How To Personalize Your Ecommerce Site For More Sales — Sergiu Cazac | Why Experimentation Fuels Growth, Why Understanding Customer Journeys Improves Sales, Why Hypothesis-driven Testing Beats Guessing, What Makes A/B Testing Effective (#425)

In this episode, we dive into why ecommerce brands should focus on experimentation over chasing traffic. 

Sergiu Cazac, founder of personizely.net, shares how mapping the customer journey, testing smarter, and learning from failed experiments can unlock higher conversions. 

He explains what makes A/B testing tools easy to use, why timing matters, and how personalization connects with testing to drive real growth.

Topics discussed in this episode:  

• Why chasing more traffic can hurt conversions. 
• How understanding customer journeys improves sales. 
• What funnel mapping reveals about hidden leaks. 
• Why hypothesis-driven testing beats guessing. 
• How failed experiments create valuable learnings. 
• What makes A/B testing tools simple and effective. 
• Why timing tests year-round uncovers better insights. 
• How personalization and experimentation connect. 
• What successful brands learn from losing tests. 
• Why smooth onboarding helps brands scale faster. 

Links & Resources 

Website: https://www.personizely.net/
Shopify App Store: https://apps.shopify.com/personizely
LinkedIn: https://www.linkedin.com/in/sergiucazac/
X/Twitter: https://x.com/sergiukazak

Get access to more free resources by visiting the show notes at
https://tinyurl.com/n44teswa

Transcript

00:00:00:04 - 00:00:22:04
Unknown
The most important thing in the testing is obviously to have traffic and to have sales. I see a lot of customers starting out and they don't really have enough traffic to have statistically significant tests. Why do so many store owners chase more traffic instead of fixing their conversions? Yeah, I think that's a very good question and it's been asked a lot and the answer is.

00:00:22:06 - 00:00:44:16
Unknown
Hello and welcome to another episode of the eCommerce Coffee Break podcast. In eCommerce, nothing standstill. Customer behavior changes overnight competition that's tougher than ever. And what worked yesterday might not work today. That's why experimentation AB testing is so important. Testing your site, your offers and your customer experiences is no longer a nice to have. It's the fastest way to get more sales from the traffic you already have.

00:00:44:18 - 00:01:05:01
Unknown
In today's episode, we're talking about why experimentation is so important. My guest today on the show, Sergiu Cazac. He's the founder of Personizely, actually a platform that helps e-commerce brands boost sales with simple, powerful marketing tools. He's a life long developer who launched personalized V in 2018 to give marketers smarter, faster ways to get results. We have a lot to cover, so let's dive right into.

00:01:05:02 - 00:01:28:22
Unknown
Hi. Such a value today. Hi class. I'm great. Thanks for having me. Excited to be here. First question Why do so many store owners chase more traffic instead of fixing their conversions? Yeah, I think that's a very good question and it's been asked a lot and the answer is not always the same. So I think people are always leaning towards the path of least resistance.

00:01:29:00 - 00:01:58:14
Unknown
So they learn how to bring traffic to the website. They know how to do that, so they learn to do the stuff that they already learn how to do and they often forget, obviously, that there are leaky funnels, the points inside where they're losing traffic when they were people, that they bring them more and more people that they bring can actually decrease conversions because in time it can be can become a race to the bottom, because you can't have all the builders in the world.

00:01:58:19 - 00:02:25:00
Unknown
And at some point the traffic became worse and worse and the conversions lower. And it's counterintuitive, but people still keep on doing this because that's what they learned to do at the beginning. That's what the marketers, the gurus are telling them. That's my opinion. Mm hmm. Okay. It makes perfect sense. Now, a lot of stores out there, or most of the stores that only convert like 2 to 3% of their visitors into customers.

00:02:25:02 - 00:02:51:05
Unknown
Why does it happen? Where does that happen? First thing is on speaking biases. Baselines is always funny because it can be sometimes. I just had a customer today that said they have a four or 5.5 conversion rate and they say they they think it is good for them. But I think this can still optimize. Another store can have 225 and it can be very good for them and it's like a top, top performing website.

00:02:51:07 - 00:03:13:23
Unknown
So first of all, his is baselines, a very customer or niche dependent. Speaking about why this stuff happens is because obviously I think the first thing that you have to think about when you think about conversion rate is you need to understand who are the visitors visiting your website for the top of before the conversion on the website.

00:03:13:23 - 00:03:44:12
Unknown
So you need to understand, are you attracting the right audience? So first of all, so even does it have anything to do with my facilitation? But if you are attracting the wrong audience, no matter how many experimentation, no matter how much fun optimization you're going to do, you're not going to get good conversion rate from that audience. So first thing is, you need to understand why targeting quite perfect customers, how you bring it on to the website can be ads, partnerships, whatever, depending on the niche, obviously.

