Latest / Inside Commerce: Ecommerce Strategy, CX and Technology Podcast / EP251: Predicting User Intent Is The Key To Ecommerce Personalisation - Made With Intent Helps You Unlock Its Secrets
Transcript
- 0:02You're listening to the Insidecommerce podcast, hosted
- 0:05by experienced consultants James Gerd and Paul Rogers.
- 0:08Insidecommerce is your fast track to digital knowledge,
- 0:10featuring interviews with leading e-commerce retailers,
- 0:13industry thoughtleys and the C-Suite of key tech companies to
- 0:16explore the latest in digital strategy, customer experience
- 0:18and tech. Discover what the latest trends
- 0:21really mean and get practical advice based on experience to
- 0:24help fine tune your ECOM strategy and tech stack.
- 0:27Follow us on LinkedIn and reach out to James and Paul if you'd
- 0:30like to discuss an e-commerce project.
- 0:32We hope you enjoy the latest episode.
- 0:35Today's episode is sponsored by Centro and Grebin.
- 0:37More about them later. Let's set the scene.
- 0:39Our topic is predicting user intent to improve onsite
- 0:42segmentation and conversion. So the principle by this is
- 0:44about, yeah, not want to treat all customers are the same, not
- 0:48expecting that all customers are coming to your side ready to
- 0:50buy. And if by understanding the
- 0:52different signals around user intent, we can be better at
- 0:54communicating and targeting those people with relevancy
- 0:57during their user journey, that should lead to better customer
- 1:00experience, better commercial outcomes.
- 1:02So we are talking to the founder of a tool that is focused in
- 1:07this area, specifically David Van Dheim.
- 1:10Welcome, David. How are you?
- 1:12Yeah, I'm all right. Thanks.
- 1:13I'm all right. I mean, we had a bit of a somber
- 1:15conversation about two minutes ago, but let's perk things up
- 1:17and let's say, yes, everything is all good.
- 1:22Well, this is going to be hopefully this will will sort of
- 1:24like elevate the the experience really positive conversation
- 1:27about what you're up to and the, and the things you're doing with
- 1:29some exciting retailers like Rafa.
- 1:31So just a bit of background before you tell us about made
- 1:35with intent. So you were previously global VP
- 1:38of CRO because you like you obviously like letters in your
- 1:41job titles at brain labs. You're currently a digital board
- 1:44advisor of some some impressive brands like current body who
- 1:47Paul knows well because he worked very closely with.
- 1:50And you're the author of the person and personalisation.
- 1:52So you're like, it's like having a mini celebrity on I.
- 1:57Wouldn't go that far. In fact, if you do, if you do
- 1:59want to add to the semi or less so impressive, see the I was a
- 2:05celebrity at one point, I would say AZ list celebrity.
- 2:08I was on the famous show Take Me Out.
- 2:11How embarrassing is that? There you go.
- 2:13So talk about. That is a first on the inside
- 2:17commerce pocket. I feel like we need a gong or
- 2:18something going off. But so let's let's start with
- 2:22your journey because before we get into the details around made
- 2:24with intent and there's going to be some really interesting stuff
- 2:26listening. So make sure you keep this is
- 2:29take us through your journey because you're exiting acro
- 2:31agency, writing a book and then launching made with intent like
- 2:34what, why, why that path? Well, it all felt kind of
- 2:40sequential, right? So the, so the agency that I
- 2:43owned use a conversion. I founded that in 2015.
- 2:47Never really wanted it to be honest.
- 2:49It was all rather accidental. But there we are.
- 2:51We are where we are. And I felt like throughout that
- 2:54process, I'm a practitioner, right?
- 2:57I really care about doing things in the right way, as most
- 3:00practitioners would say. So I felt as though we were just
- 3:04optimising for revenue. And I think when you optimise
- 3:07them purely on everything for revenue, you get the, the
- 3:10e-commerce stores that we see nowadays, right?
- 3:12You know, pop ups flying in front of our face saying, hey,
- 3:14why don't you do this for 10% off, Please buy in session.
- 3:17Please buy in session. That way we can attribute to,
- 3:19hey, Google, here's loads of money because we know we can
- 3:21attribute that because acquisition's easing the
- 3:23retention, yadda, yadda, yadda. And sorry, I felt a little bit
- 3:27of needed a bit of salt for the chip on my shoulder, put it that
- 3:31way. Anyway, when I sold that, I
- 3:34decided to explore those feelings a little bit more, both
- 3:37in therapy and in 80,000 words. So I interviewed, it's about 150
- 3:42odd. I think it was 150 odd.
