Test Your Ads on Simulated Customers Before You Spend a Dollar This BFCM

Steve Hutt: Welcome back to eCommerce Fastlane. I’m your host, Steve Hutt. Today’s conversation is timely because it’s all about getting more out of every dollar heading into Black Friday and Cyber Monday. If you’re listening right now and seeing customer acquisition costs climb, I think part of the reason is that more brands are crowding into the same feeds. For most brands, the largest ad budget of the year is coming up soon. There’s also a code freeze coming with Shopify, and a lot of apps and brands are saying, “We’re not doing anything to our theme other than promotional changes.” That’s weeks away, but I still think there’s an incredible opportunity right now. That’s why my guest today is Tamera, founder and CEO of Clarvos, C-L-A-R-V-O-S. I did a little research before recording, and she’s spent the last 15 or 16 years in tech, leading sales and marketing teams. What she’s building now is interesting. It’s about scoring your targeting and creative against simulated customers before you spend money, then launching and optimizing across channels from within the platform. It’s cool that you can plan ahead before spending your money. We’re spending a lot on ads to drive net-new revenue, so I’m hoping we can unpack what this looks like in practice, what early-access brands have seen, and why it’s different from prompting ChatGPT or Claude yourself. A lot of people think, “I’ll just ask Claude about this ad campaign and give it a folder to review.” It doesn’t really work like that. You tend to get more generic answers, whereas this is focused on what you’re trying to do. I’m excited to unpack all of this today. Hi, Tamera. Welcome to the show.

Tamera Montagna: Thank you so much for having me on your show today. I’m very excited to talk about our new platform and how it may help your listeners and Shopify merchants optimize for the holiday season coming up.

Steve Hutt: Let’s talk about your 16 or so years leading sales and marketing teams. What made you decide more recently that you wanted to build Clarvos? I want to make sure I have the pronunciation correct, too.

Tamera Montagna: Yes, “clair-vois” is how we decided to pronounce it, but we played around with it, too.

Steve Hutt: All right, got it.

Tamera Montagna: I’ve always worked with small businesses. Even when I worked with large companies, Fortune 100 companies, I was always working with small businesses within those organizations. But I would say my more recent experience with a company called Evolve24 helped me understand the needs of small businesses trying to acquire new customers, especially through social media, whether organic or paid. I saw how hard it is to compete against the largest companies in the world, and the team and I wanted to find a solution. At the time, I was leading Evolve24, a data science and professional services company. We were essentially using social media and billions of documents a day to help large companies determine strategy based on society’s values, attitudes, wants, needs, and interests. That’s how I became focused on consumer segments and how to find them for companies so they can grow. It started as a strategic focus and became a growth mission for me. While Evolve24 focused on the largest companies in the world, I wanted to solve this for small companies. I wanted to make everything we were giving large companies available to challenger brands and enable them to compete. That’s how Clarvos came about.

Steve Hutt: That’s amazing. I want to unpack the uniqueness of the platform because I love it when people solve unique problems. You could pull back a bit and say that Meta and Google already do some of the things you do for free within their own ads managers. Some people work strictly that way, and there’s a certain amount of baked-in trust. Meta has Advantage+, with targeting, creative opportunities, brand voice, and performance features. Google has something similar, although it works a little differently. But I get that Clarvos is solving a gap. Part of that gap is testing before you spend, which is not what Meta and Google do. They learn by spending your money on live tests, which is completely different from how Clarvos does it. Can you talk about that gap? It’s a big differentiator.

Tamera Montagna: It is. It’s one of the things we find so fun to work on internally. Before we talk about testing, I want to explain how it came about. We were looking at developing consumer segments specifically for direct-to-consumer brands, and we wanted to cut across demographics. Large companies can afford to advertise to Gen Z or millennials. They can advertise to large demographic groups. For our challenger brands, we wanted to cut through those demographics. Regardless of somebody’s age or gender, we wanted to find people who shared values, had similar interests, and cared about similar things. We grouped people by those values. We have 1,300 consumer segments using billions of documents a day, and we’ve created segments with similarities within them. Some of our segments have a large age range, but what brings those people together is what they’re interested in. Then we match consumer segments to a brand. Of course, it’s an algorithm, but what that means is we look for consumer segments that would be interested in buying from a brand if they knew it existed. Their values and what they’re looking for are similar. We align consumer segments with brands and tell our customers, our merchants, which segments are most likely to buy from them. That enables us to make the audience smaller, more affordable to advertise to, and more likely to buy. We didn’t stop there. We also needed to understand how to build creative that was further targeted to that consumer segment. We wanted to take everything we knew about the segment and build creative that would appeal to them. We built a platform with agents that create ads that speak to that targeted audience. Then we thought, “That’s great, but we need to prove that this ad would interest this consumer segment.” That’s when it got really interesting. We built what we call a customer simulator. We create 50 simulated people for each consumer segment who represent the attributes and characteristics within it. They have first, last, and middle names. They have families. They went to college or didn’t, depending on their educational background. Their profiles tell you what their jobs are, what they like and dislike, how they spend their time, and what they do in their free time. In fact, 3% of our simulated people have cancer. They represent the consumer segment they’re part of. Then we can take our creative and test it among those 50 simulated people, get feedback, and change the creative. All of that happens on our platform. We might build 50 creatives, and in the end, we show the ones that received the highest engagement scores from that simulated customer group.

