Podcast Episode 2: Always. Be. Testing.

Podcast Episode 2: Always. Be. Testing. Featured Image

Tap in on Verse CEO David Tal and Intellimize CEO Guy Yalfi’s conversation about the importance of “Always Be Testing” and how you are missing out if not doing so!





David Tal: Welcome everybody. I am so excited to have here with us today, the CEO of Intellimize, Guy Yalif. Guy, welcome to the show. 

Guy Yalfi: Thanks for having me excited to join you. 

David Tal: Thank you. Guy, I’m really intrigued with what you’re doing at Intellimize, but before we dive in too deep, why don’t you tell our viewers a little bit more about what Intellimize does?

Guy Yalfi: Sounds great. Intellimize intelligently optimizes websites. We personalize them for each unique visitor in the moment to meet them where they are in your buyer journey to drive more revenue, more sales and more customers, depending on the kind of business you are. 

David Tal: Awesome. One of the ways that I’ve heard Intellimize being described is “AB testing on steroids.” Tell me about that. 

Guy Yalfi: It resonates a lot because you can break a lot of the constraints you have with AB testing. But at that high level, it’s AB testing and personalization together. Both of those are often separate budgets, separate strategies, separate teams. In our humble opinion, they’re all in service of driving more revenue. And so we use both together at scale. 

David Tal: Awesome. And personalization is something that people are paying more and more attention to. This concept of, make a million people feel like one in a million. How do you put that into practice? Can you give us a couple examples? 

Guy Yalfi: Happy to. It’s similar to what we’ve all been doing in ads for the last 10, 15 years. It’s not an accident that I spent the last 20 years in ad tech, where you would give Facebook, Google, or others, five ad creative, not one. They’ll figure out the right one to show each unique visitor in the moment. They’ll show the good ones more, amplifying your good ideas. They’ll starve, the bad ones of traffic, protecting you from them and optimized for some goal you care about. That didn’t exist for websites, where the money’s made, where we work. All this spend, all this energy to get people to convert. That’s what we’ve done. And so practical examples, meet your prospects where they are. Treat each individual visitor differently, whether they’re high value, low value, whether you want to mirror the message they saw on the ad that got them there or the email that got them there, continue that on the website, or simply have two people who have a difference of opinion inside the company about how to message resolve using data. 

David Tal: I love that because so many people do disagree and they have their own opinions and their instincts. And and we just operate on the data. We say, eh, we can have opinions, let’s look at the data. And so AB testing is so core to what we do, and obviously all marketers do. That what you do is so important. People go on a site and they have Intellemize they’re leveraging Intellimize, is the copy changing? Is the imagery changing? What else changes? Do CTAs change?

Guy Yalfi: All the above. So to your point, you and I might go visit dermatologica.com today, and we might see two different things. If you showed up a week later, you might again see something different, either because you did something or because the machine learning, which updates itself every few minutes on its own with zero human intervention, got better at predicting human behavior. What might those different things be? To your point, could be headline, call to action, imagery, layout on the page, interactivity, right? Maybe you add an exit modal to capture someone’s attention. It hopefully is continuing what we know about you from before. So if you saw an ad about this product, we’re showing you that product. If previously you browse this category, hopefully we’re bringing that front and center so that even someone’s homepage feels as personalized as a dedicated landing page. 

David Tal: That’s amazing. And I think that’s what every consumer wants, is to feel that they’re being spoken to not at, right? 

Guy Yalfi: Very much with you. And the more you can be helpful to them, remember things about them. It boggled my mind when we started that we would spend all this time coming up with the right context, the right product, the right imagery, the right everything in our ads, spend so much time doing that, tailor the journey all the way, get to the website and forget all of it.

David Tal: Yeah. 

Guy Yalfi: And so the more you can remember that, the more useful you can be, and probably then be rewarded with revenue.

David Tal: And as a consumer, you’re absolutely right that you see certain things that grab your attention here at the ad space. Then you end up on the website, it’s completely disconnected, a different message, sometimes completely different branding and colors, because you have fragmented teams working on these different parts of the funnel.

Guy Yalfi: Very much with you. And you talked about how it’s different than AB testing. That can happen when you’re AB testing. AB testing’s wonderful because you’re doing data driven marketing, to your point about resolving disagreements and improving the experience, it’s less wonderful cuz you treat everybody the same. It’s AB testing is like saying, you know what? Every woman in the us has to wear size nine shoes, because that’s the average shoe size for women in America. Sounds ridiculous when you talk about shoes, but people do it on websites all the time. And it’s worse than that because you update your site. You baked it in. 

David Tal: Yeah. 