- 3:44Practitioners from around the world mostly centred around this
- 3:48concept of personalisation, because I feel like
- 3:51personalisation should all be about person being personal with
- 3:54one another. And I feel as though we've lost
- 3:57that connection between e-commerce brands and the person
- 4:00on the other side of the screen. So I don't know we all and that
- 4:04and we could talk about football if you want, But that coupled
- 4:06with my love for football kind of created this weird tiki guy
- 4:11moment. You know, this what you need,
- 4:12what you have, what you love, what you can sell.
- 4:15And I had to explore it. I had to create a hypothesis and
- 4:19see if this thing would work. And here we are today.
- 4:22It works. Greben create digital flagship
- 4:26stores for iconic brands of today and tomorrow.
- 4:29We are the number one Sentra agency in the world with more
- 4:32than 50 Sentra implementations to date.
- 4:35We are also e-commerce design specialists, platform and tech
- 4:38partner agnostic. Try our new e-commerce design
- 4:41subscription. Visit greben.com.
- 4:44Right. Yeah, I think, yeah, that's an
- 4:46interesting topic. We could do a podcast on some of
- 4:48that stuff in itself, to be honest, like the whole kind of
- 4:52like what I'm a funnel experience that everyone seems
- 4:56to be moving towards at the moment.
- 4:58Yeah, not really like building for new customer list, but yeah,
- 5:01I think we'll talk a bit more about that later on.
- 5:04So I guess first question for me, like how and why did you
- 5:07land on predicting or like the the business like how did you
- 5:12kind of because I think the thing that I found quite
- 5:13interesting was the role that you're playing and you're not
- 5:16trying to do everything. You're kind of really kind of
- 5:19specialising or honing in on like the underlying data.
- 5:23So yeah, that what was, what was the initial kind of idea?
- 5:26Yeah. Where did it come from?
- 5:27I guess that's the football route being honest, so we might
- 5:30as well jump, jump right into it.
- 5:32As a United fan speaking to a Liverpool fan, the the whole
- 5:36idea of expected goals is really one that I, I was fascinated.
- 5:40You remember when football went away, the whole European Super
- 5:43League, this little thing called COVID meant that the Bundesliga
- 5:45was the only thing on in the world.
- 5:48I started reading up on books and specifically the XG
- 5:51philosophy. So the idea of XG expected goals
- 5:55was it's not about the quantity of shots that you have against a
- 5:59goal, a net. It's about the quality of that
- 6:01chance. And by flipping that
- 6:04perspective, it meant that tactics and strategies could be
- 6:08much more focused on, you know, thank you, pet, but tick attack
- 6:11a football or the type of chance that one would have less so the
- 6:17whether you're pinging the ball from 30 yards or two foot away.
- 6:20And so the concept of XG was born.
- 6:24Took about five years to get a match of the day.
- 6:26And it really has transformed the way that football matches
- 6:29are played. You know, some would argue for
- 6:31the worse, OK, But there are more goals now than ever before.
- 6:36The football quality is the highest that it's ever been
- 6:39arguably. And the idea of predictive
- 6:41analytics is in every permeates to every sport.
- 6:45Starting in baseball. We've all seen Moneyball,
- 6:47grateful, went through to tennis, went through to rugby.
- 6:51It's in golf, it's in curling, it's in everything you do.
- 6:55And I just feel, Paul, that the the problems that we're trying,
- 7:00that we're trying to be solved with in football, the quantity
- 7:02of chances feels like the same problems that exist in
- 7:06e-commerce. The idea of we're going to shove
- 7:09everything at the top of the hopper and see what falls out.
- 7:12So that's that's kind of where the idea originated from.
- 7:16I'm going to thank football for that, all my love for it.
- 7:19And then again my mixed with the ship on my shoulder helped to
- 7:23create this hypothesis. And I guess what was your and I,
- 7:28I should just check that we're not covering this later on.
- 7:31I don't think we are. But like what was your thinking
- 7:33behind and this might just be the stage you're out.
- 7:35It might be because of kind of how you'll go to market like
- 7:38people like Bloom Reach and mentioned dynamic yield earlier.
- 7:42But what was the thinking that I guessed behind just focusing in
- 7:45on that data rather than, you know, the whole kind of front
- 7:48end sort of being presentation element?
- 7:51Is it that you just think that there's room for someone to
- 7:54really like specialized in that area and I guess, yeah, build
- 7:58out more data points and better understand segments?
- 8:00Yeah. What was the reason?
- 8:02Well-being honest it was is an iterative reasoning.
- 8:05So our go to market was exactly that.
- 8:08We thought we'd be a front end layer as well and maybe a
- 8:11personalisation platform you could cause at the end of the
- 8:13day, I literally wrote the book on it, right.
