Insight-led Personalization: Whys and Hows

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

Hello, everyone.

Welcome to Convect two thousand and twenty four, by VWO.

This is the day three and the final day of the event, and, I am feeling mixed emotions right now. My name is Rupal Bansal, and I am the senior marketing manager at VWO.

Thousands of brands across the globe use VWO to optimize their customer experience by gathering insights, running experiments, and personalizing their purchase journey.

I am honored today to have Chris Gibbons Chris Gibbons, sorry, with me here to run the third and the final workshop at Convex two thousand and twenty four. So, Chris, can we have you on stage, please?

Hey. Hey. Hi, Chris.

Really excited to be here. Thanks, Vipul.

Thanks for the introduction.

Absolutely. And, I’ve been an audience to the past two workshops, and, they were amazing, by Craig Sullivan. And yesterday, it was by Erin Weigel. And the, the level of interactions, the level of engagement from the audience levels was, you know, something that I’ve never seen before. So I’m expecting the same level of engagement and the same level of energy, to happen, just I mean, something that we can witness today as well.

Yeah. No. That’s fantastic. Yeah. I’ve got a lot of, things to go through all around personalization today. So the topic is is insights led personalization.

Yeah. And so so, yes, let let let’s kick off. Actually, are we waiting for anyone else, do you think, Paul?

No. I mean, we’re just waiting for a few more people to join in. And, yep, I see the numbers are now rising, because the calendar invites, just went out, I mean, few minutes back, and people are now, you know, joining in. I already see a lot of people have, synced where they are joining in from. But, yeah, it’ll be good to see where everyone else is joining in from. We have people from Scotland, Denmark, France.

We also have people from planet Earth, Douglas.

And No.

It it’ll be great to do a word cloud. So if if everybody does put where they’re from, I reckon we can do a really nice word cloud from all of this.

Yeah. It’s it’s, unfortunately, that’s not a feature here.

But It’s a feature for afterwards, isn’t it?

Yeah. Afterwards, definitely.

We there are many new many, names that many repeat names rather than, seeing in this third workshop. A lot of people like Marcela Sullivan, Betty Kim, Julie Frigoso, Sharon, Gabriela, Jesse.

A lot of names I I am able to recall from our past workshops.

Oh, that’s great.

It’s a good sign, isn’t it?

Good good retention there.

Yes. Absolutely. And I’m this is quite, quite a testimony that, yes, the workshops are delivering value, and, people are finding value in these workshops, and, they prefer coming back and participate and to participate in the new in in in each workshop. Sorry, guys. I’m fumbling a bit. It’s been a, tiring day today.

But, yeah, we’ll start, just in a few seconds from now.

Fantastic. So, welcome everybody. Really happy to be here.

I think this session should last an hour, I reckon. I’m gonna try to keep it to an hour, but that all depends how many questions you guys ask.

I’m gonna start off by telling you a little bit about us, but probably just for a few minutes.

So I’m Chris. I’m a chief experience officer at Creative CX, and we’re an experimentation consultancy.

But, of course, as part of experimentation, we do a lot of personalization as well. We help businesses to establish, improve, and and scale their experimentation programs. So real specialist consultancy based in London. And this picture you can see on the screen here is is some of the team. We’re about thirty to forty people now.

And you can also see just a few of the people we work with in terms of clients, some big names there, a mix of retailers, finance people, and, travel companies. So a nice mix of different companies we work with.

But most importantly, I would love to find out a bit about you.

So, actually, what I would love you to do is to, for a start, it’s been great to see where you’re all from, which is amazing.

But I’m gonna start the poll, and I want to find out first how many of you are currently AB testing.

Okay. Fabulous.

So any more votes?

And then how many of you are personalizing at the moment? Because I think it’d be quite interesting to see the comparison.

But we’ll move on. Right.

So we’ve got three parts to today. First part are introductions.

This part is an introduction to the insight led personalization theory and examples.

So what I wanted to do, first of all, is to, talk about the whole reason. Why on earth do we want to, to personalize in the first place? And and I always find this a very interesting, kind of way to talk about it.

And, actually, personalization is so often used purely for marketing and for ads, you know, very much pushing those messages out. But, actually, I believe that it can make a massive difference to the customer experience.

And, ultimately, it’s it’s about increasing overall customer centricity and improving business performance.

So while we have the one size fits all approach at the top, we have to make a lot of compromises.

We have to effectively what we’re doing with all everything we do, we are effectively, we are actually, we’re we’re sometimes making experience better for some groups of people, but at the same time, making experience worse for others. So we end up having to make a lot of compromises.

Whereas the the amazing thing with personalization is that we can actually start to, create tailored experiences for different important groups of users. And, overall, that means that we can get the overall customer centricity up. So we can have fewer compromises.

We have the ability to meet the needs of a higher percentage of users.

More effective and bolder marketing messages, actually, because if you know who you’re marketing to, it means you can be bolder to that audience. Whereas if you’re marketing to everybody all in the same time, you have to, like, basically compromise and make it bland and less strong.

We can be, we can focus on all channels. We can connect up the online and install and make it work better and ultimately drive that competitor advantage.

So some very high level, reasons why we personalize, but, actually, it’s quite important to have these things in your mind. And then if we move to the like, couple of things which are really important for personalization reminders, I call it.

So delivering personalization doesn’t mean that you can skip experimentation.

It’s it’s amazing how many, how many people I speak to who, when it comes to personalization, they forget about AB testing everything they personalize.

They assume that if they’re targeting personalized messages to groups, that it’s always a good thing and that it’s better for not personalizing at all or or better than not just talking to people in the same way.

It’s so important that you should we need to AB test.

And you should apply best practices such as you should conduct your research just like you would in AB testing. Do your research and analysis first to identify opportunities for personalization.

And what we’re looking for is contrasting needs and behaviors between key audience groups, and I’ll come back to that actually. So it’s the difference in user needs and motivations and in intents between different groups.