Guy Yalfi: Do you rerun all your AB tests every time you change your targeting, you’re messaging, your competitor to something? You don’t. But the world changes, so now it’s like telling every woman in the US, you have to wear size nine shoes from the day you’re born until the day you die, because it’s the one answer for everyone forever. Sounds totally ridiculous. You can use machine learning to give everybody the right shoe for them. That grows over time changes as their style changes. 

David Tal: That’s a super interesting analogy. And I’m sure all the women are resonating with it. So you’ve raised quite a lot of money from venture capital, from some of the most respected firms there are. What do you, as a CEO, really try to focus on and where are you placing those bets? 

Guy Yalfi: We feel very fortunate to have some amazing partners on our board and among our investors. I feel humbled to be able to work with them. Candidly, the place we focus is our customers and our culture. For us, we don’t get up in the morning thinking, “how do we do this financial engineering? How do we kill this competitor?” We’re aware of what our competitors do, but for us, we get up in the morning saying, “how do we make that marketer, that merchant, that PM a hero internally? How do we get them promoted? How do we get them up on stage saying nice things about us?” That is our true north of success. That then guides a whole bunch of decision making and our culture that is literally our top value. We also then say, we want this to be the best professional experience of our lives working here that it’s an inflection points in people’s careers. We’re also humble, empathetic, logical, and data driven, and proactively transparent. Those set of values, focus on them a bunch, cuz it makes it fun to go to work every day. Because all the mumbo jumbo, not all, but most of the mumbo jumbo we’ve all experienced in so many different places, we don’t have that. And I genuinely view it as a scaling mechanism for the company. People make better decisions at all levels of the org because they live and breathe these things every day.

David Tal: Absolutely. That’s so important. People spend more time at work than they do at home. You’re at work eight hours, you come home, there’s three, four hours before you’re going to bed or five hours. Maybe six, if you’re a late night person. But it’s interesting people don’t really think about that as much. But when someone told me that once, your employees are like your family, because they’re spending more time with you than their own family. And it brought me this sense of responsibility and care, and maybe my employees will disagree, but I really care about my employees because I look at it through that lens of, I want this to be a special place that they spend 40 hours plus a week. And when they do, and when they feel that they’re taking care of, they’re family, the productivity, the employee satisfaction, and, the, having the owners cap on is makes such a big difference. 

Guy Yalfi: We are reading from the same book, very strongly. We we say often internally, like we’re not family, but it’s not an accident that we feel like family. We want that to be the case. And we also will sometimes talk about how, like life short, you gotta do stuff you enjoy doing with people you enjoy doing it with. 

David Tal: Yeah. And you’re a team, more than family. Reed Hastings famously says, “we’re not a family at Netflix, we, but we’re like a, we’re like a sports team, where we have to be able to rely on one another and count on one another and care one another like a family, but we have to be dependent cause you can’t fire family.” And if you don’t perform, you’re not sticking around forever. We’re AB testing employees too, right? 

Guy Yalfi: Every one of us. And it’s infectious, right? You get a bunch of A players that operate at a high level. They wanna attract more A players and it’s this self-reinforcing thing where you then start attracting people. Maybe you wouldn’t otherwise be able to get, because they’re like, I want to be happy doing what I’m doing with great other people. 

David Tal: Yeah. Reed talks about that and calls it “building talent density.” Because the high performers will attract more high performers and then compete with themselves to be even higher, more amazing performers. But if you bring on a few kind of duds on the team that are a little lazier, don’t care as much, they have a bad attitude, they don’t elevate to the high performer level. What happens is the high performers average out with the low ones and it ruins the moral and their productivity and you stop exceeding expectations. And so it’s so important to keep that talent density as strong as possible internally, and that’s so important. So when people talk about AB testing, what’s the average amount of tests someone may do on an average page at any given time? So should it be called AZ testing? 

Guy Yalfi: People do sometimes call it ABN, and it really depends on how much traffic they have. The more you got the faster you can go. We do see our average customer over the last couple of years, tried 177 different ideas that would’ve taken him a quarter century to do. Then you can use machine learning to, which they did, to go optimize across the 78 million possible combinations of those, which no me mortal marketer can ever go do. And so we think the right way to go accelerate AB testing, be multivariate, and simultaneously go personalize, is to go free up time to go crawl inside the head of your prospect. Go walk a day in the life of their shoes. Go understand what gets them promoted. What makes them a hero internally? And use that as fodder and then have machine learning and your own rules, should you have some, drive what gets shown to who because your prospects will tell you your website, visitors will tell you and a machine can listen and act on that better. 