- 8:15So it feels it feels quite natural to translate to there.
- 8:19What we found though is that people started pitching holers
- 8:22in this what was effectively a pop up merchant, you know,
- 8:27nudge, nudge, nudge, pop up, pop up, pop up.
- 8:28And that's almost the antithesis of what we're doing.
- 8:31We also found, Paul, that one of our biggest, biggest personas
- 8:36was, was CRM teams. It was all about customer
- 8:39lifetime value. It's not just about what happens
- 8:41on site, but if you understand the intent after they've left,
- 8:44you could be more appropriate to them.
- 8:46So by not having a front end layer and instead by integrating
- 8:50into those platforms, our ICP widened, our personas widened
- 8:55and we became more of an authority then dare I say a pop
- 9:00up merchant. Am I going to get told off for
- 9:01that? I'm not am I?
- 9:02Nah. Should be fine.
- 9:07Centra is the only commerce platform built exclusively for
- 9:09fashion brands. Over 300 fashion brands run on
- 9:11Centra, including icons like Paul Smith, Jeff Ave. and Nudi
- 9:14Jeans. Centra solves business logic
- 9:16challenges for fashion brands like no other platform, and they
- 9:18are on a mission to help you scale up rapidly globally across
- 9:21all online and offline channels. Take your fashion brand global
- 9:24faster, further stronger today. I I think let's add a bit of
- 9:30context and quantification, David, because when, when, when
- 9:32we talked prior to to this recording, we're talking about
- 9:36the like the ubiquity of these. You're right.
- 9:38You come on the site, sign up the Newsday, get 10% off and to
- 9:42varying degrees of of like hassling and like not letting
- 9:47you go until you've closed the pop up several times.
- 9:50And a lot of retailers become dependent on them because it
- 9:52they see the data and the data says they get people and people
- 9:55spend and it's like tick success.
- 9:57But you talked about the, that less is more.
- 10:01So can you from your point of view, from what you're seeing
- 10:04and you know, from your, the strategic perspective and the
- 10:07data analysts, what are the issues with generic like 10% pop
- 10:10ups for newsletter? And how has your intent to help
- 10:13prove that sometimes less is more?
- 10:17Well those types of tactics are designed for everyone but only
- 10:21impact the few you know full well.
- 10:24If you land on a website that has a hey 10% off, sign up for a
- 10:27newsletter immediately. As you land all you see is that
- 10:31big fat X or in some cases the small X.
- 10:34Or the worst culprits are the really passive aggressive ones.
- 10:38Have you seen this? You know where the call to
- 10:39action is? No, I don't need money off
- 10:41today. Thank you.
- 10:42I'm happy paying full price. Big fuck off to them.
- 10:46I'm allowed. I'm allowed to swear, yeah.
- 10:47Yeah, I thought they honestly deserve, they deserve that for
- 10:50that. Just that.
- 10:52Well, we think it's clever copy. Is the most passive aggressive
- 10:58thing you've ever seen on any website ever a pain should?
- 11:00Release it as hell. You know what?
- 11:04Ironically, it does work psychologically anyways, So
- 11:08we've we tested a few functions. I'll give you 3 examples.
- 11:12I'll give you the newsletter example, Social proof, probably
- 11:15one of my favourites and discounted.
- 11:17So in terms of newsletters, the hypothesis here is that as soon
- 11:22as someone lands on the site and you get given a new user and you
- 11:25get given a pop up for 10% off is less really more?
- 11:30We've tested this across about between 6:00 and 10:00 retailers
- 11:33and we found that yes, although you're serving a pop up to
- 11:37somebody and it therefore has less impressions, the number of
- 11:41subscriptions that occur is inherently more.
- 11:44One example of that is where we instead of serving it as some,
- 11:48as soon as somebody lands serving it when they're likely
- 11:53to delete their intent to exit. Intent isn't just about
- 11:56purchased by the way, their intent to exit increases.
- 11:59So before the action actually happens, we saw that that very
- 12:03same pop up is served 66% less, but the number of subscriptions
- 12:07increases by 29%. That's only just one of those
- 12:10retailers. Less is more, quality does the
- 12:13quantity. I found that really interesting.
- 12:16My second favorite example will go for discounting straight
- 12:19away. The whole concept of giving
- 12:21discounts to those who are going to buy anyways feels a little
- 12:25bit parochial, feels a bit redundant, doesn't it?
- 12:28So if you're able to identify users who are ready to buy, who
- 12:32have a high intent, then we can almost exclude those from any
- 12:36discounts that are given. And yes, your conversion rate
- 12:39might ultimately decrease because some of those users of
- 12:42that segment will not end up not purchasing.