Steve Hutt: Wow. That’s very technical. I understand how you’ve broken down these consumer segments, but how do you choose winning creative based on the simulations? You’re not actually spending money yet. You’re making educated decisions about which creative you believe will be impactful for a particular campaign. I’m curious about the methodology.

Tamera Montagna: We have what we call judging LLMs and judging agents. This is an AI platform, and we have 16 agents that build those creatives. They base them on everything they can find about the brand. There’s an interactive onboarding process where our customers or Shopify merchants go through a chat-like experience. We take all the information we can find online, and the merchant enters any additional information they want to add. Through that process, we gather and synthesize information from those 16 agents to develop the consumer segments. We create ads based on what we know about your products, what we see on your website and in your shop, and the characteristics of the consumer segment. The agents actually build the creative.

Steve Hutt: That’s amazing. What I find interesting is that you have more of a neutral view. When you think about Meta on its own, I think there’s some bias in its recommendations. The pitch is often to push for more spend on Meta. It’s not necessarily going to say, “We think you should swap out this creative because it’s starting to decay.” There isn’t really that intelligence layer around it. It’s going to try its best to be efficient, but it’s always for the benefit of Meta, right?

Tamera Montagna: Sure. That’s their mission, and they’re concerned about driving customers’ results within their platform. This is where Clarvos offers a big advantage for our customers. We’re unbiased. We can look across publishers and find where the audiences are that match your brand. We can develop a strategy across different publishers so you optimize your spend and find your audience wherever they are. I’ll tell you how unbiased it is: we’re not even going to try to increase how much you spend. We’ll tell you what the strategy is, how much it will cost to reach that audience, and which publishers you should use to reach them. But how much you spend doesn’t impact our revenue at all. We don’t take a percentage of your ad spend. You pay a subscription fee. You can put as many advertising dollars as you want through our platform, and there’s no additional cost.

Steve Hutt: That’s interesting, and I appreciate that transparency. You describe these simulated customer profiles in plain language and provide a lot of detail. I’m not trying to rag on Meta, but if you contrast that with its targeting, it’s a black box. Let’s be honest.

Tamera Montagna: Yes. What’s interesting about Meta is that it’s really granular. If you look at all the options across interests, behaviors, and demographics, there are actually 9,000 combinations you could choose. When a marketer is trying to find an audience, imagine having 9,000 possible combinations. Which combination is going to drive the best results? That’s the question.

Steve Hutt: Exactly.

Tamera Montagna: That’s where we have the advantage. We can identify where our consumer segments are and how to reach them on Meta, and on the other platforms.

Steve Hutt: That’s amazing. Walk me through what happens when a brand connects its store and moves toward a live campaign. What’s the onboarding process? If a brand already has a live campaign, I assume you ingest the data and do some analysis before making decisions. At some point, I assume the platform makes recommendations for approval based on the testing and experimentation. Walk us through what happens.

Tamera Montagna: If you were listening to this podcast today and wanted to run an ad this afternoon, you could do that. It takes about 10 minutes for us to set up your account, strategy, creative, and campaigns. What’s great is that, while we’re using artificial intelligence within the platform, the merchant is the human in the loop. You can agree with, approve, or change what’s being recommended. You approve everything. No ad goes out without that approval process, so the human in the loop has total control. Because of AI, we can do it very quickly, in as little as 10 minutes. It probably takes about an hour to work with a customer to get their store set up, and after that, they can run a campaign. It’s that fast.