00:03:44:12 - 00:04:06:20
Unknown
And once you already understand that you learned to bring good traffic to the website at targeted traffic traffic that's actually interested in your offer, maybe they don't want don't buy right away, but they might buy in the future. But still they have to be relevant. And that point is where we are starting to talk about the conversion rate and what the stars are telling us on the website.

00:04:06:23 - 00:04:29:19
Unknown
And that is where we start to optimize that 2.534% conversion rate. But before that, I think that conversation doesn't even make sense. I think people have to start there. Mm hmm. Now, you mentioned before that the customer journey starts much earlier in the funnel. So first touch points, then you go, they go through the different steps until they reach to check out.

00:04:29:19 - 00:04:53:16
Unknown
And we want to talk about experimentation. Do want to talk about a, B, testing. What's your approach to when you come to a new client, to a new store website? What's your approach when it comes to experimentation? It's a very good question and it's very interesting because we are getting customers who are sometimes just trying experimentation, and then we have agencies as they always approach this thing very differently.

00:04:53:16 - 00:05:23:21
Unknown
And what we try to do, we want to show our customers to also approach the things as a zero agency would do. So first thing is to actually map the funnel. You probably won't believe this, but there are a lot there are a lot of actually successful brands that don't really have that funnel map, so they don't really understand how people are going from an ad or from a, let's say, Instagram post or a Google ad to actually to actually making the purchase or at least initiating the check out.

00:05:24:02 - 00:05:51:18
Unknown
So first thing is to actually map the final and understand what is the point that we can intervene in and and actually improve and optimize once you have the final if you want to have the conversion rates between the steps of the funnel. Another thing is definitely the customer analysis, the traffic analysis, because for example, if we're speaking about need to ask about the hypothesis.

00:05:51:18 - 00:06:15:00
Unknown
So you probably know the one of the most popular things is distrust business or social proof. This is like a go to thing, but it doesn't work for all audiences because if your audience is, let's say, 50 plus year old customers, for them, some of those trust signals might not be worth the same values and say, I work for someone younger.

00:06:15:02 - 00:06:35:18
Unknown
So you need to really understand like who is your audience and how are they actually using your website. And I think what we are trying to do with our customers when they just come must come to us and say, okay, I want to just appetizers of the cart. Copy this. Okay, let's just take a step back and see why you want to do this.

00:06:35:18 - 00:07:02:05
Unknown
Let's see how a customer driving to the website. Is this the right stage to be optimizing and whether the change itself is relevant to the audience that you want to implement it with. And for this thing actually like a lot levers framework from conversion agency. So there's a really great conversion optimization agency. They renamed the game now and they have a Levers framework.

00:07:02:05 - 00:07:34:09
Unknown
I recommend everyone to go and read it as it's an amazing piece of content where they describe how you can actually map the changes that you want to experiment on your website and how when you have a good change or a good experiment, you can maximize up on that. And even if you have a bad one, you can use it to actually understand what are the points, what the the learnings from that experiment to actually apply and maximize the revenue in the future Experiments.

00:07:34:11 - 00:07:52:01
Unknown
Mm hmm. I think that's a big difference, what you just mentioned between just guessing and testing. I think a lot of brands out there, they're just guessing what they need to change and they're not really tracking it. And I find it very interesting that you mentioned that the brands out there that could sort of understand that have no idea how their funnel looks like and where the traffic is really coming from.

00:07:52:01 - 00:08:10:20
Unknown
And so there's a lot of data points. Obviously now when we're talking about experimentation, I think a lot of brands out there, they're sort of scared or they think it's too complex to do a B testing. I want to dive a little bit more into this. How would you how do you get started with a B testing on the new customer side?

00:08:10:22 - 00:08:35:02
Unknown
Yeah. So I think like the best the most important thing in the testing is obviously to have traffic and to have sales. So I see a lot of customers starting out and they don't really have enough traffic to have statistical description of tests and I just don't and you probably have to go back a little bit and maximize the traffic that you're getting because otherwise you're not get any any traction.

00:08:35:04 - 00:09:01:20
Unknown
So this is the first point where customers are just starting too early. And if they're coming to the right time and sometimes have said they are afraid, they're afraid that maybe the stats are going to lie or their phrase is going to be hard to implement, does so with us and our tool. And actually, I think there are a lot of tools on the market where you actually just have to have the right hypotheses and you choose the tool for that for the purpose.

