compound the debt.

[00:17:50] Sabir Semerkant: Oh yeah, yeah, definitely. I mean, it just keeps on going. Like, it's not just site performance and then customers; then you move on to product. Look at the product. So many people have invested so much in inventory, and they came up with 16 variations because your competitor had 16 different colors. “Why did you do it?” “Oh, because our competition has it.” “Okay. How is that inventory doing with the other 14 variations of colors that you introduced for every SKU?” “Oh, it's not doing so well. That's why we are coming to you.” “Did you not realize that your competition that introduced those 16 colors, maybe, that guy who came up with it got fired for a bad inventory decision?” There are things like that. Like, it's very obvious that, for me, fundamentally, you need to fix the product in that sense. Right? And in most cases, I mean, there are so many examples of it. For one of my nail polish brands that I had worked with in California, I actually literally walked through their warehouse, and I pointed towards, you know, a set of boxes that were sitting there. My question was very different. The founder misunderstood me. I said, “What is this inventory?” “Oh, it has this thing in there.” But like, “Oh, how is it doing?” “Oh, I guess I can pull up the report.” I said, “I don't need to pull up the report to see that it's collecting dust in the warehouse.” Right? I don't need to know. I don't need to have a nice Google dashboard, you know, a pretty dashboard to tell me that you have a problem. This section of your warehouse has not even been visited. You have not even cleaned it. So, I can tell you that this has been sitting here for at least 2, 3 years, right? We need to do something about this. And then, after my walkthrough, literally, what we did was we identified all of the inventory that needed to have a flash sale in order for us to take that money that was literally burning, sitting in the warehouse, taking up space, and square footage costs money in a warehouse, right? Literally, we need to take that and flash-sale it. We want to collect the cash so we can deploy it towards your top 10 bestsellers. You are making your best sellers suffer because you don't have cash flow, and you have this garbage sitting in here. And unfortunately, by the way, that stat follows the 80-20 rule. 20% of your inventory does 80% of your revenue profitably, right? And then 80% of your inventory is eating up space, cost, everything, including your time. Because now you're thinking about, “How do I sell this thing?” Because it's been sitting there, and I don't have cash. I need to turn it into cash. But unfortunately, some parts of that 80% could be revived because it didn't have a proper product campaign, product launch and stuff. But some part of it, a good part of it, needs to be just getting rid of, like, you need to TJ Maxx that thing or TK Maxx it, if you understand what I mean. Right? You need to do that in order to get your cash. Because that cash, oh, it's $50,000. It’s $100,000. Are you serious? Yes, because I can take that $50,000. What's your ROAS right now on Meta ads? Oh, it's 3. That $50,000 can be $150,000 in revenue from your bestsellers. That's it. And it's been sitting here and doing that. We move from product to a brand. When I say brand, brand storytelling, right? A lot of founders think that pretty legends on a product page about their return policy or that it’s 100% silk is good enough. No. Why should I care as a consumer about your brand? Why did you come up with it? They're missing that part of the story. The entire thing is missing. And when that's missing, everything that feeds that - any content that that feeds it: blog posts, Meta ads, your social posts - everything is wrong. You don't have that foundation in there. Then you're doing just surface-level stuff that you're not seeing results for. And you go, like, we don't get engagement on our Instagram. Yeah. You're missing the fundamentals of brand story building that you have not done.

[00:22:07] Kailin Noivo: No. This is really, really insightful. So, bringing it back to some of the AI front. So, I mean, obviously, AI doesn't help with what you just described, right, in some cases.

[00:22:19] Sabir Semerkant: No, it does. It does. Yeah. So, if you have the right workflows, that's where the guidance comes in. So, AI is amazing when you have someone like me, right, guiding it. Right? So, in the Rapid 2X Program, everything that I just said - like, these are just the 4 parts of the program. We go through 9 in a year, right? It takes us about 6 weeks each to work on each one of them. And everybody works on it. And once we are done, we're pretty much done. You just continuously do improvements after that. And it compounds. Everything I'm talking about is compounded growth. It's not a one-hit wonder. Like, you just do one ad, and it performs, and next time you try to do the same ad, it doesn't perform. You know, it just keeps on compounding. What AI is missing is that context, that intelligence, somebody who can steer it to say that this is what good looks like or what best looks like. When you have that in place, then the output, AI is amazing. I have a glove brand right now. She is a product designer, a fashion person with zero technical skill set, right? She is managing these AI coding agents now, you know, because she sees that, and she's running like 5 different projects at the same time. You know, that's how effective and efficient you become when you're giving AI the right context, the right workflows, telling it what correct looks like, what best looks like. Like, every task that AI works on, it has to go through what I call evals, right? Evals means that, okay, this is my problem, this is what I need you to solve. What does success look like? And these are the exact parameters you need to hit in order to say that I ran through these 50 tests and this task is completed from my perspective. And then you feed that to the other agent to say, “Oh, this guy says that they did the job. Can you find any gaps that they missed?” And then the other one finds it. Like, Sol finds Opus's mistakes, right, and then tells Opus, “Oh, yeah, you missed these details, these edge cases.” And Sol is amazing at edge cases, you know. “You missed these edge cases. Go and fix them.” Once it fixes it, we do three rounds of comes back and gives it for re-audit. And when you go through that, then you don't have the typical problems that most people see nowadays, right? They go like, "Oh, AI didn't work for me, or it gives me the wrong answer.” An example of that is when they do analysis and insights and stuff, right? Did you know that AI, just like human beings, does a lot of mental math? So, when it comes back and goes like, “Oh, 156 out of 1000 customers, blah blah, blah, blah, blah.” Some of it, it may have done. You say, “Are you sure?” It goes like, “Oh, I got that wrong. I miscalculated. That real number is 178.”

