No Two Journeys Are the Same

NOTE: This is a raw transcript and contains grammatical errors. The curated transcript will be uploaded soon.

Hello, everyone. Thank you so much for joining us for one more session at VWO Convex twenty twenty five.

Today we receive a very special guest, Dorian Gonzalez, manager for experimentation and personalization at UI. And we are going to talk about No two journeys are the same, which is very intriguing. Super excited to receive you, Dorian. Thank you so much for accepting our our invitation. Please introduce yourself. Tell us a little bit about your experience experience, please.

Thank you, Daniel. Hi, everyone. My name is Dorian. I currently work as a manager at Before that, I worked for Samsung as a manager for personalization and experimentation. So my focus is in improving customer journeys.

Before as I mentioned, I worked for Samsung for several years. And before that, I worked in another e commerce company in Netherlands. Currently, I moved back to Mexico and that’s the reason why now I have clients for the LatAm country and for the US.

Nice. Super nice. We definitely choose the right person to to be with us and and talk about this very exciting topic of personalization. My name is Daniel Nedef. I’m gonna be conducting this this session today. I’m a partner manager for Latta at VWO. And I can tell you for sure that personalization with AI, I’ll put these two topics as the the super hype hype today in LatAm, the the experimentation world.

And it’s nice to see that you’ve been working we’ve worked in campaigns for Samsung, and you have a lot of experience in this. So just please share us share with us what’s your vision for the future of personalization? Now, how will AI and the experimentation reshape how brains connect with people in the next, let’s say, five years?

Even though five years is a long time to So what I think that is going to happen is that we are going to move from manual personalization to autonomous experimentation, which this I mean that the AI is going to start to help us to not only to segment, but also to optimize.

AI is going to be faster than anything in testing and optimizing. And therefore, in five years, I see that we are going to have journeys in which every touch point is going to be personalized and the personalization is going to be seamless. So the vision that I have is that because AI, their teams are going to stop to focus in these repetitive tasks that we have, And then there is going to be more space for productivity over the journey.

That sounds super exciting. That sounds super nice.

One thing about personalization, let’s say one challenge that teams and companies usually face is about how do you transform this bunch of data that’s most of the case fragmented data into more seamless and personalized experience for user who expect more every year, right?

Before jumping in how I do that, I want to say that in most of the companies that I work, I found that the main problem is that we work in silos. So, you have the traffic, you have the tooling, the budget, but because every department pushing different strategy, sometimes because we are not working parallel, we lost opportunity. So the way in which we do, knowing that we still have this problem in many of the big corps, is that we start small. We start experimenting locally.

Even though we have the data that is globally, we prefer to start locally. And then once we build a strong basement about what we do or region or locals, we try to expand that local learners to the global areas. But yes, I think that how we do is we first start small. That’s the key that we do.

That’s nice. Let me just add something because that’s a super interesting view, which also we face a lot in companies, in with our projects at VWO in LatAm, is the silos, right? And both the silos and how to expand this from a small group or a local group experimenting to a more global project.

How do you feel in your experience? How is this expansion received, let’s say, from the local experimentations to transforming these into a global program and eventually getting to personalization? How does the leadership, let’s say, receive the idea and how can you really personalize these in the field?

What happened a lot in these big companies is that, yes, we have leadership and sometimes leadership just drop ideas that sometimes are not based in data.

An example that we have is that a long time ago, they did a UX revamp in which they released this new card that looks nice. It was beautiful. It even looked better than the previous one, and it looked at that was easier to use for the user. But then what we start to see was a drop in the card.

And then how we took that to leadership, we always For leadership, we need to understand as a team that we are data team, but they are revenue driven people. So, you always need to tally money to them. So, you need to be like, Okay, this is happening. But if you arrive with all your data and your dashboard, they are going to be like, Yeah, cool story.

It was web. So you need to tell them, This is what is happening, and you are losing all this money for this. So I recommend that before releasing UX Rebounce, what if we just test it or experiment it or do something with it. So, you need to approach these people talking about money, then once they realize what is the losses that they are making, they are going to start to listen to you.

