Latest / Inside Commerce: Ecommerce Strategy, CX and Technology Podcast / EP302: Confessions of A Customer Intent Expert: Lessons In Experimentation & Why Context In Data Is Everything, With FoundIt! CEO Warren Cowan
Transcript
- 0:00We think Elton John's in the test side, we think he's a
- 0:02skewing factor in the reason why the AOV and the RPB tanked or
- 0:07appeared to tank so heavily, because actually it was
- 0:10unnaturally I and possibly had a look at in the data and there he
- 0:13was in the in the control. So basically he'd spent all his
- 0:17money in the control side of the test, which made the experiment
- 0:21look really bad. You're listening to Inside
- 0:25Commerce, a weekly ecom strategy, technology and CX
- 0:28focus podcast hosted by experienced digital consultants
- 0:31and re platform specialist James Get and Paul Rogers.
- 0:33Get insights into ecom strategy, leading brands, hear from
- 0:35industry thought leaders and leading tech vendors with
- 0:38practical advice to help you improve your online CX.
- 0:46This episode is sponsored by Shopify.
- 0:48Got an interesting topic today? It's, it's all around data and
- 0:53the context of data and being able to understand data to make
- 0:55better decisions for ecommerce websites and drive performance
- 0:58improvements with a kind of dual part.
- 1:01We're going to start on confessions of an
- 1:02experimentation expert. That's that's a nice hope to get
- 1:05your interest. And there's some intriguing
- 1:07stories about where tests gone wrong because I'm some real
- 1:11outliers and why it's so important to understand the
- 1:13context of that. Then we're going to evolve into
- 1:16a e-commerce platform that is helping brands improve like
- 1:21contextual marketing, merchandise and on the site.
- 1:24So that's the intro, that's what we're trying to do today.
- 1:27And a warm welcome to our guest who's Warren Cowan is the CEO
- 1:30and founder at Founder. How you doing, Warren?
- 1:33I'm fantastic. Thanks.
- 1:35We appreciate you having me and thanks for calling me an expert.
- 1:40Only an expert because I've got so many things wrong.
- 1:42Well, but this is the thing isn't I having, you know, saying
- 1:45just before we start recording? I do just done a big re
- 1:49platforming project and I've heard people say so is that your
- 1:52expertise as well? I, I still get things wrong.
- 1:55I still learn and this is the most important thing is that if
- 1:57you don't get make mistakes, you can't learn so.
- 2:01Yeah. Yeah, exactly.
- 2:04So I'm really looking forward to it.
- 2:05So let's before we get into. The the, the the.
- 2:08Questions. Be really interesting for you to
- 2:10share your background and founder story and I guess how
- 2:13you've got to your current technology project.
- 2:18Yeah, of course. I mean, I, I ran a digital
- 2:21agency for 20 years. We were talking about green
- 2:24light. It's that that was the agency
- 2:26that I ran. So I found it that built it, ran
- 2:28it. That was a 20 year journey
- 2:30focusing on, you know, we started on SEO, we've got some
- 2:33other performance marketing and other sort of digital, digital
- 2:37methods. After that.
- 2:38And about halfway through that journey, what we found was that
- 2:41every time we wanted to you know, improve the business's
- 2:44digital marketing, we had to go back and change things about the
- 2:47website. And the more we did that, we
- 2:50started to think, wouldn't this be better if we built the sites
- 2:53the way that we kind of wanted to use them at anger.
- 2:56So we actually had a bit of a step back moment.
- 2:58We thought you need to start building sites.
- 3:00So we started building websites and we started to, we built our
- 3:04systems integrator and economists integrated.
- 3:07We started doing SAP Hybris work.
- 3:09We started doing so do more aware as well our Salesforce
- 3:13commerce class and you know, so we worked with different
- 3:16platforms. We started to build the sites as
- 3:18well. And we spent the first part of
- 3:20our our careers always looking externally thinking how is it we
- 3:24help customers outside the business find the business, find
- 3:27the products that the business sell.