The second point is know that personalization only works when it’s solving a genuine problem. So, again, a bit like when we when we try to do when we with our experimentation programs, of course, we need to focus on problems because, if there’s a problem and your solution actually fits that problem, you’re more like more likely to solve it. The same is true with personalization as well, except that the problem is is a really big difference in user needs.

And when you’re planning your personalization campaigns, make sure you got a problem statement and a hypothesis. It might sound obvious, but it’s it’s amazing how much personalisation is done, and there’s no, there’s no, hypothesis at all.

And and the last thing is always test your personalized experiences against the control.

So some quite clear things, but, but I think they are very important reminders.

And going back to the the other point here that this is before we move forward, the bigger the differences in user needs, motivation, intent, the bigger the opportunity for personalization.

So this is a really important one. And, actually, if there’s no difference, there’s no point in personalizing. It’s not that this it’s a waste of time, actually.

So this is a really important thing to get in your mind. So when you start looking for opportunities to improve the experience through personalization, look for those differences. Look for those contrasting needs and behaviors between one group, say, new visitors or, say, existing customers, for example, and look for where they have completely different needs. And that’s and that’s why you should need to zone in.

So is that making sense for you? Obviously, in in the chat, I’d because I think I think personalization has been, misunderstood and has different kind of different meanings sometimes.

I I think it’s a lot of people, it’s very tool driven. It tends to be in very kind of quick win, very tactical, where in our experience, the the best impact has come from being insight led and actually thinking about your customers or being customer centric.

Okay. Great. And I thought it’d be it would be worth talking about some, some themes. I always find this quite helpful.

So what we’ve done and think of it as examples.

And all of this is going to be inspiration and fuel for you for the next interactive session.

Exactly. Like Amy says, problem and solution. Exactly. So some really good good messages here.

K. So the first one, this is a really good theme to think about, and it’s it’s very fitting for mobile especially. But this we call it session continuation.

In other words, it’s picking up where you left off.

So, don’t make users start from scratch again. So think about it. And when you look at a lot of, data like we do, actually, people buy or people book in the second or third or fourth or fifth visit, not in the first visit because, you know, they’re on their mobile, they’re walking down the street, they’re at the bus stop.

They’re not and they’re gonna be distracted. They got other other things in their life.

So it’s it’s so important to have bear that in mind. And then this is where personalization can come in because you can make it easy for them the next visit and the next visit to continue their journey rather than starting all over again. And this is absolutely key. So on the right, we can see with this fashion brand, the next time they came to the site, they would still see they’re still interested, and it would show some of the products they’re interacting with. And we’ve also experimented with a lot of ideas around, helping people to continue where they left off with their search. What were they doing last time? And, actually, you can see on this occasion with this brand, where for users where they have something in their basket already, this overlay would appear to remind them that there’s something in the basket, and they can continue their journey of buying that item.

Another really good one is the, you may have seen it on travel sites before. So when we’re talking about session continuation, sometimes on travel sites for the second time or third time you go back, you can easily kind of redo the whole journey again. You can you can again it it it is it’s smart and it and it remembers you, which is great.

So here’s an example here where you can actually, on on this website, when you come back again, there’s a continue your search up here so I can straightaway look and start off looking at Los Angeles, which is the search. So I don’t have to start brand new again. And and below, we’ve got still interested in these properties. So, again, what they’re doing a really great job of personalizing is to is to put the properties there, which I was looking at before. And if you think about it, actually, that’s cut out, like, five of my steps or five of my actions as a user.

The next theme is the is around implied, yeah, so implied preferences.

So this is all about implied personalization is different to explicit personalization.

Implied means that, through your actions as a user, you’ve told the system, you’ve told the brand that you like to, sort your listings page from low to high because you’re very price sensitive. And, actually, if you’re smart with personalization, you can learn that, and then you can you can take that information and tell the rest of experience because you know it’s this group of users is now price sensitive.

And then, there’s other things like color preference. Let let’s show them this color product by default from now on once they’ve selected the color. So it’s the it’s the changing the default and the implied preferences.

Or if people click on assembly and packaging on, you’ll know that they are more interested in the details. So, therefore, you can default that next time they come to another product. Again, this is cutting down steps, and this is improving the overall journey.

And then we got explicit preferences, which are especially important for logged in. And you can see here, this is the BBC trying to encourage me to log in by signing in. But the the interesting thing with with explicit is that you can ask your customers sometimes. Sometimes you can ask them if they’re a business user, for example.

And this is a great way of finding out exactly what they want. You can ask them what type of, what type of things that they need and they want. And and then you can, you can be very transparent with personalization as well. It doesn’t have to be implied.

Then when we go to prioritization of content, this is all about, the order of things. So as we can see as people who use heat mapping, as if you use the VWO heat mapping tools, for example, or other tools, where things are positioned make a massive difference to how much, how much interaction they get.

Makes a massive difference. So, therefore, when you can when you know through implied preferences what type of things they like, like the like the pricing, for example, you can change the order of of pages after that. You can put more price sensitive content higher up. So you can improve the effectiveness of landing pages by changing the order. And the same is true from from simplicity and decluttering.

I bet, a common problem with everybody is the fact that they got too much content to put on their website.

And as a strategy with personalization, it can allow you to even remove some websites, remove some of the clutter if you know that is not relevant to that particular user because of the information you know about that user.

And that’s a very well known thing, targeted sales and promotions and content.

So this is probably the more well known one where you can be more you can tailor the message to the audience when you know who you’re talking to.

And, therefore, you can rather than, again, put a thousand messages to everybody, just tailor it. And you can, like I said before, be stronger in your messaging.

And the most famous one of all is recommendations. And I know people almost a lot of people who who do personalisation only do product recommendations because they have a recommendations tool.

And it’s you just let it go sometimes, don’t you? But, actually, there’s it’s it can be really important. It can be.