And another key is that you don’t make big changes at the beginning, start small. Be like, Okay, according to my data, the problem is in the car. We don’t know yet exactly what is something in the car. Start testing.

Keep them informed about what you are learning. And for example, in this example of the car that we found is that when they did this more beautiful car, they highlight more the feel of the coupons.

So what was happening is that people was dropping off because they were looking for coupons. And then what we first we we start testing by making it more high data, hybrid. And what happened is that we reduced the drop, but still there was some drop. And because there was a global release, locally we cannot suggest such as I mean, we can, but the fact that they listen to you is very hard because this idea is coming for the headquarters.

So, what you do is start to negotiate with them and it’s like, Hi, guys, can okay, we are going to keep your huge fill of the coupon, but what if we add now next to the coupon above that promote a promotion? So, the user don’t need to leave anymore because they’re next to the fill. You have the promotion already there. And at the beginning, we start with this small test.

At the end, this turned it in a personalization. Now, it’s more complex because linked to that promotion model, now we have a recommendation model that what it does is a propensity model linked to promotions. So, what happened is that according to the user, it’s a different promotion. And then if we that one, we for sure reduce the drop, then we show that even instead of having drop and losses, we start to have even revenue.

So, what we did is show all these results, show the data, but not deep data, but more about in terms of money to the leadership. And then the leadership start to like the idea. And that’s how a local idea can become a global idea.

That’s very nice. That’s very true. We have five to eight minutes into it. And I think we have two very nice takeaways. Start small and expand and talk money.

Exactly. Talk money. Talk money. And also storytelling. Storytelling is sometimes I feel that it’s a skill that a lot of people don’t think how big it can be.

And I have teams of really smart people that works very hard, data analysts. But what I have found is that while the people get more into data, more hard for them is to explain what they found. And sometimes they come with this huge deck full of dashboards. And let’s be honest, leadership doesn’t have the time to sit and check all your numbers. So try to make it just point. This is what I’m seeing and this is what is making you in terms of loss.

Nice. Super interesting. Super nice. And what’s what’s one, let’s say, one experience or or an experiment or a tactic that you’ve seen in your experience that took personalization from something that’s nice to have, which right now, as I mentioned, personalization is kind of hyped. So everyone’s talking about it. Everyone thinks it’s nice to have.

But I feel in my experience that many people don’t genuinely saw and really understand the transformational power of having personalized campaigns.

One experiment or tactic that you’ve seen that took personalization from these nice to have space to genuinely transforming both business and customers.

One experiment that is very simple was, okay, in Samsung, when you sell a mobile in the website of Samsung, you don’t get a charger. And this happened in many companies of smartphones. I didn’t know that. Of course, if I didn’t know that, users don’t know this. So what happened is that a lot of people place the order, put their smartphone in their car, they inspect the smartphone with the charger. When they get the smartphone without the charger, they get angry. So we get complaints, returns, people angry, screaming in your chat, putting Trustpilot reviews, all negative.

So, what we started to do was testing to add more blocks explaining, hey, there is your charger. At the end, this turned in a personalization that was really impactful because what we found is that in the recommendation model for the products, so right before you After you place the order, you have this ads on page in which you can add accessory.

We check it. Okay. Only people that is new, so those users that didn’t buy a smartphone before from us, so they didn’t know about this no charger situation. And also users that didn’t have yet a charger in the card get a pop up in that accessories page. Well, that was not a pop up, was more like an extra card that was in page, filtering page, where you have the listening I don’t remember the name, but it’s a page in which you have all your products, all the accessories. In that page, instead of a pop up, what we did is move all the products and then you have this extra card of product in which you have this mention that say, Hey, don’t forget your charger that is not added with your packet of the smartphones. And then this is a very simple one, and it turned out to be very successful because it reduced the returns, complain.