- 3:29And the more we did it, the, the job of building the sites, the
- 3:32more I started to uncover this idea of customers are struggling
- 3:36to find products on the site. The whole customer experience
- 3:40piece is, is, is an enormous part.
- 3:44And you think when you, when we put a lot of websites together,
- 3:46you know, we love an accelerator, don't we?
- 3:48Anything that can help us get all those things out-of-the-box
- 3:50just that extra bit quicker. Standardized grids, standardized
- 3:53templates, cookie cut, you know, the the dynamic system helps our
- 3:57sites stand up out-of-the-box. And that was when we started to
- 4:01tune into these finer issues. That actually was still a real
- 4:03problem for customers to find the right product on the site
- 4:07and the noise of thousands to navigate through PLPS and do
- 4:11search and find the right arriving on APDP trying to
- 4:14expand their view. So we encountered all these
- 4:16other issues and so we built a lot of technology to help
- 4:21optimize websites for their external projections for things
- 4:25like Google and stuff like that. And so we started to bring that
- 4:27technology to start to focus more internally on the websites
- 4:31as well and creating a better experience that was more attuned
- 4:34with what we could understand people were looking for.
- 4:37And that technology stack just grew and grew and grew and
- 4:39eventually became what what is now founded, which is basically
- 4:44a, a technology platform that can help a retailer understand
- 4:49customer intent, work out what it is that customers are looking
- 4:51for on any page, any touch point, and work out how to
- 4:55optimise the journey to make those things easier to find or
- 4:59buy. Whether that comes in through
- 5:01the PRP or home page or search or it's APDP, wherever the
- 5:05customer is, you always want to be able to go.
- 5:07I think I know what you're looking for.
- 5:08It's this way. Let me help.
- 5:11And that is what we kind of felt was the the real missing piece
- 5:14in, in, in the e-commerce operation.
- 5:17Everything was built for customers to come into the site
- 5:21and find their way muzzle through as as we call it.
- 5:25But actually, it was never ahead of the helping, the guiding and
- 5:28beating. I got you.
- 5:31Here's what you. Oh, you want a camera back?
- 5:33No problem. We got 20 of them.
- 5:34They're right here. Yeah.
- 5:35Let me help you. Let me help you break that down.
- 5:37There was nothing that was doing that.
- 5:39And so that's. Really where we started to apply
- 5:41our focus and more. We did that all.
- 5:43We saw the experience improved, sort of conversion improved and
- 5:46that's really how that the founder platform came to be.
- 5:50Yeah, it's really interesting actually because having having
- 5:53been in the industry an equally long time over a couple of
- 5:56decades, I, I came from a marketing background initially
- 5:59and intent was really strong in terms of marketing strategy and
- 6:03understanding and an e-commerce star.
- 6:05It was very functional, wasn't it?
- 6:07And then you're starting to see the the the application of those
- 6:10multi principles into the E com technology stack.
- 6:13And now I think as the techs advance so much that the ability
- 6:16to do this is is like is increased exponentially.
- 6:19So I'd find it really interesting to see how strong
- 6:21the tech is evolving in in this digital space.
- 6:25Yes, to get into a bit more details.
- 6:27But before we get onto the platform itself, I want to talk
- 6:30about the contextualisation of data because obviously the
- 6:33critical thing about what you're you're doing from a technology
- 6:35point of view is using data signals to understand intent to
- 6:39improve context. But I wanted to start based on a
- 6:43few posts you put out on LinkedIn research I loved which
- 6:46was about it's tests. That you run.
- 6:49So you're obviously using all this data, you're finding
- 6:52opportunities and ideas and you're testing them and learning
- 6:54what works and what doesn't work.
- 6:55But posting out some really interesting ones where you've
- 6:59had what looked like really good tests that have been impacted by
- 7:03outliers or, you know, really unexpected issues creeping into
- 7:09test or control cells. And I thought this would be a
- 7:12really good starting point because I don't, I don't think a
- 7:14lot of people necessarily think about that.