Steve Hutt: That’s amazing. A lot of founders are using Claude as their source of truth right now, connecting MCPs to different tools and using that agent environment to help run their business. They might be querying Shopify, Google Analytics, and Webmaster Tools. A lot of SaaS companies are creating MCPs now, so someone might say, “This is my source of truth, with the full context of my business, and I’m connecting it to dedicated platforms.” What can they get differently from your platform versus querying ChatGPT or Claude directly? Even if their ad accounts are connected through an MCP server to Claude, what’s different about running through your platform?

Tamera Montagna: I use Claude and ChatGPT myself, but I don’t use them to create personas or strategy. What’s different is that we use billions of documents for marketing purposes. Our activity is specifically focused on marketing, and the social media information we use is fresh and updated in real time. We don’t just ask a tool to create a persona, as you might with Claude or ChatGPT. A lot of the information those tools provide is interesting and useful. What we do differently is use all this information to understand your brand, your competitors, your vertical, what’s happening in society, and what people are talking about today. Then we use that to create everything on our platform. It’s different from creating one persona in ChatGPT. We use those documents to create 50 personas for every consumer segment. It’s much more scientific in that way.

Steve Hutt: As an analogy, suppose you use Claude and create a health project. You connect your Apple Watch and upload your blood work. It can give you some reasonable information, but there’s a reason specialized tools exist for doctors. General tools have a lot of information, but startups are building dedicated applications and skills for specific industries, with the queries, answers, and resources those professionals need. Medical information can be scientific and complex. A doctor wants information that helps them say, “That corroborates what I was thinking,” or, “I hadn’t considered that.” General models are great, but they’re not necessarily going deep into those specialized resources. That’s why these startups are emerging for specific people and use cases, rather than the more general use cases we see with ChatGPT and Claude, right?

Tamera Montagna: Yes. Let me pause for a moment. I don’t think my answer was very good, but yours was excellent. You might be able to tweak my answer a little and then include your explanation, which I thought was beautiful.

Steve Hutt: Absolutely. I’ll tweak it a bit. I’m trying to make the conversation flow smoothly and explain the differences. These chat solutions are incredible generalists. They’ve ingested a lot of information and are great at many things, but they’re not subject matter experts in everything. I use those tools myself with my Apple Watch and blood work to look at things like my LDL, but I wouldn’t want my doctor relying on Claude just to double-check and confirm something. The comparison with your platform is that you have all this data, so you can run simulations against it, make educated judgments with AI about personas and creative, and then make decisions. Knowing all that, my next question is about early adopters. You have brands using the platform, and I assume you have real-world case studies. I’m not sure whether they’re public yet, but can you share anything anecdotally or call out any brands? I’d love to hear what their life looked like before your platform and what it looks like after, or how they’re using it.

Tamera Montagna: Yes. We’re excited because we’ll be able to put case studies from our early-access partners on our website shortly. We have a team of performance media managers who work with our merchants, and in this case, our early-access partners. We work with each early-access partner for three to six months. What we love is hearing their feedback when they go through the platform for the first time and can’t believe how easy it is. For instance, they’ll say, “I love that campaign and the creative you built. I just wish it were on a sunny day.” Then we show them how quickly they can change it. They type, “Change the background to a sunny day,” and it changes. Their reaction is really fun. Another customer thought everything was perfect at each step and loved the creative, but wanted to adjust the content and call to action. They said, “I want it to sound like I’m speaking to my sisters.” They typed that in, and it changed. They loved it. So much of what we hear is about how easy it is to advertise on social media, either paid or organic. We love hearing that affirmation. Some customers don’t want to use generative AI creative, or they want to use less of it. They can upload their existing creative and use our tools to adjust it for the consumer segment. They can change it however they want, test it against the consumer segments, add it to the strategy and campaign they’ve chosen, and launch it from our platform.

Steve Hutt: Amazing. For a lot of brands and marketers listening, I think the word I want to use is “taste.” Brands have a specific taste and know what they want. That’s why the human-in-the-loop concept is a good analogy. Generative AI can do a lot, and we can target things toward the segment we want to reach, but sometimes it’s just not quite right. That’s part of the context loop: “I like this,” or, “I want to change that.” You make those changes, and then you’re ready. The nice thing is that this also becomes a context loop within the platform for that brand. It learns that they like sunshine or prefer calls to action written a certain way. Somewhere in the background, there’s a record saying, “When we build creative for these customer segments, we need to avoid this because the brand says it’s not for them. They prefer this.” It’s interesting that there’s an ongoing learning process.