It’s something you should AB test, and you should always optimize the algorithms, the placements. For example, a a a really good one that can work is for those coming from Google Shopping.

So you know people who if you have a really high bounce rate on your product details pages, it’s often because of people coming in and landing on that page is but then they’re in a completely different, mode of a mindset to those who are, say, coming from a PLP or coming through the normal traditional journey.

These people from Google actually are not that bought into your brand yet, and they’re just comparing products. They’re using Google as their landing page. So they they they hop on your site, and they go back and check another site. And every single time, it costs you money. And that’s where things like recommendations can come in as long as they’re higher up the page and very visible to people.

And then this is a really interesting one around journey intervention intervention and assistance.

Because, actually, it’s it’s really interesting around what we can do with the technology these days. And I think this is where things are going to move, where we can start to rather than pick up on insights and then at a later time, optimize the website for their for their, specific needs.

What we’ve been experimenting in this screen grab up here is actually we connected the analytics tool that picks up on struggle.

We connected to their live chat so that when people are actually chat are actually struggling on the page through their actions, it opens up a live chat and provides contextual help around the area of the page they were struggling with. So it’s a really it’s a much more the intervention is a really powerful, arm of personalization, which has a big future, really, because why not solve the problem there and then?

And then finally, the omnichannel side of things, which is, which is extremely important, especially when when stores are so important for people at the moment. So it’s it’s it’s so important to think about the whole end to end journey and not think of things as a single channel at a time. And some of the biggest opportunities are are all about, we’re connecting up these two areas. So for example, we’ve been doing some interesting work around QR code journeys in store and making sure that people can sign up really well and join up the whole process.

We’ve also been doing our user research across usability testing online, but also then following up in store as well and doing real world sessions in store. And there’s fantastic opportunities for not only connecting up the experience, but also for just improving, what people are doing on the website.

Because not everybody goes to your website with the intent on buying there and then in a single session. Actually, a lot of people will be there to do researching or that or that to research and plan their trip into store. And, again, through personalization, you can pick up on certain implied implied behaviors, and then you can tailor that experience afterwards.

So that’s a lot of the theory for now, and I wanted to go through that and and try to hold all those information. And, actually, it’s all on the Miro board, so you can go back whenever you want to.

But this is about, inspiring you for the next stage and for the insights led personalization process and exercise.

But before I carry on, let’s have a little bit of a breather. Has anybody got any questions?

I’ve seen a few on there.

I think there’s a question around is AI being used for personalization yet? I know VWO has AI for testing and suggestions.

So, absolutely, AI well, more machine learning is used for recommendation engines a lot, Joel.

And it’s also starting to use in testing tools for coming up with content recommendations.

I know we use AI a lot for, we we use it a lot for, our research and survey and analysis. Sometimes it could be really useful when you’ve got, like, thirty or forty thousand responses in a survey as long as you know how, how to use all the prompting.

I think, okay, I think one of the one of the challenges for personalization is is about how detailed you get, how small the segments are, how small the audiences are that you that you personalize for. And and I hope that AI will make it possible to actually, address more groups’ user needs. Whereas if you’re doing it in in a traditional, traditional way, you have to focus on the biggest groups. But in the in in the future, I believe you can be more, granular. And I think someone was talking about that too, which is great.

As Ajal was saying, thinking more granular per customer versus a broader AI recommendation. Absolutely. I mean, I think for the it’s just that granularity, being able to get closer to their needs. Because, of course, just imagine if you’ve got a thousand groups, you can’t possibly manually create different messages for everybody, but that’s where you can use tools like AI in the future to actually be a bit more intelligent with how you how you deliver.

Yep.

Sorry. I just, I just came to inform that, we had published the poll on Zuddl itself.

And, if you want me to show it on the screen, I can actually open it up on the screen for you to see.

Fantastic. So so let’s let’s do this poll now and see how it goes during the briefer.

So eighty eighty one percent eighty percent now have, you know, voted in the favor of yes. They are currently experimenting and AB testing.

Seven percent, plan to do it this year plan to start it this year rather, and six percent plan to start it next year. Eight percent are not currently, experimenting in a b testing.

So I think it’ll be interesting to know, Chris, because it’s just three minutes three months, sorry, before the end of this year.

What, any sneak peek that people might want to give into, their plans for starting, AB testing?

Why do they feel the need of starting it in the first place?

Yes. Yeah. Absolutely. Like to speak up or share their observation in the chat.

And meanwhile, I, I see an interesting question that I’ll put up on the stage for you to answer, Craig.

Does in app personalization make sense, or is it enough to have feature flags on or off in most cases? We’ve had so much success experimenting and personalization personalizing on app because, more users are are logged in. You got more data to to use.

You know more about the customer when they’re especially when they’re logged in on an app. And the other advantage of an app is that if it’s it it depends on how often, how frequently they go back to it, but it’s it’s a huge opportunity actually for personalizing on apps, as well as experimentation.

That’s a really good question. Or was it enough to have feature flags on or off in most case? I mean, when it comes to when it comes to experimentation, the whole process of server side testing for apps is is a little bit different to web because you need to think about when the app’s released. But there are some really good methods around. We do a lot of training to train teams on how to test and personalize on apps because you have to use variables a little bit more. There’s some clever ways, but I mean, you have to think about how an experiment is put together a bit more different a bit differently.

Okay. Second poll. This is so this is gonna be an interesting comparison, isn’t it, if we got eighty one percent AB testing to see how many people are personalizing yet.

And quite a few are planning next year, but we got thirty two percent yes at the moment.

But I think, twenty seven percent is also a very significant number.

Personalization takes a lot of time to actually execute, and, almost twenty seven percent of the people I assume that people are not selecting random options, but this is the true data that we are looking at. So twenty five to twenty seven percent of the respondents are actually planning to start it this year itself.

I think this is, I can safely assume this could be because of the approaching festive season that businesses are feeling the need to personalize their, the customer experience. Right?

Yeah. But here’s here’s a question for everybody.