- 12:46But guess what, you're saving your margin because the vast
- 12:49majority of them will. So we, we were with an
- 12:51electronics retailer and we managed to save a, a very
- 12:55specific discount for them. We managed to save their margin
- 12:57by about 42%, 42% decrease because you're identifying users
- 13:01who are ready to buy. The adverse happens by the way.
- 13:04So instead of just looking at users who might be ready to buy
- 13:07and excluding something from them, think about the adverse of
- 13:10users who are never going to buy.
- 13:12So why would you remarket to them?
- 13:15One of my this isn't a client of ours, but one of my favourite
- 13:17stories is that of Land Rover, where this is years ago, by the
- 13:21way. So forgive me if Land Rover are
- 13:23listening, I'm sure they've changed this now, but to get
- 13:26through to a test drive you have to go through the car
- 13:28Configurator and they were remarketing to use all users go
- 13:33through the car Configurator who hadn't completed a test drive.
- 13:36Now guess what? Can you guess the average age of
- 13:39the audience that were going through the car Configurator?
- 13:43It's 14 year old boys pissing about during their, their, you
- 13:46know, dream car they have no intent to purchase whatsoever.
- 13:49They are literally tyre kickers. You could call them tyre
- 13:52kickers. So yeah.
- 13:54So think of those who are ready to buy and those who aren't
- 13:56ready to buy. I could give you another example
- 13:58of social proof, but you can start to see how less is indeed
- 14:02more. That makes sense.
- 14:06So I guess moving on to the kind of how you're segmenting
- 14:11customers, so you mentioned earlier that you have around 250
- 14:15different signals for predicting intent.
- 14:18How are you bucketing or like grouping those intent against
- 14:23customers? Like what are the different
- 14:24stages that you're looking at? So we, we just call it context.
- 14:29At the end of the day, all we're doing is we're collecting
- 14:31context from users that are more appropriate to how people
- 14:36actually buy. So the very first thing we do is
- 14:39that we bucket them into an objective because guess what?
- 14:43Your website isn't all about conversion.
- 14:45So we have engage, build intent, maintain intent, and then
- 14:50convert. Those are generally the four
- 14:52objectives of what any website should be achieving at any point
- 14:55in time. We have segments per objective.
- 14:59So there are different segments that we would book at users into
- 15:01when you need to engage them versus different segments of
- 15:05users when you need to build intent, you know, persuade,
- 15:09remove anxieties, that type of thing.
- 15:12And within each of those we have different states, things like
- 15:14focus or abandoned or struggle. And those segments are generally
- 15:18named things like focus, shoppers, last chances, the, the
- 15:23types of people that have a really high intent to purchase,
- 15:26a high intent of leaving and a low intent to return feels like
- 15:30your last chance to grab them, right?
- 15:32So we take all those signals, we model them together, we schmooze
- 15:35them together in a, in a lovely LLM and we basically bucket them
- 15:39into those different segments that are more appropriate to how
- 15:43they buy and therefore your objectives.
- 15:44Does that makes sense, right? Yeah, absolutely.
- 15:49And I have a sub question on top of this.
- 15:51I guess you've probably got like a bit of a framework.
- 15:53Well, you've definitely got a framework and but how custom is
- 15:57each implementation? So that I have an example of a
- 16:00client at the moment, they sell high end furniture and I'm
- 16:03trying to get them to test more kind of lead Gen. orientated
- 16:07approaches and like, you know, micro conversions and, you know,
- 16:11maybe taking people through different journeys rather than
- 16:13just trying to get the transact on like a 10,000 LB piece of
- 16:18furniture. And I'd imagine like the the
- 16:20kind of event structure would be quite different to maybe like
- 16:24the average fashion brand. Like how do you how would you
- 16:26approach that? Or like, is that is it a case
- 16:29that it's more standardised? Or is it?
- 16:30Is there a lot of like bespoke? Now it's completely standardised
- 16:34again. We tested these approach
- 16:36rigorously and ultimately we fell back to what happens in
- 16:39football. Whereas every game is the same.
- 16:42There's 90 minutes with 22 players on the pitch, two goals,
- 16:46usually an offence and a defence.
- 16:48The structure and the attributes of a football game are exactly
- 16:52the same, just in the same way that the structures and the
- 16:56attributes of an e-commerce site is exactly the same.
- 16:59OK. There's always a series of
- 17:00products that require adding to baskets that is slightly more
- 17:03complex. Yes, granted, that requires a
- 17:06purchase. So you need to find those
- 17:07products and filtering etcetera, etcetera, etcetera.
- 17:10So the structure and the attributes are the same, but the
- 17:14weighting is different. So granted, you know, you're
- 17:17right, the the need for filters for a furniture brand might be
- 17:22different than that of like an ASOS for example.