[00:25:25] Kailin Noivo: It needs guardrails for sure. And like, you need to build that systematically with agents, where, like, you have an agent check another agent's work, and that gets complex. Like, how has AI, in your opinion, changed how you recommend to your clients that they interact with traditional agencies? And what I mean by that is, like, the 40-hour-a-month retainer agency that does, like, maintenance, CRO, maybe some marketing. Like, how does AI change that in your opinion? And how are you now, like, has that changed how you advise your clients to interact with them?

[00:25:59] Sabir Semerkant: So, here's the scary part, right? Unfortunately, because I've dealt with a lot of agencies throughout my career. Unfortunately, I believe that 80% to 90% of agencies are going to die because they were doing commodity work. They were not doing high-intelligence, high-quality specialty work. Right? Did they get away with it? The answer is yes, because, you know, the bigger agencies would not take you on if you were a smaller brand. So, unfortunately, smaller brands or even medium-sized brands got hurt because of that. Right? So, that's one part of it. Because now AI, with a program like Rapid 2X, right, can replace your agencies. And we have had several founders in the program literally replace freelancers and agencies and stuff like that they didn't need. They just didn't have any need for it. Now the people who are still in business and who are not going to be suffering are the ones that had real experience, real quality. They really understand, people like Gary Vee and VaynerMedia. Right? They understand branding really strongly because they've been doing it. They've been running thousands of tests. They're really good at that. You know, platform optimization, like Meta and Google and so on and so forth. Right? Those kinds of agencies are going to thrive in the age of AI. The ones that were doing remedial work and commodity work, those are going to fall through. Right? So, the founders, once they become really efficient in our program, at least you know that they see that they can do most of it themselves because they have that feedback cycle a lot quicker by talking to an AI with the guardrails. Right? They can get the results a lot quicker than waiting on a weekly call. “Oh, next Monday I'll talk to my agency to see what happened.” Right?

[00:27:52] Kailin Noivo: Yeah. You open up a Jira ticket, and they charge you to do a Jira ticket. No, it's very antiquated. No, it makes sense. Well, Sabir, this was awesome. I really appreciate you taking the time to do our show, and this was really, really valuable. So, yeah, thank you. Thank you so much for hopping on.

[00:28:09] Sabir Semerkant: I do have a gift for your audience. I would say that if they go to growthbysabir.com/book, I have the Rapid 2X Protocol book complimentary to your audience; they can go and just grab it.

[00:28:21] Kailin Noivo: Amazing. Okay, cool. Thanks so much, Sabir. It’s very generous of you.

[00:28:24] Sabir Semerkant: Thank you for having me.

[00:28:25] Kailin Noivo: Thank you.

[00:28:25] Kailin Noivo: Before you go, if today's conversation has you thinking about what you'd change on your own store, let our agents take the first pass. Until now, a store only got better as fast as the people working on it. That limit's gone. Our agents find what's slow, broken, confusing or leaking revenue, build the change and hand it back for your approval. Toolbox listeners get an exclusive free trial that runs until you have two changes, live and measured. Head to noibu.com/podcast-audit. That's n-o-i-b-u.com/podcast-audit. Thanks for listening. See you next Wednesday.

[00:28:58] Outro: The Ecommerce Toolbox: AI in Retail is brought to you by Noibu. To find out more about Noibu and how we unify error monitoring, site performance and experience analytics to uncover growth opportunities and skyrocket your revenue, visit www.noibu.com. That's n-o-i-b-u.com. And then make sure to search for The Ecommerce Toolbox: AI and Retail on Apple Podcasts, Spotify or anywhere else podcasts are found and click subscribe so you don't miss out on any future episodes. On behalf of the team here at Noibu, thanks for listening.