00:09:01:22 - 00:09:26:09
Unknown
And that tool usually allows you to do the changes without any technical involvement. That's what we pride ourselves in. So precisely, we have very easy to use visual editors that integrate very well with with websites, especially with Shopify website, because we have a deep integration with them and we optimized our product since the beginning and we pride ourselves in that.

00:09:26:11 - 00:09:49:20
Unknown
And just as a customer has to be able to basically sign up and set up a desk and let's say 5 minutes, but still a lot of times in advertising we see that we still need to be near that hold their hand because some websites are very oddly built sometimes the changes that they want to implement are not actually worth implementing.

00:09:50:02 - 00:10:25:17
Unknown
So I think if you already have the experience, you can just log into the tool list of hypothesis creative campaigns and start testing. Just not too many, just as a time for sure. And if it's something that you're just learning, I recommend to first as a side, take a step back, analyze what you know about your audience, Learn something about David Dawson from frameworks like the Leverage Framework, and then you can then when you log into an application like ours, I would think it is like eye opening because it's so easy to implement.

00:10:25:17 - 00:10:47:13
Unknown
Our fingers actually make sense when you understand what you want to experiment because sometimes people go the wrong way is they get the tool to experiment because they know they need to experiment, but they don't understand what they can do and they just take it from that what the products can actually do. So they say, okay, product offers, but experimentation.

00:10:47:17 - 00:11:13:23
Unknown
Let me just ask that. But that's not the right approach. You go back from the hypothesis and then you look for the tool, for the purpose and the tool that makes it as easy as possible to implement those changes. And even though that might sound counterintuitive, coming from the founder of a tool like this, I really love and customers are coming to us with actual ideas and hypotheses and we happily brainstorm with them unless against what or when they come.

00:11:14:01 - 00:11:31:01
Unknown
And they just know they have to have the tools because they heard somewhere about it. Mm hmm. Now, I think a lot of brands out there, they have heard about the concept of AB testing experimentation. But as you said, they don't have the experience, they don't have the background. Obviously, you help with that. Talk me through your tool.

00:11:31:01 - 00:11:57:18
Unknown
What kind of different areas can you test within a website in the store? So I want to take a little step back and tell you that's actually precisely started as a personalization tool. And it was we focused on pop ups and looks like personalization and we just thought and I still believe it is that personalization is a very important, very important part of the customer journey.

00:11:57:20 - 00:12:21:18
Unknown
But to be honest is that any personalization, any pop up, any widget on the website has to be a bit tested to actually prove that they are impacting your website in a good way and you're not losing conversions. And I was just listening to another podcast of yours today and I saw about I heard about this popup opt in pop up that actually convert to conversion even that you think you are building go this you can have the conversions.

00:12:21:20 - 00:12:46:03
Unknown
So I think that's when we understood that we have to add a B testing at every step of what we building precisely. So first we added personalization over things and we added pop up a bit, think and it's not just pop up Davidoff and it's actual pop up baby testing where you can control against a control group, where you can control does, whether without a pop up or without a pop up generates more revenue.

00:12:46:03 - 00:13:17:23
Unknown
And this is very important because in advertising, I think conversion rates is good at some stages in the funnel, but in the end the bottom line should do revenue and even better profit. And that's where we pride ourselves as well as we report on all of that. And fewer years back, we decided that we have to actually step up our game and expand our product offering into a full fledged conversion rate optimization suit where you can actually avoid any piece of content, any price on the website.

00:13:18:00 - 00:13:38:16
Unknown
The URLs flipped over to us. So we pretty much have it all when it comes to advertising. For some customers, it's important to have a script API. We got that. So you can test ads code level for some customers, they just want to redirect between two offers. We got that you're seeing. You're off the test for some customers.

00:13:38:16 - 00:14:00:04
Unknown
They just want to just do different Shopify templates and it's very handy right now because the ship file theme editor becomes better and better and they can just build them there. And you first love to run the does. Some customers are migrating from an old theme to a new theme and have some agency telling them that they're going to triple that conversion rates.

00:14:00:06 - 00:14:27:23
Unknown
But yeah, we, we offer that film testing so that they can make sure that the revenue and the conversion rates are actually better. On the new theme and besides that, obviously we have our simple editor that you can use to run content us and actually use a visual builder to change images, copy buttons. And in all of that, what I've told here is obviously goes back to my to my idea.

00:14:27:23 - 00:14:49:05
Unknown
That's we give you all this tools and in the end, if you have a hypothesis, you will know which of this tools you will need and which ones will work better for each specific use case. And it's going to be very easy to implement when you know what you want to do. Mm hmm. I want to talk a little about timing and to.