- 17:25So it's the prioritisation of those attributes that are
- 17:27different. We found by having a universal
- 17:30large language model that's trained on over 75 different
- 17:33retailers. So there's no degradation in
- 17:36quality versus a standardised LLM that predicts intent,
- 17:40certainly not an accuracy or prediction.
- 17:43And in fact is allows you to benchmark at the same time
- 17:47against other people who you know, how good are they at
- 17:51building content? How good are they engaging?
- 17:53How what's their intent to purchase on users who come from
- 17:58paid media campaigns, that type of stuff.
- 18:00So for us, the answer's no. A universal large language model
- 18:05is usually we find has no degradation and that's that
- 18:08those are the benefits. So yeah, that's that's where
- 18:13we've where we've taken things. Something I come up against a
- 18:17lot of the time, and I'm sure you'll be the same, is that a
- 18:20lot of e-commerce retailers have short term trading mentality.
- 18:24They've got targets pressures and that leads to a lot of this
- 18:28ubiquitous like 10% off, 20% off blanket discounting.
- 18:32And it can therefore sometimes be hard for them to change
- 18:37approaches. What's so you work with lots of
- 18:40different people. So I'd love to hear from this,
- 18:42what's the business case? Because how what's the
- 18:44proposition? Then how is user intent analysis
- 18:47going to help them demonstrate to you know, whether it's AC
- 18:52suite or a director or a head of they're actually just constantly
- 18:56slashing prices and doing discount and isn't always the
- 18:59optimal way to to to convert. Yeah, you bank on, I mean,
- 19:03that's the enemy, isn't it? Short termism, the enemy isn't
- 19:06really conversion rate. That's just a measure of success
- 19:09of the enemy. The it's the same in in the CRO
- 19:13agency that I found it, you know, that's, that's one of the
- 19:16problems that I felt is that we were always optimising for the
- 19:19conversion. I know that's what the name
- 19:21suggests. It's not what process is
- 19:23intended to be though, because you're meant to convert to
- 19:27support you. Sorry, you're meant to optimise,
- 19:29to support users through a journey, right?
- 19:31Not just, oh, let's shove some social proof here, there and
- 19:35everywhere sticky out to cars, fill your boots.
- 19:37That's generally what happens, isn't it?
- 19:39So, yeah, our, our mission, James, is to change the
- 19:42perspective of e-commerce for the better.
- 19:45OK, there's a few things in there.
- 19:47Number one, to change behaviour, you need to 1st change attitude.
- 19:51So this concept of change in perspective, it's all, it's all
- 19:54us #2 we call ourselves Moneyball for e-commerce.
- 19:58You know, self-proclaimed. That's what happened in
- 20:00Moneyball. Someone has to dance first for
- 20:03others to dance, right? And I feel as though we
- 20:07trademark the, the term XXC, see it's exactly the same as
- 20:11expected. Could goals.
- 20:12We trade well the term XC for expected conversion because we
- 20:16feel by having a more predictive real time almost two way
- 20:21measure. Of customer satisfaction and
- 20:24business objectives rather than something that's binary and
- 20:27retrospective and aggregated like conversion rate.
- 20:30We feel that's the start of the movement.
- 20:33Does it sound quite culty? It's not intended to be, but
- 20:36it's definitely intended to be done for the better, right?
- 20:39For the betterment of e-commerce in general.
- 20:45Great. And so you work with around 25
- 20:49customers. I've got written here that
- 20:50earlier you said 75, so it might be a bit higher and that you
- 20:54know includes brands like Rafa and Benson.
- 20:57Yeah. Can you talk us through maybe a
- 20:59couple of case studies around kind of how people are using the
- 21:02product because going through the site you know it's really
- 21:05interesting like interactive examples of like how you can
- 21:09build up the data. Could it be interesting to know
- 21:12how some of these brands are then going on to use or whether
- 21:15it is kind of, you know, the CRM teams using it or, you know,
- 21:18they're doing more on the front end or, you know, there's a team
- 21:22of analysts just trying to understand customers.
- 21:25So yeah, if you've got a couple of case studies.
- 21:27Yeah, of course. So we generally have three use
- 21:29cases, right? Number one is insights.
- 21:32It's fairly obvious, but you can identify and prioritise segments
- 21:37based on user intent to purchase, to exit or or return.