For those who are not personalizing yet, why is that? What challenges are you having getting started?

Is is is it, for example, that you’re just not ready or is or are there technical difficulties or are there, like, cultural difficulties in getting going with personalizing or or some or actually, you just don’t know where to start, or something else. So it’d be great to find out, what people think.

But, yeah, there are people who are, sending in their inputs using the chat option itself. I see. Where to start. Right?

So Where to start.

Yeah. Technical competence, technical hurdles on the client side. Okay. On the client side. It must be someone working in an agency, I think.

There’s a lot of technical kind of Technical difficulties. Yeah. Aren’t there.

Jackie has also pointed out, active regulatory difficulties, which is also an important one.

Mhmm. Yeah. Technical and regulatory difficulties.

Yeah. That’s an interesting one, isn’t it? I suppose, if I’m just thinking some of our insurance clients and things, you need to be a bit more careful. I mean, you need to record everything that is live on the website because when somebody calls up, for example, on their call center, they need to know what they were looking at, what price they were seeing, what all the details were, and it’s, of course, it’s very highly regulated as it should be. So there are definitely difficulties. You couldn’t just go rogue and and do all kinds of, personalized messages on a site like that.

Right.

Okay.

Yeah. I do not see anyone who has raised their hand to to join on stage. I think everyone’s a bit camera shy today. Okay. Yeah.

I’ll let you Let’s continue then, and I would talk to you about the exercise.

So I’m going to I’m gonna start with the steps that, we go through with our clients. And today, I’m going to use a, not a client of ours, but just use a hotel site because they’re similar to many other hotel sites. And we do have we we work with travel and hotel and airlines, etcetera.

But I think so this is a really good one because I think it can you can get your head around it quite well.

And step step one is always it’s about identifying the key audiences.

And it remember, it’s which are most likely to have contrasting needs and behaviors.

And this is actually we we normally do it through a lot of stakeholder interviews, talking to people in the business, as well as a lot of, analysis and looking at existing research of of a client has. But a lot of clients a lot of people will know this already. So for example, and and this is, like, key audiences of interest. So in these this hotel site, I put three pairs of interesting audiences, and I’ll talk about them a bit later. So there’s guests without children and families with children. Actually, these are very different behaviors. So they’re of of interest to the company, and they’re of interest because they’re they have very contrasting needs and behaviors.

The other one are existing members. Membership and and, reward points are massive in the hotel industry.

Really, really important for people.

So and and there’s a very big difference when it comes to behavior and the user needs and the motivations for someone who is an existing member who’s collecting points and everyone else who are nonmembers.

And we actually see in the data, we see in the user research, we carry out some quite big differences in behavior there.

And then the third one, for the example today, we’re all gonna get our heads around, is ought to do with this kind of session continuation, really. So it’s new visitors in their first session, and then it’s return visits, visitors who have already been searching but not yet booked. And these two people these two groups of audiences also have very different user needs in our experience working with other clients.

So that’s the first step is to and and, of course, if you’re doing this for real, you need to there’ll be quite a few audience groups. The important thing is they need to be meaningful sizes of audiences.

And and and the clue here is that which are most likely to have contrasting needs. So this this is based on existing data. But as you do more research analysis, you’ll discover new interesting groups potentially, for example, in a financial sector, we see very different behavior from people who have come come from comparison websites to those that have gone direct to the brand.

Again, they’re they’re landing on different parts of the journey, but also their behavior is completely different. They’re much more, the ones going direct are loyal to the brands. They’ve already brought bought in. The ones coming from comparison websites are a lot more fickle.

So they’re they’re they’re more likely to to bounce. They’re more likely to go back and check another brand. So, therefore, you they need a different experience ideally.

So okay. So that’s the first step. The second step is to conduct opportunity discovery research analysis to identify exactly the contrasting needs and behaviors at these different steps of the journey. So this is where, we do our user research.

This is where we do our data analysis. This is where tools such as VWO insights and other heat mapping tools, for example, come in extremely useful because you want to look at the differences in behaviors and really drill into what’s important to these different groups. And then what I’m going to do now is to zoom in, and we’ll get another chance to do this. I’m gonna zoom in on a few on the guests without children.

And these are the type of things that we find out. And and, again, this is contrasting behaviors with families of children versus without children.

So, we with the families, we find all kinds of things. And then we got this a lot of this from survey data as well, from customer survey data. So we find out that families struggle to find the ideal room for them and their children. That seems to be a real need of them to find out what room is gonna suit them if they got children. And then there’s, what was coming up quite a lot was the thing around connected rooms in hotels, and that seemed to be quite a pain point and a difficulty for them.

The other aspect is a large number were interested in hotels with good facilities for their children, so, obviously, to keep the children occupied during the stay. The buffet breakfast options were coming out really highly of important with this group, but not with the other group. So this was because kids are often fussy eaters, I think, with the with this, brand.

And behavioral analytics was supporting a lot of these findings. We’re just looking at the amount of interaction with filters. You can gather a lot of information from what people are interacting with in filters or in in different parts of the page, which feed into the and when there’s a big difference between them and guests without children, there’s your opportunity for personalization.

So guests without children, in usability testing, many were put off if they saw too many pictures of children, which is quite a problem, isn’t it? Because you want beautiful pictures on your hotel website, and you want people maybe you want pictures of, like, yeah.

Fussy kids are under an understatement. Absolutely, Alice.

If, for example, you want to attract one type of user to your website, what pictures do you put up? You put a picture of a of of a family enjoying themselves and the kids playing, which could be fantastic for some people, but, of course, they’re gonna have the opposite effect for other people, like the business user, for example, who just wants peace and quiet.

So if we move to the guest without children, then, they want peace and quiet. A large number of these users were business travelers who need a desk to work from, so they’d be looking at the room information to find out if there’s a desk in there, if there’s a place to work. And things like reliable Wi Fi will be really important for this. These type of users and other facilities like business suites were coming up a lot in the research.