So we show and the NPS numbers were up. So we show, like, three results, and it was very simple, honestly.

That’s that’s interesting. That that I think that’s a nice message because sometimes people that work with experiments and personalization that in the CRO world, I was trying to find a big idea and something like really right? Really revolutionary.

But sometimes the the very simple ideas when you think about when you have the the the user journey in mind and you’re backed by the data and you’re thinking about talking money to your leadership, all these three together can can really create a a successful way to experiment and and campaign.

Yeah. But I think that a lot of people is not seeing that before personalization, there’s there must be experimentation. Because in order to have your personalization, the best personalization, the more optimized one, you first need to experiment lot with that personalization. And also, before personalization, also needs to be segmentation because you need to segment every user to have a different experience. So once you have your segments, you start to test in every segment different experiences. And then when you find the best experience for that specific segment, then you can start to call it a personalization.

That’s nice. That’s awesome. So the to have the the the comprehensive process, right, of of CERO, that’s a message that we always send to to to people we talk with is the importance of having a methodology and and really following the entire process instead of just jumping into one or another, really starting to have insights and then testing a lot and then personalization now comes almost like a consequence, right, of good testing and having all this bunch of data.

Yeah, because also when you experiment, you have results. And then from those results, what you can do is segment your results. So see, okay, it was a winner, but it was a winner for desktop and mobile or only for desktop. And then in that way, you start to personalize before personalizing because maybe it was a winner for mobile, but a loser for desktop. So why are you deploying this for all visitors when it is only good for one segment of those users?

That’s true. That’s so true.

Thanks so much. It’s been super insightful so far.

Let me let me just get to a topic that I’m really keen about, like, about creativity. You mentioned a bit you you you spoke a lot about creativity when we we talked about AI, how AI is going to open space for people that are working and getting rid of more manual routines and be able to focus on more creative ones. You also mentioned the importance of data when talking to leadership and showing data in order to convince and to expand programs and to adopt.

How do you how do you blend creativity and data to design personalization strategies that, let’s say, surprise and delight, right in rather than just recommending, let’s say, what’s expected? How do you blend creativity in in data?

I I see data has the what. So what people is doing.

And then I see the activity, the why. Because at least that you do a UX research and then you see all your users and then you ask, why are you doing this? Maybe you can get some insights, but you don’t have all the insights of all the work. So that’s when you need to become creative. And then through data, you see what is happening. What is happening is that users are leaving my car.

Why? Then there is creativity. You need to start to become creative and be like, okay, I think that it’s because this, I think that it’s because that. And then you create this list of assumptions because all of them are assumptions and hypotheses that you need to test first.

And then there is the creativity part. All these assumptions are purely creativity. And, of course, they are backed up by data because you know the what, but the why is creativity.

That’s nice. Let let me just just open another tab in this topic, which is the tab of Test. I I don’t wanna I don’t wanna say failed tests, but tests that do not reach the significance. Yeah, not even the significance, but tests where the variation do not perform better than the control.

Right? So yeah, it’s let’s say it’s a failed test. How important is that for in this process of creativity? You know, like being able and having room to test your ideas and eventually to fail in order to reach a good or winner one, let’s say.

That’s something that you need to develop a lot when you are working in this area, resilience. Many of your ideas, and this role made you so humble, because many of your ideas you come with this, oh my gosh, I’m sure this is what is happening. No, there are a lot of things fail, completely fail. So, I have becoming very humble and empathic.

I have more people in my team, so if they bring an idea I need to know to be a fail, I never shoot them down and imply, Yeah, it’s a fail. You can never come with ideas because they fail. No, I start to ask questions. Is failing in all the segments?

Again, start the segment. Maybe it’s failing only for this specific group of visitors. It’s honestly very weird that an idea fails in all the segments. Sometimes what happens is that it fails in one segment and what happens in the other segments is that the others are flat, so that there are no significance.

And then what I ask is like, don’t drop this idea. First, iterate in it. So, start like, okay, now I have the same assumption, but then now I will test in a different way. Now I will test only for the ones that were flat.