Tamera Montagna: Yes. It’s an intelligent platform, so whatever information you give it, it gets smarter and smarter. The same is true of campaigns. When you’re running them, the guidance and suggestions you receive continue to get smarter. It’s exciting to see customers astonished by how easy it is and how they remain in full control. Some of our early-access partners are busy business owners. They’re running and growing a business with a fractured marketing solution. They have a lean team, if they have a team at all, and it might consist of a small tech stack with some freelancers. They’re trying to stay on top of marketing while driving the business, managing supply chains, and building products. We want to provide one platform a Shopify merchant can use for social media advertising. Shopify is their store, so that’s one platform. They might have a CRM to contact existing customers, which is another. Then everything they need to execute a social media campaign happens on Clarvos. Currently, they might use five tech solutions and work with freelancers to navigate Meta and other publishers. With Clarvos, all of that is on one platform. You log in to one place, manage campaigns, make changes while they’re running, and see campaigns across publishers. If you’re a Shopify merchant, you have your store, you might have a CRM, and then you have Clarvos. It makes your life much easier, and that was our goal.

Steve Hutt: You’re connected to the ad platforms, so campaigns would be pushed from Clarvos to Meta, for example. There would be analytics within Clarvos, but a brand might also use other tools to visualize data directly from those ad platforms. It’s not necessarily two sources of truth. Where the money is being spent is the source of truth, and you’re the underlying layer making intelligent decisions about creative and pushing it to the ad platforms. I want to make sure I understand that people can choose where to push their campaigns and where to see the results.

Tamera Montagna: That’s true. They can go directly to the publishers or see the results through Clarvos. They can view them wherever they want. What’s interesting is that we’re looking to help customers grow, and to do that, they need to acquire new customers. We talk about customer acquisition costs, but we also want to help merchants understand the lifetime value of a new customer so they can determine whether the acquisition cost makes sense. For example, maybe it costs $20 to acquire a customer, but that customer stays with you for an average of three years and spends $1,000 over that period. I’m making those numbers up as an example, but that’s what we’re helping brands understand and see on our platform. It’s more than ROAS. It’s about the cost to bring a new customer to your store, how long you’ll keep them, and their lifetime value. Then you know how much you can spend to acquire them. We want to guide merchants in making those decisions and understanding whether it makes sense. It doesn’t make sense for every business, and Clarvos isn’t for every business. We help Shopify merchants understand whether it’s right for them.

Steve Hutt: I wonder about testing beyond creative. I’ve had people on recently talking about bundles, different offers, and pricing, which are interesting things to test. Could a brand test 20% off against a free gift with purchase? Could the simulation help figure that out? Margin is another interesting one. Could you tell a brand whether a discount will cost them margin? There are a lot of potential what-ifs. Can you share any insights?

Tamera Montagna: You’re going to see a lot of that in the future. Right now, we need more data to make those predictions, and our roadmap is filled with those kinds of outcomes. We want to get to an outcome-based solution, where a business can tell us, “I want to sell 100 of these items next week,” and we can say, “That will cost X dollars. Here’s your whole plan.” The idea is to help them understand what it would cost today, given everything else happening and everything else they’re doing, to sell that amount of merchandise. To get there, we have to answer the questions you’re talking about. Right now, we’re not using CRM or customer data to inform our platform. At some point, we may go in that direction, but today we’re focused on helping our customers find new customers: people they don’t know who would be interested in their brand if they knew about it.

Steve Hutt: That’s amazing. We’ve talked about different ad platforms, and this might be something for your future roadmap, but a lot of listeners take a multichannel approach to driving net-new revenue. Amazon comes up a lot because of its organic reach and AI visibility. It tends to appear in organic search, so it can make sense to have at least some of your product mix on Amazon, even if you’re trying to bring people back to your home base. You want to own the audience so you can retarget and sell to them again, whereas Amazon owns that customer relationship. Where do you think Clarvos is headed with a multichannel approach? Right now, we’re talking about Facebook and Instagram, driving creative toward conversions, and considering lifetime value. What about other platforms where people advertise? Could that creative be used elsewhere, and how does Amazon fit into the mix?

Tamera Montagna: Absolutely. Amazon is on our product roadmap. We’ll be expanding to Amazon and many other channels. We’re excited, but we’re taking it one step at a time. We’re going to grow alongside our customers.

Steve Hutt: That’s amazing. I’ll make sure I follow up on this. One thing I love about this podcast is having conversations annually. Some people have been on five or six times because there are always product updates and new things happening. It’s nice to ask, “What have you been working on for the last six months?” and hear what’s changed. As you said, there’s a customer feedback loop. You can share what you’re working on, what success looks like today, and eventually how you’ve expanded into Amazon, what that creative process looks like, and how you track it. It’s definitely a crawl, walk, run approach with AI right now. I love it.