00:14:49:06 - 00:15:17:15
Unknown
A lot of versions are right now still in the time of changing things on the website. And once it goes into Q4, they go into a sort of code freeze and try just to concentrate on selling. What's your approach? When's the best time to do testing and for how long would you let it run? Yeah, So in terms of timing, so my opinion is you the up to my opinion is you have to be always experimenting, especially your side.

00:15:17:17 - 00:15:43:04
Unknown
Q4 So you have Christmas like Friday, Sunday, Monday. People are kind of afraid to do major changes, but at the same time, the stuff that you learn during the other parts of the year might not be true during this time because people who are buying during Black Friday, Cyber Monday and so on are usually people that might be dormant for the year in terms of buying and they activate at that time.

00:15:43:06 - 00:16:05:21
Unknown
So it's still important to run a test during that time to extract learnings and apply them to next year promotions or maybe other promotions, because in the end any of it is a learnings that you can then apply to actually maximize the revenue. And thus this learnings are as valuable in Taiwan as they are available into free. So I think the tasks should be run always.

00:16:05:21 - 00:16:32:23
Unknown
Obviously don't start the dusting from scratch when you're just learning in November. It's probably better that you start well and start to learn the process and create actual a process in your company of experimentation earlier than that, so that when you come into this period, you already already but still don't stop experimenting because it's very important. And in terms of how much of to draw.

00:16:32:23 - 00:16:58:04
Unknown
And so I think there's various things that are circling around. So my suggestion is at least two weeks because obviously a week, some time, I'd say a week because you have Monday through Sunday. So all of this anomalies from day to day are being taken care of. But during two weeks, I think it's going to be enough to make sure everything is statistically significant.

00:16:58:06 - 00:17:25:12
Unknown
At the same time, obviously you need to have enough traffic. You need to have some say, at least 200 orders obviously depends on the market, depends on on your stats. But I think that's a pretty reasonable number to have in mind. But these are just like baselines at the beginning. Once you get into it, once you get working, once you get started, you already will know when to stop, when will be the time to start a new task.

00:17:25:14 - 00:17:53:02
Unknown
And this things will come naturally to you as you already have an idea of how your website, how your business responds to this experiments. Okay, Can you share some success stories or case studies of businesses that have worked with your solution and what kind of results they so, yeah. So as I said, we actually have a lot of case studies from the past with customers migrating from other platforms and using pop ups.

00:17:53:04 - 00:18:27:13
Unknown
Since this year we started working on more case studies in terms of advertising and what we loved was were Epitaph well pop ups with can be a little bit counterintuitive for some people because they think that they already optimize it to the maximum they use tools like leave you to run the pop ups, but then the and we have customer coming from quite a different personalize for pop up always was they have the ability to actually track the impact of the widget on their on the revenue.

00:18:27:15 - 00:18:53:16
Unknown
So we actually had interesting days where the people understood that they're actually worsening the conversions and bringing less revenue. So for me, this is still a good US, is still a good learning and you just look at this and try to understand is it still worth writing this pop up? Should I update maybe the message, maybe create a better offer or maybe hide the offer altogether, or maybe change the discount?

00:18:53:18 - 00:19:17:17
Unknown
So a lot of this question start to arise in your head when you understand when you see this data. And I don't say that you don't have to run this pop up for this customer from from, I'll work with him. You're still capital running the pop up. We just adjusted the kind of pop up, the kind of offer that he had so that it doesn't hurt the conversion, but at the same time, he still is building his list.

00:19:17:19 - 00:20:03:18
Unknown
So I think that's an interesting learning. We also have, again, a lot of customers running conventional tasks, so a lot of tasks like hero sections on the landing pages offerings. So for example, subscription versus non subscription and how well how well it converts. And what's interesting here is for subscription a bit testing, I shouldn't be taking into account only the first sale, but all the sales that come after and we had customers who have run successful subscription days and even if at the beginning they show us losing one because usually when you offer a subscription you will give a discount.

00:20:03:18 - 00:20:33:22
Unknown
So they so they kind of jump onto your subscription. At the beginning, you look at the data, it's a losing test, but as time passes, they make the second purchase for purchase, first purchase, fifth purchase, and then you see that the revenue is much, much higher than for for a simple, simple test. So this is an interesting finding that we had with our customers and actually lots of other interesting cases.

00:20:34:00 - 00:20:56:01
Unknown
But as I said, what I also love to see besides winning test is I love seeing losing taste and then using that learning to create much better taste and understand why it was losing and maybe maximize on to what we learned from here. It just made me smile. And it's good that you highlighted this. Obviously, you're learning also from losing to us and from losing results.