- 7:16How do you, how do you triage tests and make sure and
- 7:19scrutinise the so you've properly understood that the
- 7:21data and the results are reliable.
- 7:23So I want to start. I want to start an issue with
- 7:25one question which is is it easy to miss retest results and make
- 7:29compromises? Incredibly easy to miss retest
- 7:32results. I mean I I should?
- 7:33Probably preface, what you know what I'm saying is that I didn't
- 7:38build out an expertise in conversion rate optimization and
- 7:42testing. We took testing on as a
- 7:44commitment really it found it to prove the methodology of what we
- 7:49were doing. You know what we're, what we're
- 7:51telling people to come in and change their websites, orientate
- 7:54them this way, change your operations, change this part of
- 7:57your stack. It, it requires a lot of
- 7:58justification for, for anyone to do, you know, you can't get
- 8:01anything out done on a road back without business case and so on
- 8:04and so forth. So it, it became absolutely
- 8:07essential for us to prove that if you design your site as we we
- 8:12call it, structuring it around customers rather than
- 8:14catalogues, you will see an improvement in the business.
- 8:18So we hope that that will mean customers more of them are
- 8:21finding a buying and you'll see that in your engagement rates,
- 8:24in your add to cut rates in your basket rates.
- 8:25So we had to become experts in testing what it was that we were
- 8:29doing. And so you start by testing, I
- 8:34suppose the most easiest kind of the most obvious kind of
- 8:38metrics, which pretty much go right up to the pool, which is
- 8:41does the conversion rate go up? It's going to be very hard to
- 8:44justify a significant project, 10s of thousands dollars or
- 8:48hundreds of thousands of dollars pounds if it's not going to
- 8:51improve any kind of a core metric that tries to distance.
- 8:55So, so we started focusing on on the conversion, but then people
- 8:58always want to know about the other metrics as well, the AOV
- 9:02and the RPV and specifically the RPV, because the RPV is, if you
- 9:06like, often held up as that gold revenue per visitor.
- 9:09That's the gold metric, right? Which is the AOV, the conversion
- 9:13rate and the, the users all kind of smashed together into what?
- 9:17Which kind of makes it a really dangerous metric because when
- 9:21conversion, you're measuring conversion conversions either
- 9:23happen or they don't like they're quite binary.
- 9:25So it's very, very easy to test and to get a good statistical
- 9:29measure on that. Whenever you start to bring in
- 9:33AOV, I've showed a value or the basket value, and then you start
- 9:37to use that as a, as a amplifier or a factor against the others.
- 9:43You're not in a binary. Did they convert?
- 9:46Did they, didn't they convert? You're now did they buy 100
- 9:48bucks worth of stuff? Did they buy 250?
- 9:51Did they buy a 2000 LB catch? You know what, what was it?
- 9:56Was it a $10 lipstick? What, what was it that they
- 9:58actually bought? And so now you're trying to
- 10:02track a metric which is going all over the place, do massive
- 10:05peaks and troughs over different days.
- 10:07It's never the same one in a given moment.
- 10:09And that has the ability to sway your, your reading of the test,
- 10:16particularly as you go to the RPV, which is the revenue per
- 10:18visitor. So you might have increased the
- 10:21conversion rate, but there's just been a few more sales or a
- 10:24few less sales, particular type of products, particular type of
- 10:27price point. And now that AOV is skewing the
- 10:31revenue per the number of visitors that you had that you
- 10:34converted and you can see very, very different results.
- 10:37So sometimes that the gold metric that they want to hold up
- 10:40of RPV is often times one of the most dangerous metrics and the
- 10:44easiest to misread when you get the test because you can
- 10:47increase the conversion rate, everything can look fantastic,
- 10:51but then everyone wants to know, well, you look like you made a
- 10:53few pounds less there on the on the average order value in that
- 10:56test segment. Potentially, yes, there's no
- 10:59statistical indication of whether that's significant or
- 11:02not. But any any focus on a, on a, on
- 11:06any of these metrics that seems to go the wrong way often gets
- 11:08misinterpreted in the light of what is obviously the.