If also all the flat turned negative, then that means that the idea was a fail and it’s okay. We are learning. So because it’s like when you have thousand of ways to find a solution, but there is only one solution. So if you start to mark all these thousand ways, there is going to be a moment in which you only are going to have ten, and then you increase the chances of find the winner.

So what we do is, okay, if it’s a fail field segment, if it’s a fail everywhere, then we start to make it clear this is a fail because we learned that adding this model create drop in conversion. But if DOS is not failing in all the segments, test it again until you prove me that it’s failing in all the segments. But because what happened many times is by these interactions, they turn flat or they turn positive. And then we start to learn, Ah, it failing in that specific segment because that specific segment don’t like this.

And then it, again, turn it in a personalization. For that segment, let’s not do that. But that doesn’t mean that you don’t do that for all visitors, only for that specific segment.

Nice. Super nice. Super And what what what kind of signals do you look for in this in this noise of data and hundreds of thousands of traffic and users in the website?

Where do you identify new personalization opportunities?

Any moments where user behavior will open your assumptions? How do find these, let’s say, golden nuggets hidden in the noise?

I look for contradictions. So I am like, okay, for example, the card. All global teams were thinking by this because, of course, they send the design to all the local teams and all us we sign off that we were okay with the card.

So, we were expecting good results. So, when it drops, then what is happening? So, this is the moment in which you look for opportunity. Okay, if the data is acting contradictory to what you were expecting, then this is the moment to go deeper in the data and find why it’s contradictory to my expectations. And then in that moment, you find a lot of opportunity to optimize something.

That’s interesting.

That’s interest. And how how how brand because we are talking a lot about segmentations, right? I think segmentation is really the core of of having a successful personalized campaign. It’s properly segment and then test a lot of testing should deliver a good journey for the segmented audience.

How brands can move beyond just simple segmentation to a more dynamic, let’s say, living personalization that adapts in real time situations?

What is happening currently right now is that segmentation is static to what happened is that the segmentation is done by the user is from Brazil, the user is thirty years, the user is male, so show this banner to the user. But this is a static. So as companies can move more like in personalization is by real time data.

So you start to track what the user is doing right now. So if the user is looking for this specific product and then the user did that specific user behavior, then deliver this. So I would say that in order to move beyond in personalization, real time data is what is the key.

And that’s nice. And what role does automation play in this scaling scaling up personalization while keeping human? Right? Because I think now nowadays, everyone, it’s already kind of I trained it to see when something is just came out of out of nowhere from from a AI agent or or just some kind of bot or some kind of automation.

I think we all know we all receive emails and we all receive even LinkedIn approach and that’s that’s clearly like unpersonalized, tries to be personalized in in the format. But, you know, in the content, it’s not and it’s just it’s just so it becomes so evident that creates a negative feeling. Right? You automatically lose your your focus and your your attention. What how does automation play here in into scaling, using all this tool and all these this AI now for scaling personalization, but also keeping it human? Like, can you share a success or pitfall case on this on this on this front?

Okay.

There are different ways in which personalization, sorry, automatization can help personalization while keeping it humans. I see this as two different areas. One is inside your team. So, for example, automatization can help you to automatize a lot of the things that your team needed to do, like data analysis or, for example, reporting.

All these repetitive tasks can be done by automatization, and then your team has more time to be creative, which is something more human. So, this is in the team and now have a user. Automatization can help you to distinguish those segments of users that are more constrained by, more worried about privacy. So, there are users that don’t want to see promotions.

There are users that don’t want to see banner.

Okay, then you create this segment and then you offer them an experience that is less marketing aggressive, and that is more careful with their privacy and that asks less questions.

Nice, nice. And we are heading to Unfortunately, we’re heading to the final questions of our talk.

I feel we could stay here talking for three hours.

Unfortunately, this won’t be possible today, maybe one day. But again, looking ahead, how privacy and customer trust evolve in world of hyper personalized digital journeys. We just spoke about automations and how some people in some company are using AI just to mass shoot, message and emails and everything.