00:20:56:01 - 00:21:14:22
Unknown
I have to my in my own business a while ago I tried to optimize and it completely backfired. So the first version was much better. And that's that's a learning curve and that's why it is a B testing. So for our listeners, it's not always getting a better and better and better version. Sometimes you get over as it's just not as good and then you have to restart and you learn from us.

00:21:15:00 - 00:21:55:14
Unknown
Now who's your perfect customer? So our best for customers. Did you see brands that have more than 50,000 visitors per month on their websites, usually doing more than a hundred k k per month? Most of our most of our customers are in U.S., Europe and Australia. So we try to work with this kind of customers. A lot of times customers that are on Shopify that use Klaviyo and tools like all of our and others of a signal that this customers might be a good fit.

00:21:55:16 - 00:22:19:15
Unknown
So we use this data to kind of understand that these people are already advanced in their journey. They already optimize a lot of things. They already track things very well. So here comes a step where they need to optimize the actual stuff that's happening on the website. What we saw the typical onboarding process of a new brand, what steps are involved, how long does it take to get up and running?

00:22:19:17 - 00:22:43:02
Unknown
Yeah, so as I said at the beginning, it's one of the things that we pride ourselves in. So we always we are a people in team and we always thought about our product as a self-serve, as self-serve as possible. So you can I get started with personalization? We at each step of the way, we have kind of like a wizard where you choose what you want to do.

00:22:43:04 - 00:23:11:13
Unknown
So at product level, we try to make it as easy as possible to start the task. So it's actually taking just 2 to 3 clicks to set up a simple task and oftentimes that's enough. But again, with Avatar advertising, we love to be proactive a little bit with love to kind of jump in and help the customer in and see whether what they're trying to do actually makes sense.

00:23:11:15 - 00:23:42:17
Unknown
So we try to be like an extension of this self-serve onboarding process where as a side, they can start a task in 5 minutes. But we try to hold the ad and make sure that what they're doing is correct. Yeah, I like this approach because I think not everyone comes with a full blown experience on what they want to do and what kind of idea they want to test, and then having someone on your side to help you with that is a good idea.

00:23:42:19 - 00:24:09:02
Unknown
Now tell me about your pricing structure. How does that work? Yeah, so our pricing is based on the number of visitors that you have on your website. And we have basically two tiers. So the essential tier is based basically for customers that want to just get started rather than pop ups are on some little widgets on the website, maybe surveys, maybe surveys their customers and understand those data points that are very important.

00:24:09:02 - 00:24:37:16
Unknown
In the end, as a startup. And then you have the premium plan where you have everything included. And all of this is tiered by the number of visitors that you have. So the more visitors you have, it's going to be more expensive. And we a lot of times have the objection that we should build based on the visitors that actually see that us, because some customers are only running the decimal single page.

00:24:37:18 - 00:25:00:10
Unknown
And what we answer to that is that you should be actually testing as much as a website as possible. And if you do that, this number is always going to be very close to your total number of visitors. And with customers, it's actually use our platform to the maximum. I think. It doesn't make sense. Mm hmm. So, Joe, before coffee break comes to an end today, is there anything you want to share with our listeners that we haven't covered yet?

00:25:00:12 - 00:25:35:12
Unknown
I don't think I have like a specific golden nugget for them, but I want to just say that to all the eCommerce operators and marketers to always be experimenting, to always take care what tools to use. And as I said, always try to find the right tool for the purpose. Try to look at your data. Don't lose faith when you see a losing experiment because it's still a win unless you stop after it and keep on experimenting.

00:25:35:12 - 00:25:57:05
Unknown
And then in the end, I think we can maximize the value of the business as much as possible. Totally. Where can people go to find out more about you guys? Yeah. So you can find precisely at Bruce Lively dot that we all saw on the show five store. You can follow me on LinkedIn, so do Kazak and on Twitter or Excellent.

00:25:57:08 - 00:26:16:05
Unknown
I'm happy to connect and hopefully teach you more. I will put all the links in the show notes as always. Then you're just one click away Central. Thanks so much for giving us an overview. I'm a big fan of testing and experimentation. I do this every day and the learning's coming every day and even I'm doing this for 25 years.

00:26:16:06 - 00:26:32:11
Unknown
There is always space to learn something new about your customers, about your funnels, and about the things that you offer. Optimize and backfire. So I think a lot of people just should reach out to you, learn about it and learn about their own business and hopefully get better conversion rates. Thanks so much for your time today. Yeah, thanks a lot, Class.

00:26:32:13 - 00:26:33:00
Unknown
That's a great.