- 11:13The should be private. On which is.
- 11:15More people bought, we now have more customs, people get caught
- 11:20hung up. And yeah, especially if it goes,
- 11:22especially if it skews heavily in the wrong direction, which I
- 11:24think could, that's what's asking about.
- 11:27What do Mattel, Victoria, Beckham, Sennheiser and Glossier
- 11:32all have in common? They power their businesses with
- 11:35Shopify. Shopify is the most innovative
- 11:40and scaled commerce platform on the planet.
- 11:43It also happens to have the best converting checkout on the
- 11:46planet. Shopify, powering the future of
- 11:51enterprise commerce. Find out more at
- 11:53shopify.com/enterprise. Yeah, exactly.
- 11:57There's there's three examples you shared which I love.
- 11:59And for those listening, it involves celebrities.
- 12:02What more do you want on the podcast?
- 12:05Like insight and expertise on how to be better at doing
- 12:08something whilst discovering how random celebrities can scrap
- 12:11your AB testing and keep. It yes I want to get 2 of them.
- 12:15This is just one of, you know. Very pretty much isn't.
- 12:18But yeah, you, yeah, you're absolutely right.
- 12:21Celebrities, things you did, things you did plan for, things
- 12:24you couldn't see, couldn't have seen coming at the top.
- 12:28And I think everyone who's worked in E com for long enough
- 12:30and done enough, like projects and tests will kind of test the
- 12:34fact that something weird can happen where it just screws up
- 12:38all of your planning. So let's let's go into these
- 12:40examples, because I'm going to start with one that that
- 12:43shouldn't have made me laugh because I know it involved like
- 12:45caught screwed up work and having to rethink things
- 12:48through. But can you explain to our
- 12:52listeners how Elton John killed a Test for matches fashion?
- 12:55He did, yes. I'm not clearly, but maybe he
- 12:59did. They brought him in specially.
- 13:03So obviously we were running a lot of tests on on matches at
- 13:06the time and everything was looking very, very good.
- 13:10Actually. The conversion rates were going
- 13:11in the right direction and yeah, Aov's kind of holding very, very
- 13:15similar levels. And then it got to the end of
- 13:19the test. I think no one had looked at
- 13:20the, you know, you should always let the test run.
- 13:22People get very caught up in looking at every little bump in
- 13:25the road run a test. But we let it run and we got to
- 13:28the end of the four week period. I think we ran the test for and
- 13:31but halfway through the conversion rate looked like it
- 13:35was really, really strong in the test.
- 13:38But the AOV had tanked, at least in the test segment.
- 13:43But what had actually happened was the AOV hadn't tanked in the
- 13:46test segment. It had gone up in the control
- 13:49segment and what had happened was that Elton John, who was a
- 13:54really big fan of matches fashion and I'm sure people see
- 13:58this, we've all seen his wardrobe on the news.
- 13:59The man is a a fan of eye watering Lee Price clothing of
- 14:04all sorts. He was a really big fan of
- 14:06matches fashion. He was one of their VIP
- 14:07customers and he had gone to the site or his entourage maybe had
- 14:11gone to the site and they had bought 10s, hundreds of
- 14:15thousands of pounds worth of clothing and items over a series
- 14:19of different sessions. And what had happened was that
- 14:23that user had gone into the control side of the test.
- 14:27And so now that user was now creating a very, very small
- 14:32number of very, very high value sales, which obviously overall
- 14:35makes the average order value against all the users that far
- 14:39higher based on the total revenue on that side of the
- 14:43test. And actually makes the
- 14:45experiment looked quite poor. But never even knew this until
- 14:49after the test had finished. Because when the test had
- 14:51finished, that high AOV had led to a very low AOV by comparison
- 14:58and a very low revenue per visitor in the test segment.