Tamera Montagna: What we really want is to compare contribution margin by channel. When we can do that, we can optimize further. Companies can already do that, of course, but the businesses we’re working with right now might have $1 million to $10 million in revenue, and we’re focused on their immediate needs. We’re choosing platforms and channels based on what will drive their growth right now. As we grow, expand our offering, and serve larger SMBs, we know contribution margin will be key to the strategies we create for customers.

Steve Hutt: I mentioned at the top of the show that there’s a code freeze coming up with Shopify, and a lot of app companies are doing the same thing. They want to let the dust settle and get through BFCM without introducing too many variables through code changes. At the time of this recording, we’re four weeks away from Black Friday, so there should still be a solid three weeks or more before a code freeze and the Super Bowl of commerce. The rest of Q4 is still great, too. It doesn’t end with Black Friday and Cyber Monday. People buy things they missed, restocks come in, and the season continues. Do you believe there’s time for a brand listening right now to get involved with Clarvos and launch a campaign? These small and mid-market businesses are trying to figure out their edge. Could they at least run a simulation now, get ready, and launch a week before or during the event? I want to make sure there’s enough time and that we’re not rushing anything. You’ll have better insight into the next steps based on where we are right now.

Tamera Montagna: Absolutely. Our expectation is that we’ll bring on a few hundred customers in those four weeks. Because we can do it so quickly, we can meet and beat the timeline you’re describing. We also have performance media managers who stay with a brand for three months and answer questions about social media, not just our platform. They help optimize the brand through our platform, but also help with organic and paid social media overall. That’s at no cost, and it helps a brand get up and running quickly. We stay with them to help make sure they’re successful.

Steve Hutt: That’s amazing. I love that hands-on support from performance managers because they focus on this every day and know what great looks like. In my old role at Shopify, I understood the platform intimately. I knew which features were toggled on or off and which beta flags were enabled because I had access to the back office. I could be intentional and say, “You should try this beta,” or, “There’s a partner I believe can help you, and here’s why. Can I make a warm introduction?” I’m proactive that way. That one-on-one relationship is valuable, especially in the early days as you get used to a new platform. It’s nice to have guidance from someone who understands it well and knows what great looks like.

Tamera Montagna: Yes, and that’s why any brand wanting to join Clarvos in time for this season would be able to do it. We can move as fast as our customers want to move.

Steve Hutt: It’s not all or nothing, either. There’s no reason a brand couldn’t pick a collection of products and run campaigns based on Clarvos recommendations for creative and budgets. Then, when the dust settles, they can say, “Interesting. Here’s what I picked, and here’s what Clarvos picked.”

Tamera Montagna: Sure, absolutely.

Steve Hutt: Suddenly, you have an interesting dialogue around the results, and you can continue from there. It’s not all or nothing, and that’s an important caveat.

Tamera Montagna: That’s a great point, Steve. Clarvos is part of a solution. It’s not just a tool; it’s a platform that makes it easier for our customers to advertise and grow through social media. It’s not everything they’ll do, but it does a lot.

Steve Hutt: That’s amazing. For those listening, what are the next steps? I believe people can book a demo directly on the site to see what it can do and kick the tires a bit. Are there any other next steps, or somewhere else you want people to go?

Tamera Montagna: Going to our website and signing up for a demo would be a great start. We’re also going to put a tool on the site shortly so people can try it without doing a demo. They’ll be able to try the simulator, take their current creative, and test it against different consumer segments. They’ll start seeing tools they can try before signing up for the platform.

Steve Hutt: Amazing. I’ll have everything in the show notes. I’ve followed both the company and you personally on LinkedIn, so that’s another way to get in touch. As I said, I didn’t even realize a platform like this existed. You hear about platforms for A/B testing creative, pricing, and other things, but not with the billions of data points you have because of your experience in this industry. You’ve developed all these customer segments and can run creative through a simulation against those simulated people. It almost seems like science fiction to me. Then it comes up with winners, and you’ve proven with your beta testers that those winners are working because of the data you have and how you’ve segmented it. It’s an interesting concept. Kudos to you and the team for coming up with the idea. I wish you continued success and a wonderful holiday season. Hopefully, people will at least book a demo and try it. You don’t know what you don’t know, right? Thank you for coming on the show today.

Tamera Montagna: Steve, thank you so much for inviting me. I’d love to give you updates as we progress and expand our offerings. I’d be happy to join your show again in the future. Thanks again for this opportunity today.

Steve Hutt: My pleasure. Have a great afternoon.

Tamera Montagna: Thank you.