So it’s this. You can start to see, hopefully, that how the behaviors, how their user needs are very different, and that is where personalization comes in. It’s your ability to treat to improve the user experience for both of those pretty big segments, actually.

Okay. So that’s and we’ve we’ve done the same for, existing members.

So this is I always love this. The, with hotels and with airlines, well, the points are so valuable.

So, people are very, very driven by reward schemes and the points in the hotels and airlines, and they would do a lot to get more points.

But that is those existing members who collect the points and who are ready in the group. So we see really contrasting behaviors between existing members and nonmembers.

So for example so the problem statement here is existing members of reward scheme are extremely driven by points and the benefits that this gives them. However, many nonmembers don’t see any reason to become a member or or are worried about the hassle as well.

The business has highlighted a huge opportunity to help encourage regular bookers who are not yet members.

So there’s some really big difference. Like, when when surveyed, members said that the biggest benefit was the reward nights, which they can spend on their family. The comments was twice as common amongst business travelers, interestingly.

Retention rate is three times as high for members compared to nonmembers, which is kind of obvious, but it’s good to have in the data.

And third thirty percent of members spend more through upgrades in order to get more points, which which the analytics shows. So, actually, what I mean here is that they will spend more money to get more points. It’s such a driver, actually, and we’ve seen that in the data. Whilst nonmembers don’t see the point in becoming a member often, so they haven’t they haven’t been persuaded yet.

They don’t see the benefits. And there are several commenting that it’s too much hassle to sign up, and they’re worried about being pestered by emails and messages if they do sign up. And then messages if they do sign up. And then twenty percent of all traffic is from people who regularly book hotels, but who are still not signed up to be a member.

So you can start to see the opportunity here.

Okay. So I’m building up these these, like, problem statements here because, you know, you need a problem statement for personalization, and that’s why we do it this way. Okay. To the third and last one before we start brainstorming, before we start thinking about what the hypothesis should be, new visitors in their first session versus return visitors who have already been searched but not yet booked.

Okay. So the research and data analysis has shown a clear need for shortening the process, especially on return visits and especially on mobile.

So returning, most users do not complete a booking in their first visit, especially on mobile. In the survey responses, a large number of comments from people frustrated that they would lose all of their previous search criteria.

So people kind of quite annoyed that when they’ve you know, when you for steps you have to go through sometimes, putting all dates, put in the hotel, the destination, research, and then perhaps you’re showing it with your family members. And, of course, then you need to start again the next day, for example, is that’s a really frustrating experience and time consuming.

The user interviews indicated that many decisions were made with the partner, which meant that searching again for the hotels when they got home later on on in the day. And in usability testing, there were several comments about the process feeling quite long, actually, which is a common thing.

And then and only three percent of users scroll down to the home page to see other content. This is interesting. So returning, hardly anybody was scrolling down.

But for new visitors, it was, more people were actually scrolling down and looking for offers on the page.

So some very different behaviors there. So so these are our our use cases for today. So we got our, our problem statements. And, again, the problem statement is all about the differences in the these user needs.

And then if that’s all if anyone has any questions, what we’re gonna do is I’m gonna talk about the actual journey and the task.

And if you like, this is why not go to Marriott and actually and have a little play around and then kind of familiarize yourself with that journey and have in mind these kind of different scenarios.

What I’ve done on the Miro board here is I’ve put screen grabs of each step on mobile so you can see the home page with the, where to be searching and the content below. You can see the next step of searching, the destination, the dates, etcetera, and putting in your details. And then you’ve got your results page, which is where you list all the all the different hotels are there. There’s also a map view. I didn’t put a screenshot, but you can flip to the map view.

I’ll flip back.

Then what you do, if you want to, you don’t have to, but the user if you click view hotel details, they’re taken to this overlay, and it has more details about the individual hotel.

And then, so this is the scrolling down version. And then if you continue, you’re taken to the whole select a room page. And you’ve got different tabs with different options, and you’ve got tons of different types of rooms. Just thinking back to the to the families with kids here, you know, the importance of the room information.

Oh, yes. Samira is shared, actually. Maybe we share the link again.

People, if you want to put the link in. There you go.

Because that’s gonna be really important because we’re gonna do the we loads of post it notes on here. So okay. So select a room, then you get this membership overlay which appears if you’re not a member, and then join now. And then you’ve got, and also if you are a member but not signed in, you can sign in now. And then the last thing is the is the guest info and the payment. So this is the journey at the moment, but I would advise if you can, if you’ve got your mobile and got Internet, that you go to the Marriott, and have a little play around actually to get familiar familiarize yourself with the journey.

And and it’s a very it’s it’s a very, kind of common, kind of journey in the hotel industry, and it’s quite complicated too. That there’s a lot of things to think about for the user and and a lot of, you know, that there’s a lot of possible improvements for different types of users, and that’s the key thing. That’s why it’s such a big opportunity for personalization.

Okay.

Take a breath. Right.

So what I would like to to do right. I’m going to I’ve got these underneath the journey.

You’ll see there are three frames, three areas, and it corresponds to the journey and, the journey above.

And on each of these, I have put the I have repeated again the, the the audience groups, and I repeated the problem statement I’ve gone through. And then what I want everybody to do, and I can keep on filling up the post it notes, is to write your hypothesis, your ideas, at the appropriate place in this journey on this board. So for example, get write it here and say and and write what what you think would you’d want to do.

And then I want you to continue, and and I can make the board bigger if we want to, but I’m gonna be, I’m gonna let you do this for for the next twenty minutes or so. I think twenty minutes should be fine, I think, actually. So maybe if we have the counter for twenty minutes.

And and then what we’re gonna do, we will revisit and walk through all of this process and then ask a few questions.

And then hope, but please shout if you have any questions on the chat as we go through if you’re not sure what to do. And then enjoy and have fun, but keep on keep on thinking about the problem, which is the contrasting needs and behavior between these groups of users. Okay?