How do you feel this, like looking ahead, how these topics of privacy and customer trust will evolve in this world of hyper personalized digital journeys.

I think that in terms of data and privacy, users are going to be more open to share their data if companies start to be more open about what they are going to do with their data, and also if companies start to offer better experiences by using the data. So I think that the companies that are going to start to lead are those companies that are clear with the user and they stop to use these small text and boxes, all in gray, that don’t let the user to know what is going to happen if you provide all your information.

But if they are clear and they specify, Hey, we are going to use all your information to provide you a better, I don’t know, better category of products to make your life easier than when you want to find a product, it’s going to be dead in your face. So, if companies start to be more clear about what are we doing with their data, then users are going to be painful, and then even users are going to start to provide. For example, in my case, Amazon sometimes ask me a lot of questions, and in order for me to have a better recommendation model, I’m all the time replying, you know, I need this data. And I’m open to provide my own data because I know what they are doing with it. So, I think that if the companies start to stop to see privacy has a wall, but more has a principle, then those companies are going to Companies and users are going to start more They’re going to build more trust regarding privacy.

That’s that’s a super, super nice, super nice insight, Doreen.

Thanks so much. And for teams that are starting now, let’s say, and teams that wants to build productive personalization, what are the foundation skills? What’s the most have or skills or tools that you believe are game changers for the next few years ahead?

I think that because before focusing in tools, we need to build serious communication, as I mentioned before, because currently in all these big corps, which I work, the problem is not the budget, neither the tooling, the silos. So, I think that clear and clean communication is something that is a strong skill that these teams needs to have. A second one is storytelling, as I mentioned before.

Another is statistics, experimentation, literacy. So, before personalization, understand why things are happening, the statistics behind that, is it causal or is correlated? Why this is happening? And the last one, I will say that you need to have human centered thinking to be empathic and to think as a user before thinking as a worker or a company.

That’s awesome. If you have to paint a picture of personalization done right in twenty twenty eight, let’s say in three years, what does it look like? How for brands, for customers and for for the teams that that are building it?

For me, it looks like invisible, seamless, trusted. So it’s more like a friend that doesn’t need to ask you when is your birthday or what do you want for your birthday. It knows.

So that’s it looks like, yes, someone knows you and you don’t feel like, oh my gosh, why my friend is asking things to me? No, they just know you.

That’s awesome. That’s great.

Dorian, thank you so much.

Just gonna make one last exercise while you were you were sharing your insights and your experience, Preacha, thank you once again. I think it was super helpful. At least for me, I just learned in thirty minutes a lot, a lot of insights, many things that I will take to my daily discussions with companies and with people working with CRO in LATAM. But I just took some a few notes. So tell me if you agree or not, if this could be some takeaways from our session. Is that good?

Okay.

Start local start local glow global. Is that a good good sentence?

Yeah.

It’s a good Gain leadership by talking money and turning data into storytelling.

Perfect. Yeah.

Look look for contradictions.

Yes.

Use real time data.

Yeah.

Let humans be creative. I love this one. Be true to your users and have privacy as a principle. I also love this one.

Yes.

And if you think as a user, personalization done right looks invisible.

Yes.

Is that it? Do you wanna do you wanna add a final message? Dorian, thank you so much for your time, for your presence. I do hope to see you again in some other session.

Just leave us a final message Thank you for having me.

And it will not it will be nice to collaborate in the future. And let’s see. Maybe our paths are going to cross in the future.

Thank you. Thank you so much, Darian. Thank you for everyone who attended. I hope you’re enjoying Convex twenty twenty five.

Super exciting to bring the best minds and experience in the market. I was my name is Daniel Nedev, partner manager here for VWO in LatAm. I was here with Dorian Gonzales, manager of experimentation and personalization at UI. Thank you so much, Dorian.

Thank you. Bye.

Thank you. Bye bye.