- 21:42So understand the intent of visitors is a really nice case
- 21:44to do sea salt, sea salt, Cornwall, Cotswold outdoors
- 21:50their users in conjunction with our content Square and Hot Jar
- 21:54integrations, for example. Because if I were them and they,
- 21:58they've said this themselves, if I were them, I don't care just
- 22:01about how users generically use APDP or a PLPI, care about how
- 22:06they differ between users who are browsing versus those users
- 22:10that are committing, OK. And so I think that really helps
- 22:14them prioritise what to do in their experience and
- 22:17experimentation programs. Use case number two is, is all
- 22:22about in session intent, fairly obvious, but you can start to
- 22:26target users in session who have a high intent to purchase versus
- 22:30those who have a low intent to purchase.
- 22:32And your objectives should be different, right?
- 22:36You can either target visitors who are likely to exit, like the
- 22:39the newsletter example that I gave you, or you can convert
- 22:43visitors who are ready to buy. You know, in that case you can
- 22:45just leave them the hell alone if you really wanted to, or you
- 22:48can prioritize social proof. Well, probably one of my
- 22:52favourites is in Rafa did this. You can start to build a in tent
- 22:58with users who are already engaged on different visits.
- 23:02So for example, if a user comes back to your site, it's not
- 23:05enough just giving them a recently viewed mechanism.
- 23:09You recently viewed these products.
- 23:11OK, why do if they have a low intent to purchase, it might be
- 23:16useful giving them more than one suggestion or talk about
- 23:20categories rather than if they have a high intent to purchase,
- 23:23talk about that single product more.
- 23:25OK, that makes sense, right? If they have a really high
- 23:28intent, it suggests that they're past some of the stages that you
- 23:32would do in product discovery. So talking about that one
- 23:35product really help them engage not just with that module, but
- 23:39overall for that segment of users and it was like A-77 or 8%
- 23:44increase for that segment in revenue.
- 23:47By the way, we don't really measure conversion.
- 23:49And finally, probably our most popular use cases are those post
- 23:53visit. So prioritising and engaging
- 23:57with prospects in your e-mail database or your remarketing
- 24:00activity. So we, we integrate really well
- 24:04with like Meta for remarketing stuff, you know, like the Land
- 24:07Rover example I gave before, or we integrate really well with
- 24:11say a blue reach or an Amatria. Because at the end of the day,
- 24:15the way you speak to customers with a low intent should be
- 24:17different than a high intent. So I can give you examples of, I
- 24:21gave you the discount example before I can give you a raffer.
- 24:25Exclude all low intent users from their remarketing activity.
- 24:29And only when they have a high higher level of intent do they
- 24:32start to speak to users in their remarketing on meta campaigns
- 24:37and social campaigns or indeed they change that message.
- 24:41They have a low intent. Talk about the brand values.
- 24:44If they have a higher intent, well, talk about the specific
- 24:46products or the logistics of the purchase.
- 24:49I think it's all about what we call appropriateness.
- 24:52So be more appropriate in how you communicate with your
- 24:55audience. And if you can segment that
- 24:57audience based on how they really buy, then you can start
- 25:00to create a lot, a lot more impact in those communication
- 25:03methodology. Don't assume everybody's ready
- 25:06to buy, which I think feels like a a normalized statement, but
- 25:09it's true. It is and and the the whole area
- 25:14of like custom communications around emails, SMS, whatever
- 25:16channel people are sending out their regular comms in.
- 25:21There are a lot of platforms that are really cool for flow
- 25:23base and have all of the tool in to, to do the, the advanced
- 25:27segmentation, but not necessarily the the data insight
- 25:30to create those meaningful segments to tailor the message
- 25:33in. So I think that's really
- 25:35interesting. Let's let's drill this a bit
- 25:36more because so you know, you don't, you don't do the
- 25:39personalisation. You are basically a data intent
- 25:42modeling engine that gives other tools the intelligence to be
- 25:46able to execute it. So be ready to know, you've
- 25:49mentioned a few vendors during the podcast, but what are the
- 25:53key vendors you currently integrate?
- 25:56And then what are the options? So if somebody really loves the
- 25:58idea of the tour but they've got a different third party, what
- 26:03are the integration options for for people you don't have a pre
- 26:05integration with? So we surface everything on the
- 26:09front end in the data layer. So if you're an integrator, a
- 26:13vendor, just grab it. Pretty simple stuff.
- 26:17But we have native integrations. I think we're about seven
- 26:20different vendors now. We're only babies, right?
- 26:23But Blue Reed, Cimetrio, Clavio, Meta tend to be the post visit
- 26:29stuff and then in the inside stuff it's hot jar con, hot jar,
- 26:34hot jar content Square. Same thing nowadays.
- 26:38GA, Adobe Analytics and then in Session.
- 26:42Actually don't think we have any native integrations of in
- 26:44session. You could count Blue Ridge as
- 26:46one of those I suppose. AB, tasty, dynamic yield,
- 26:50they're all pretty easy to integrate with regardless, we're
- 26:53just waiting for a use case. And I get just a sub question on
- 26:57that. I was just interested.