Excellent. And please, also, start to fill out the hypothesis on the second one. So feel free, and even if you’re struggling now running out of ideas on the on the first one, start adding some ideas on the second one, which is around existing members who collect points versus nonmembers.

Okay. You’re doing a great job.

And then for some of you, do you want to start, also working on the third one as well? I can see Julia was working on that. Nice work.

So this is all around the new visitors in their first session versus those returning who have already been searching but not yet booked.

K. I think we’re we’re I think people are are getting there with with all these, with all their ideas. I can’t see too many new ones.

So I’m just I think it might be time to kind of go and visit some of the ideas that they put on there.

Yes. Absolutely.

Okay.

There we go.

Really good work, everybody.

Some fantastic ideas. I was having a sneak peek as well.

So amazing job. I hope you enjoyed the exercise too because I think it’s it’s a it it works for us and works for our clients working this way, actually. So it’s a good it’s a good exercise, keeps us all grounded in real data and real, solving real customer problems, and away from very much, like, solution first approaches.

And it’s it’s a great way to create more user centered experiences for your customers through personalization.

But I thought what we would do is just pick up on a few a few of these ideas as we go through.

So the the you guys are really busy. So this one, if this is without children, we’re seeing images without children, they’re more likely to book and convert.

And I think with this one, for example, this this one around the imagery, you can sometimes just prioritize and move them around, on the carousel, for example. So you can change the default first order. You don’t always have to remove imagery completely, but, hey, you can test that too. So I think that’s a a great one.

I think there’s one here. By pre selecting amenities filters based on the the, number and type of travelers, we can reduce the time to book by displaying only those results with tailored amenities, images of families, and personalized sort. This is really important, actually.

Because I think if you look at the, filters on mobile, you you only get to see two or three in view. So if you can start to, display filters and even preselect them, I think this is a great idea based on what you know about them.

After all, they’ve they put in that they they have children in the search, so you know that about them. So, therefore, you can start to prioritize. That’s a really good idea there.

So a lot around the filtering, which was great. By featuring USBs and featuring USBs such as gym facilities and fast reliable Wi Fi in the search results for users who have selected that they are booking for only one or two guests, we will observe increasing click through rates. Yeah. Absolutely.

That’s a great one. It’s really solid information what people actually, some clients have put in extra questions in their search box purely because it allows them the ability to tailor the experience a bit better later. So you may have noticed that some clients some, some websites actually ask whether people are business or leisure, for example, and that’s so that they can improve their data.

So we got some fantastic ideas there. We will review everything afterwards, and let’s look at the existing members and nonmembers. This is always a fascinating one.

So, yeah, someone’s around offering a tangible welcome bonus for signing up that would be applied to the first trip.

So that trying to get people to log in if we offer some perk related to logging in, e g free cancellation, we’ll boost sign up rates. Absolutely.

There’s some extra data we’d need to figure out and drill into in terms of the value between one off traveler versus frequent traveler.

Cool.

Single click sign up for for the regular booked nonmembers.

Absolutely. Because the nonmembers at the moment, but regularly booked is a real key kind of focus.

So incentivize regular nonmembers with a value complimentary gift when they sign up. That’s the business objective here. Absolutely.

But, again, I I hope you can you can start to see that if you approach these strategies for everybody, then it would really upset some people. It just wouldn’t be effective, and that’s the whole point of personalisation.

But if you sometimes, teams make the mistake of they get an idea and they for personalization, and you only targeted at one user type or, actually, that improvement could benefit everybody. So not everything needs to be targeted.

Okay. Let me go down. Okay. So tons of ideas. Let me go to the last one, which is new visitors and returning visitors who have already been searching.

So let’s have a look at that.

If we notice visitors returning very close to their entry date, we could display the same filters chosen plus a better price to help them navigate faster and also create urgency.

Two in one hypothesis.

Nice work.

By using cookies, show their last step, last destination plus date, selection on the home page to speed up booking journey. Absolutely. Yeah. I love that idea.

And you can take that you can take that even further. That whole kind of, just shortcut. You know, just put up those three or four shortcut links to their last searches, and that will save them tons of time. They don’t have to put in all the dates again, which is a which is a pain, isn’t it?

Really good ways of improving the user experience through personalization.

Right? Add latest search feature for returning, and the guests recommend and return to us because yeah.

Amazing ideas.

Oh, if we allow returning visitors to quickly save their search criteria, I. E. Entering just their email address or simple checkbox to save a session cookie, we will increase second visit bookings. Yeah. So this is interesting. This is almost like, explicit when you’re asking them to save.

You you don’t have to do this, but this is another option. Many sites, you can just save by default and assume it’s gonna be good for them.

But okay. I think we will start to wrap up now, but I think it’s an amazing idea. And, of course, these mirror boards, I’m gonna leave open for you guys as well.

I just wanted to touch on the last couple of steps. If we were doing this for real, the next step would be to prioritize the hypothesis based on the size of the problem they’re attempting to solve. This is also based on the size of the segment as well. So it it’s really, as with all experimentation, we really believe in, problem first prioritization, which means that you should focus on the problem you’re trying to solve and how big that is, not how not subjective prioritization of how good you think the idea is, but that’s probably for another day.

So it’s next step is prioritization, and then it goes into the feasibility check, design, and build the personalization experiments.

Key thing is here, they need to be experiments. You need to control. You need to AB test them. Don’t make the mistake of just assuming that they’re gonna work because it’s personalized.

And then, eventually, you need a really good balance. You need a you need a a a team or, this needs to be in your normal experimentation or product roadmap. You should always have a have a level of personalization because the rewards are absolutely there.

So I would like to, kind of finish up with with the last few questions, really, just just to see, I’d love to find out more around your challenges in getting going with personalization. So if anybody hasn’t added anything yet into the chat, it’ll be great to to figure out if there’s anything stopping you. And, hopefully, today’s session focused on insight led personalization has been helpful and is something that you can take away, with with you. And and, of course, do get in touch. I’m always on LinkedIn. So send me a message if you want to chat further, and we could talk about this some more. We can go into a bit more detail around the practicalities.