David Tal: What I love about what you do is it’s measurable. You can actually measure the on page, the time on page people spend, the conversion rate on that page. And so it’s not just this should work, but it’s did it work or did it not? And your whole business is just around data. So there’s little argument, it’s just trying different things out, testing, and then using intuition and experience and, real skill to know where to focus and what to optimize. Because if you look at one page, take a homepage or take a pricing page or a demo page, a landing page, there’s I can think of so many elements there that you can try to optimize and probably all at the same time, but at the same time, if you try to optimize too much at once, is it tricky to understand the data? Is it better to piece it, start with the header, change the image, change the layout. But if you do too much all at once, can the results come back in a noisy way? That’s hard to decipher? 

Guy Yalfi: It’s a great question. And it’s a great use of ML. I’ll take a half step back and say amen to data driven marketing. And it is remarkable when you have truly data driven marketers who are like, I am certain, this is gonna work. That it doesn’t. They’re like, okay, that’s what the data said, as opposed to no, they just gotta be wrong. No. This is some something’s off here to directly answer your question. You can use ML to tease that out. You can use machine learning to understand. Hey, I ran this highly multivariate thing. Using ours, you can also start and stop stuff midway, which you cannot do in an AB test while you’re waiting for weeks and months, use ML to understand what was the impact of this individual thing? What was the impact of these things together? And tease that out in a way that is statistically valid. You can not muddy the results and still move more quickly. 

David Tal: That’s super interesting. I wanna talk to you after this, about, about Verse and how we can leverage you. We’re always AB testing. Sure. All the time organization, it’s a constant. And but you bring such a level of experience there and a, just a different level to the game from everything else I’ve seen. That’s it’s impressive. So I’m excited to talk to you after this separately. Not trying to over promote you. I’m genuinely interested as a marketer myself, marketing our product all the time, so 

Guy Yalfi: I’m grateful. It fires you up. I must admit we are. Totally conversion obsessed as a group. And it is nice to talk to other conversion obsessed markers. 

David Tal: Yeah, absolutely. Our name verse comes from con converse and conversion together conversations drive conversion. And that’s what we’re all about. And I. 

Guy Yalfi: That’s a brilliant name.

David Tal: And I’m curious. Intellemize was taken, so we had to we had to keep looking. When you are analyzing someone’s page, are there key elements like to our viewers as like advice to give them, where should they start on optimizing? Is it header? Copy? That’s the most important? Is it imagery? Is it the CTA? Is there an answer? 

Guy Yalfi: In my humble opinion, if you don’t know anything about your users yes. Start with above the fold prominent stuff. If you know a bunch about your users, your prospects. What’s getting in the way of them moving forward on that page? That’s the place to start that concept of, oh wait, we have a low trust moment here. Oh wait. This is complicated. That will fuel stuff that’s meta above. Is it a headline or is it right? That concept of, oh, I wanna simplify this. Maybe reflected in layout, imagery a copy all at the same time, because you’re bringing that concept to life. If you don’t have those insights, cuz not everybody does all the time, then totally above the fold prominent stuff tends to yield more fruit because more people see it.

David Tal: All that’s good to know. What do you see as the future of optimization? 

Guy Yalfi: I’m almost certainly biased in the answer because having spent 15- 20 years in ad tech, I watched everything go from paper iOS, which is just, I want to buy this static thing. To me that feels like a, be testing to rules, which was the early days of search where you’re like, if they search for this term, I wanna show them that thing to programmatic advertising, which is, gimme a bunch of creative and I’ll figure out what to show whom and show the good ones more and the bad ones less. To me, it’s natural for the next part of the funnel on the website to go the same way. And in my humble opinion it will, it’s it’s a better answer to meet every prospect where they are in the funnel with you. And we’ve all talked about one-to-one marketing forever to be able to practically do it. I, I think we can generally do it in ads to be able to do it on our websites, I think is the next great thing. And so then you great, you get that, then you need to marry that with great ingredients. You need to marry that with the prospect in sight that’s both qualitative and quantitative. I think that’s chocolate and peanut butter.

David Tal: Yeah, I like that. so let me ask you this. You guys are working really hard on page, right? Com optimizing what happens after, so you improve conversion. Okay. More people now complete this form on this page. Do you track what happens afterwards? And what are some of the things that you see, or maybe challenges that you see companies face after you do all the amazing things that you’re doing and flush out a better way for them to generate prospects, which is what they want.