- 26:59And so at the moment, I'd assume a lot of your implementation is
- 27:02either like custom JavaScript or like by GTM or something like
- 27:06that. And do you have any intention to
- 27:08build out like platform specific integration and like move into
- 27:12like Shopclues ecosystem more for example?
- 27:15It's probably not within our core ICP right now, you know,
- 27:19start up just over a year old. We obviously prioritise.
- 27:24So currently, in fact just yesterday hots off the press, we
- 27:28have finally got our GTM native template all set up on the
- 27:32store. We didn't even get an e-mail.
- 27:34We've just been waiting months and months and months for
- 27:36confirmation. But we we we've given templates
- 27:39out for for a while now. So largely just GTM template
- 27:45based integration. It's pretty simple.
- 27:47We don't collect any PII, it's just tracking behaviour at the
- 27:50end of the day. It's the stuff in the back end
- 27:52that's the sexy stuff. You know, the LLN that models
- 27:55everything out into meaning, really into context.
- 28:00That makes sense. And then the the last question
- 28:03for me before James asks the last question, overalls and what
- 28:08size of the business need to be to have enough signals to really
- 28:12like get value from the product. And I can see you've got a bit
- 28:16of a mix of like bigger and smaller.
- 28:19This is on the website but is do you think it's a case of you
- 28:22know you need a certain level of traffic to really feed the model
- 28:26and get the value from it? It's less so the amount of
- 28:29traffic I would say and it's more so your maturity because
- 28:33everything at the end of the day, it's a universal model.
- 28:35So it's all propped up in some ways by everybody else's data,
- 28:39not just yours. And it's more so your maturity.
- 28:43So for example, our ICP is generally, we're not talking
- 28:47data scientists with far fetch, right, probably a bit too
- 28:50mature. We're also not talking
- 28:52shopifying on the pops who do a bit of side things on Etsy.
- 28:56We're talking about ideally someone would have ACDP or some
- 29:00kind of ESP that's a little bit more mature.
- 29:03And what I would say in the post visit stuff, we take somebody
- 29:06from immediate to to advance, intermediate to advance.
- 29:11And with the in session stuff, they're likely doing some AB
- 29:14testing, as you can imagine. But let's be candid here, the
- 29:17majority of people doing rather basic AB testing, you know,
- 29:21hands up if you've thrown a sticky CTA on your site in the
- 29:25past two months or you're working in pages, PLPPDP, basket
- 29:30checkout. That's how you optimise is page
- 29:32templates, which really should be thinking about stages,
- 29:36browsing, committing, evaluating.
- 29:38So we do say #stages, not pages. And so, yeah, so the that's
- 29:45generally the data maturity is more important than the traffic
- 29:48I would say and. What's I would like this
- 29:53question for, for like founders and where they're headed
- 29:57strategically in vision is what is your vision for the future
- 30:00like where are you headed? Is this just about continuing to
- 30:04do what you're doing or is it about evolving the product?
- 30:09You know what, It's not evolving the product at all.
- 30:11It's evolving the market. And again, I appreciate that
- 30:13sounds culty, but just appreciate like the journey at
- 30:18least, which is I've gone from an accidental agency odour to
- 30:24cathartically writing out 80,000 words in a book with a hell of a
- 30:29lot of research behind it. Having, you know, let's be
- 30:33candid here, having loads of panic attacks in my previous
- 30:36business because I had conflicts about I had to own this thing
- 30:40that I didn't want to own. The one thing I can tell you
- 30:44with this is that there's so much purpose behind it.
- 30:47I can't, I can't begin to describe it.
- 30:49And OK, like the the love of football, the inspiration behind
- 30:52that really props that up as well.
- 30:53That's why I called it my icky guy moment.
- 30:57So the, the vision, the vision of being honest is for over half
- 31:02of all global retailers to be using intent in some form.
- 31:06And we call this money goal for e-commerce.
- 31:07I know for well, others will quote UN quote copy this.
- 31:10It doesn't really exist right now.
- 31:12Spotify uses intent as the personalisation framework that
- 31:16they have. So do Netflix, Google, what do
- 31:19you think rank frame was in 2019?
- 31:21Intent is everywhere and what we do, it's just not an e-commerce.
- 31:24I want to bring it to e-commerce, so copiers by all
- 31:28means, we're just here to be the authority in it.
- 31:32I like that I like that bold statement and earlier mover
- 31:36advantage is always, always useful and beneficial.