But Yes.

Absolutely. Feel free to connect with Chris. He is very active on LinkedIn.

And, send send him your questions, and he’ll be able to answer.

One question just rolled in, Chris. Yeah. I know it’s from, Mary.

Let me just, wait. Did I yep.

Examples of personalization for ecommerce.

Absolutely. There’s so much you can do for ecommerce.

The the the session continuation one is something I really recommend exploring because you can for ecommerce, you can on on their return journey, you can surface what they were searching last time. You can surface if there’s something in their basket.

So there’s many and also for brand new users, you can surface the more brand information, introductory information, and offers. So there’s some really key starting points for ecommerce.

But, yeah, happy to follow-up on that one as well. But, yeah, ecommerce has really lends itself well to personalization.

Yeah. Just just a quick reminder, guys. You can also, come up on stage and ask your question with more context and have your follow-up questions answered as well.

Carol has, an interesting question. Carol, would you like to come up on stage and chat? You can send in your request using the join tab.

But, yeah, here’s the question from Carol. She’s asking, they’re asking, brother, if there’s time for one more question, could you comment on using personalization tools versus building it in house?

Yeah. Exactly. So this is a big this is a big one actually for personalization because you tools give you the flexibility to test out personalization in the in in the easiest way. They give you the agility.

They give you the ability and to to to test things out properly. But eventually, it is in the ideal world, you want to build it impermanently into your platform, the actual so when you’ve discovered the winning solution, you have two choices. You either keep it going in in the in the experimentation, in the testing personalization platform ongoing, or you build it into the ecommerce platform. And that’s the more that’s the better long term solution.

So I’m a believer in using experimentation tools like VWO for exploring the opportunity, exploring the personalization campaigns, and when you find a winner, if possible, building it into your platform.

Right. But, I’ll be interested I mean, we all will be interested in knowing your thoughts as well, Carol.

I saw that your request just came in. I have approved your request and hey. Hi, Carol. How are you?

Hi, Carol. How are you doing?

Hi, everyone. Thanks for answering, Chris. Yeah. I mean, so I’ve been, working on personalization for a bit more than a year now. And as I was, like, getting up to speed with everything, I saw a lot of research comes from the tools that is focused on personalization. But in house, we’re, mostly doing, like, recommendations, and we’re building, like, kind of our, setup to deliver recommendations to API and and stuff like that. And then, we we have, like, a testing platform where we do some, we we do use more for experimentation, but I was, I was curious, the the personalization engines and tools available online, available, internally.

For especially for bigger companies, I I think that the the implementation of such a tool can be, bit problematic and and costly, and, and a lot of companies decide to because they have also a researcher to do it in house. So, I was I was curious if it’s worth for bigger companies to explore also the personalization tools or if it’s enough, like, having, AB testing tool, and on top of that, the the in house solutions?

I think for the long term, in house for permanently serving is is is the best approach. But for for a lot of people, it’s quite tough to build their own AB testing tool. So so and have it in but for it’s it’s only some of the biggest companies go that way.

But I think a good a good approach is to do your exploration through an experimentation tool, find the personalized experiences which are working well, and then hopefully look to build that into your platform as a, as as a normal experience for people. So say, for example, the ideas around the get letting people jump to their last search really easily.

Maybe you you put a little shortcut underneath the search box that says, you know, the the last one, Los Angeles, you know, these dates, and it’s a quick way through.

You would eventually be better to build that into the actual CMS if you can target it with those users to build the code in and do it that way. Other otherwise, you’re reliant on the testing tool always to kind of continually serve that winning experience.

So, eventually, I think that combination works really well in my experience.

Okay. Thanks a lot, Chris.

Thank you.

Thank you so much, Carol, for your hey. Hi, Jane. How are you?

Are you on mute today?

We cannot hear you, Jane.

The mic is switched on, by the way, but no worries. I’ll just pin your message to the stage. Sorry, Jen. This is the question that Jen has asked. Would you prioritize a lower or higher percentage segment? I would think ROI would drive that decision, but curious what your thoughts are.

That’s a really good question. I think the, the the important thing here is I would prioritize, a fairly big segment, but you need to take into account what the differences in the needs are as well. So it’s the the upper size of the opportunity is almost equal to the size of the segments, but also the difference in their user needs. So if there’s a real big contrast, like, if somebody wants something completely different than someone else and they’re also big segments, that would be top priority list.

And there’s also feasibility as well. If the segments become so small, it, and it’s hard to hard to have such an impact.

So I hope that helps answer the question.

Sorry. I couldn’t hear you.

Yeah. Unfortunately, we couldn’t hear you, Jen.

But thanks for your question.

I will now see you at the stage, and let me see if there are any new request.

There are, no other questions.

Let’s see. There is a question again.

There is some long post by Emmy.

Let me just put that up on stage as well.

Prior to a second, they’re on a problem they are attempting to solve. I prefer to include also confidence how much we have data from not totally. How much we have data about problem and solution solving problem and ease. How difficult is it?

Just we’re trying to solve something huge, which might in the end take too much class to you or take too much issue. Okay. No. Good question.

There’s when we talk about problem first prioritization, there are two aspects to it. There’s prioritizing the size of the problem, and actually confidence does come into that. It’s it’s not about how much data. It’s how how much you understand the problem you’re solving.

So if you’re really if you have some really good quantum call, that means you can improve that confidence for the problem. But then what we prefer to do on the solution side, we prefer to think about the the problem solution fit.

And and we do it in that way because we don’t wanna give up on problems. And, also, some solutions might be simple, like, lesser for other solutions might be bigger. So it’s a very much doing both of things is really important.

This is an enormous topic in itself. This is a we’ve been I’ve been speaking quite a lot about problem first prioritization recently, which is a really good such an important topic.