Guy Yalfi: Thanks for asking. And the amazingness comes from their ideas, right? It’s their ideas and their customer insight activated with this machine learning. That’s the great combination that tends to yield better results. We will then help them optimize the entire journey on the. Shape of the journey, the length of the journey, where they want people to go. That’s one set of things. And sometimes the entire journey is on the site right on. E-commerce sure they, they bought it right then and there. The meta challenge there, I think is the same as the meta challenge you were talking about of, Hey, conversion happens offline in B2B sales, enabled B2B. It’s almost always like that, right? You generated a lead and now a sales team needs to engage. We’ll listen to those conversions, we’ll use them as well. Then you’ve got a trade off of where in the funnel do I want to optimize for? Do you wanna optimize for the near end thing? Oh, I got them from the landing page to the product detail page. Do you wanna optimize for the ultimate goal? They purchase something. We started the company optimizing for this. We’re like did you want revenue?

David Tal: That seems like the natural thing. 

Guy Yalfi: Totally natural. The more we dug in, the more we’re like, hold on. These results don’t make sense. Data driven marketer is fine. The data said they don’t make sense. This is what we saw so much of it that we came to appreciate that sometimes. What you put on the homepage actually has no correlation with whether or not they purchased. Blasted me, one would say it did have correlation getting you to the next step, but when you gotta enter your credit card here, or your social security number, when you’re doing a financial services application, that has a much bigger impact on whether or not they make it all the way to the end. And so there’s a set of things you gotta do there. And so we started shifting people towards optimize each step to get to the next. Okay, that did help. As long as you do every step, otherwise you might have pushed a bubble of unqualified folks down and they’re gonna drop later on. Okay. So you optimize every step for the next version which we have quietly running in the background is optimize for all of it at the same time, have the ML figure out, Hey, what’s the relative waiting between these? How can I optimize both to get somebody. So the next step and all the way at the end and take into account everything that’s happening in the journey. Whether that final step is online in e-commerce or offline in sales. In particular, if it’s offline in sales, you want a near end signal on the website. Cuz if that sales cycle is a week, a month, a year, you need a signal to optimize on the phone. 

David Tal: You just load that in later and start to understand it. You need signals along the way.

Guy Yalfi: So that you can iterate more quickly on your ideas. Totally with you, David. Yeah. 

David Tal: Awesome. Let me ask you this. When you’re trying to help a customer and you’re telling ’em all of these things, what is the biggest hurdle or challenge that you deal with, with those customers to get them, your clients, to really get them to adopt and understand the power of what you do? 

Guy Yalfi: Interestingly the- 

David Tal: are they afraid of changes cuz they’re stuck or they just love their slogan or they just, they’re just used to things or it’s the thing, it’s a way it’s always been done, or… 

Guy Yalfi: Fortunately most of them are very data driven. And so there’s part of that comes more easily, depending on how you apply the machine learning. If you recognize that the world keeps changing, there’s this really difficult if you’re really set on AB testing mindset challenge of, there isn’t a winner. There is a winner. You can reach statistical significance and use that in your other forms of marketing. Use it in your sales materials. But that answer might change, you run a promo, you change your messaging, your competitor does something. In reality, the answer might change. We never acknowledge that when we’re doing AB testing. Cause the test is done. That is a big mental shift. To say, hold on a minute, I built my reputation in this company by finding winners, my career advances, because I found winners. Now you’re telling me my winners might go away. That’s super hard to reconcile. And just like people took their own organizations from intuition driven marketing to data driven marketing to appreciate how important winners are, it’s then another level of organizational understanding to say, “yes, I found winners and they might change.” 

David Tal: Yeah. 

Guy Yalfi: And that’s not a bad thing. 

David Tal: You gotta keep evolving and adapting. Yeah. Awesome. What advice do you have for the average business owner or marketing team for this part of the process? 

Guy Yalfi: It is probably similar to advice a lot of conversion rate optimizers give, which is go test. Go run. Go understand your prospects and- 

David Tal: always be testing.

Guy Yalfi: Always be testing. Yes. A B T. Always. Be. Testing. 

David Tal: I couldn’t agree more. Yeah. Do you have a quote or a saying that really resonates with you or that you repeat a lot with your team? 

Guy Yalfi: Customers are true north. We live and breathe. It’s very trite, but to actually live it as an organization is a thing. And with our customers, we’re often talking about “meet every prospect where they are in their journey with you.” It’ll produce good results over time. 

David Tal: I love it. Guy, thank you so much for spending time with us today. And please check out Intellimize if you’re interested in optimization that’s on steroids. We’re really talking to one of the leaders in this entire industry. So thank you Guy so much for chatting with us today. 

Guy Yalfi: David pleasure to spend time with you. 

David Tal: Cheers. Thanks for having me. 

Guy Yalfi: Cheers. L’chaim! 

David Tal: L’chaim!