- 31:40I've really enjoyed this. I've always I've enjoyed both
- 31:43chats we've had in the recent weeks and it's a really cool
- 31:46project. I'm hoping that I can find the
- 31:48right project to work with you guys on.
- 31:50It'd be be a lot of fun. I I agree, I think the principle
- 31:53about predictive intent is really, really important and not
- 31:57treating people all the same because yeah, the, the, the
- 32:00historical view of you optimize a page, a template versus that
- 32:05there is different user journeys for different types of browsers
- 32:08based on their intent. And that actually a template
- 32:12isn't just a fixed thing, It can be a dynamic thing depending on
- 32:15who's viewing it and for what purposes.
- 32:19I think that scares a lot of people off, but I, I guess my
- 32:23final question and link to that is so things sometimes people
- 32:26think, Oh my God, this is like really complex.
- 32:28How how do people, how do people start simply to not over
- 32:33complicate things? Where what's your advice to be
- 32:35when people come and go? But this feels like it's
- 32:37optimised an entire site for all of the different users for
- 32:39intent. I can't possibly execute that.
- 32:41What's your advice? How do you start small and
- 32:43actually prove it and get some benefit?
- 32:46Well, exactly that number one. You either A, treat it as a
- 32:48framework. I the way I'm going to
- 32:51personalize or segment the audiences in my site is not
- 32:54arbitrarily on page templates. That's ridiculous.
- 32:56That's not how people buy, nor is it on attributes like new
- 33:00versus returning. I'm sorry, that doesn't exist
- 33:02anymore. It's based on their intent to
- 33:05purchase. It's guiding users and holding
- 33:08their hand throughout that journey.
- 33:10So either A use it as a framework or B, you're right,
- 33:13start small, Take it. How about this and challenge you
- 33:17take an AB test that you've ran if you're there, segment it by
- 33:21intent and see what happens. I can give you evidence right
- 33:26now if you'd like that. If you were to segment your
- 33:28latest social proof experiment that you run to put fifty people
- 33:33have also purchased this on the website and you segment that
- 33:36experiment by the intent of the user.
- 33:39What you will most likely found, as we have found is that that
- 33:42type of messaging works really well for users who have a high
- 33:46level of intent that need a level of persuasion, but really
- 33:50poorly to the point where detrimental for those users who
- 33:53have a low level intent, they are not there yet.
- 33:55That's like turning around to a user and saying, you know what,
- 33:58I'm wearing some lace Nicki's, Nicki's store, some Nicki
- 34:01knickers. I'm wearing some lace knickers.
- 34:03It's just an invitation intends to more prioritize the content
- 34:06that needs to be prioritized. Test on intent based on what you
- 34:10already are doing. There's a good chance good to
- 34:13end a podcast episode on a challenge.
- 34:16And if anyone listening, is anyone listening wants to like
- 34:19discuss any of this challenge, you find out more, talk about
- 34:23how they could actually do that in their testing framework.
- 34:26How do they reach out? I'm usually ranting, probably
- 34:30with a drone on LinkedIn somewhere, but made of intent.
- 34:34dot AI is our URL. We have a lot of nice concert
- 34:38around there. Today we literally released a
- 34:39learning centre, which I think is quite cool, Some articles and
- 34:43playbooks. So yeah, through those
- 34:44mechanisms. Obviously on our landing page
- 34:48when you're listening to this, that's inside Commons dot FM and
- 34:51there will be a link to David on LinkedIn as well.
- 34:54David, really enjoy this. Thanks.
- 34:55We could definitely sit here and talk for hours, but I know that
- 34:58that people would eventually fall asleep listening because
- 35:01podcast listeners have a finite attention span.
- 35:05So yeah, thanks for taking the time and the latest news on the
- 35:11journey to world domination. Haven't got that far, but
- 35:15thanks, James. Thanks, Paul.
- 35:17Cheers and thanks everyone for listening.
- 35:18Make sure you're following us on LinkedIn and have subscribed to
- 35:21the Inside Commerce newsletter. We drop it every two weeks.
- 35:25Take care everyone, you're listening to the Inside Commerce
- 35:28podcast hosted by experienced consultants James Gerd and Paul
- 35:32Rogers. Inside Commerce is your fast
- 35:34tracked digital knowledge, featuring interviews with
- 35:36leading e-commerce retailers, industry, Thoughtleast and the
- 35:39C-Suite, key tech companies to explore the latest digital
- 35:42strategy, customer experience and tech.
- 35:44Discover what the latest trends really mean and get practical
- 35:47advice based on the experience to help fine tune your ecom
- 35:50strategy and text. Follow us on LinkedIn and reach
- 35:53out to James and Paul if you'd like to discuss an ecommerce
- 35:56project. We hope you enjoy the latest
- 35:58episode.