There’s another interesting question by Alan. Alan, would you like to come up on stage?

Oh, I can read the question otherwise.

Speak right now. But, yeah, he’s not able to speak, but I’ll just put up his question on the stage. So Alan is asking, how are you dealing with the current problem of privacy and third party data to make a good targeting strategy for personalization?

Is there any tool for this?

Yes. So we we have integrated CDP sometimes with the the personalization and testing tools to help with targeting.

But this is this is a feasibility question. This is like when it comes to the step five feasibility checks. I the most important one is can we actually identify these people, which is the feasibility that happens early on because, obviously, that that’s that’s absolutely key. Some of the segments we’ve talked about today are easy to identify because it’s through the actions actually on the website. Like, for example, are we adding children or not? So it makes the segmentation quite easy.

But, of course, we still have to store it in the cookie, and cookies don’t last forever on different devices, etcetera.

But I think it’s when it comes to privacy, it depends what’s data you’re storing as well.

If it’s really personal data or not, we’re actually the behavioral data is anonymous. And I think that’s the key thing with the with the data you’re storing. If you’re if you’re obviously, you need to be careful if you’ve got all customer data. That’s the thing you have to be worried about and and just and just make sure you use the appropriate tools.

But that’s what CDPs are for.

Ah, okay. Got it.

So from web design planning standpoint, when a group is working to plan and You can try John.

Fine. Oh, yeah.

He’s here.

Is John there?

So he’s quite skeptical about his audio, working on that.

Okay. Oh, we we got his question here anyway, haven’t we?

Yeah. We got your question anyway.

That’s right. Teams. I’m in news and wants of marketers, but I’ll look into trying to assist.

Plan and develop a new website. How when should they start thinking about personalization?

It seems to my it seems to me there is a value between UX, UI teams and the needs and wants of marketers, okay, that are looking to drive business objectives.

That’s a great question.

There there is a the what we want to do with all apps and websites, we wanna do what’s good for the customers, but in line with business needs, don’t we? We want to do what’s good for both people. Actually, helping a customer through the booking process is good for the customer, and it’s also good for the business, of course.

And so that’s in in some ways, that’s a bit of, like I know marketers are looking to push things. That’s the problem. The key thing is we need to think about the end to end experience. And sometimes when you push too much content in front of people, it ruins the experience. So it’s just about these product teams and UX teams working alongside marketing. I think we’ve when should we start thinking about personalization?

I think you need to think about getting the UX right first if it’s a brand new journey. You need to do a lot of usability testing, making sure it’s accessible, usable, and then works really well.

But then personalization is directly related to to what you know about your customers and to the different people you need to serve. And it’s just it’s like it’s it’s the next level up. I look at it like a level on top of the core experience, to really, really make that extra impact.

But, yeah, I think it’s with with marketing teams who are challenged with driving the revenue all the time, I can completely see the problem. It’s just they need to be aligned with you can’t make you can’t really and you shouldn’t really make users do something or manipulate them. It shouldn’t work like that. You need to you need to you need to provide entice them, rather than rather than the other way. But it’s a great question.

Right. At the end, of course, it comes down to building a consensus between different teams and different team members. Right?

Different teams, different departments in the organization have different KRAs. They have different KPIs. Right? There are different kind of an ROI that they’re looking at. Right? Marketing team is looking at a different ROI.

UI UX team is looking at a different ROI. Right? The investment is different. The returns are different, sim proportionately.

So I think, the entire, you know, conflict sort of even that even if that exists, it it, sort of is resolved when all the team members come together and they figure out, you know, what are the areas what are the specific areas of their responsibility and, you know, who should contribute at what stages.

So I think that should help you out.

Right. But, I think that’s the end of all the questions. I do not see any more people who have raised their hands. So I think, we’ll end the workshop here, Chris. Is are there any other stuff that you?

No. No. I think just just keep in touch. I’m really interested to see how you get on if if you if you apply this to your own websites and apps, actually. So, yeah, do keep in touch. And if you have any problems, know where we are, so we’re happy to help. But thank you so much, everybody, for for, all the brain work that went into the ideation session and your attention.

I really loved the engagement window, on the chat and also on the my report.

People have really put in their thoughts and put in their efforts to, you know, work on all these exercises. So thank you so much, guys. It wouldn’t have been possible without you. And, of course, it wouldn’t have been possible without Chris as well because he’s the one who has built this entire workshop. So thank you so much, Chris, for putting in your effort to build this workshop and, you know, tuning in at this hour. I know you you have a really, really busy schedule, but you still made out the time to, to run this workshop. And, also, guys, do not miss the two other presentations slash discussions that Chris has, done with, Vicky from EasyJet and Annette from Primark.

So the recordings of those conversations, those presentations are now live on, the Convex two thousand and twenty four landing page. I’ll just request, my colleague to drop in the link to the landing page where you can scroll to the agenda section, and you can watch just all the recordings from day one, two, and day three as well.

And, of course, there is a quiz coming up. And by the way, Chris, thank you so much for your time.

Yeah.

I will now send you off the stage.

And feel free to hang around, to, you know, look at the quiz as well. Else, if you have another meeting set up, of course, feel free to to jump off as well.

Will do.

And don’t forget to do a survey as well.

There’s a survey in the last tab as well, so don’t forget to do that.

Yes. Yes. Forgot to, mention. There’s a survey in the last tab.

We we are looking forward to your feedback. Your, your feedback is what helps us keep improving, and we will be sharing your feedback with Chris as well. So just mentioning the feedback that this is particularly for, for me, VWO, or, for Chris. We will share the feedback accordingly, and it’ll help us to build better, you know, webinar experiences in the future.

So please do take out the time to, you know, fill up the survey as well. But, yeah, thank you so much, Chris, for your time. And No worries.

Bye, everyone.

Looking forward to having you in the next webinar maybe sometime shoot sooner. Awesome.