- 15:01And so everyone thought that comparing the two that the test
- 15:05had hurt sales. So even though that the
- 15:08conversion rate had improved overall, the business had made
- 15:11less money and the business wasn't interested in anything
- 15:13that could make them less money, at least in the comparison
- 15:17between the two segments of tests.
- 15:19So, you know, we got kicked into touch.
- 15:22We're sort of getting our getting our, our coat ready to
- 15:25sort of leave through the side door.
- 15:27And then it's almost like you couldn't have planned it.
- 15:32The the television was on in the office while we were working.
- 15:36And yeah, you had Sky News playing on or something like
- 15:39that going on in the background. And Elton John's on, you know,
- 15:43we start sort of could hear a well, which one's talking about.
- 15:45And he's talking about his antics, the fact you got some
- 15:48upcoming tour dates and he's just had a massive splurge at
- 15:52batches fashion and like, wait, you should spend how many
- 15:57thousands of pounds of batches fashion?
- 15:59But no, it couldn't. That would have been just over
- 16:02our test. I was, I wonder which segment he
- 16:05went into. And that's when we brought this
- 16:08idea back to, to, to the, to the testing team.
- 16:10When we think Elton John's in the test son, we think he's a
- 16:14skewing factor in the reason why the AOV and the RPB tanked or
- 16:19appeared to tank so heavily, because actually it was
- 16:21unnaturally high. And of course, we had a look at
- 16:24in the data and there he was in the in the control.
- 16:27So basically he's spent all his money in the control.
- 16:31Side of the test. Which made the experiment look
- 16:35really bad just for that that short period of time.
- 16:39I curse him in that I did. I'd like to think that there's
- 16:43Elton John late at night on the match's fashion website and he's
- 16:46just sat there with his credit card out getting ready to place
- 16:49it rather than his team who probably taken our I.
- 16:53Think he's I think he's more of a PayPal man.
- 16:55But so the the, the, the the other one that you which I get
- 17:04not so celebrity, but but equally interesting is about how
- 17:09the New York Police Department went but big on Ray Bans and had
- 17:14another impact on a control segment.
- 17:16Yeah. Yeah, exactly.
- 17:18Very, very similar situation. Yours finest just chimed in on
- 17:21one of our tests. You know, we were testing on on
- 17:24on the Ray Ban site and conversion again, conversion
- 17:29looking good, consistent up all the way through the testing
- 17:33fact. However, again, just like Elton
- 17:36John, someone in the New York Police Department had decided
- 17:40that they were going to come in. So you'd think they'd have a
- 17:43procurement ports for this sort of stuff, but but apparently
- 17:46not. Someone had come in and
- 17:49basically purchased several precincts worth of Ray Ban
- 17:55sunglasses. All I can think of is chips, you
- 17:57know, you know, with the, with the sort of mirror and aviators,
- 18:00but basically purchased hundreds and hundreds of pairs of these
- 18:03sunglasses and they all went into that control side of the
- 18:07test, which of course meant that the AAV sword and the RPV
- 18:13obviously looks really, really strong.
- 18:16And of course that meant that the RPV in the experimental side
- 18:21of the test has obviously appeared to tank.
- 18:24And so again, nobody wanted to hear anything about it.
- 18:26It was one of those a test has lost US money.
- 18:29And I think that's the hard thing that I think there's that
- 18:32that's the misread. I think that's the core of the
- 18:34misread. This test has lost US money, but
- 18:36we've got loads of more customers, but somehow we've
- 18:39made less money. And that that paralyses decision
- 18:42making. It, it creates A fierce state,
- 18:45especially when it's, it's something that no one appears to
- 18:48be able to explain as well. Because when you do a test, you
- 18:52often get to the end of it and it's like, oh, this one won or
- 18:55this one was appeared to outperform.
- 18:57And there's a we have a 95% chance of statistical chance
- 19:02that this was didn't occur by by chance.
- 19:06And but you don't have a lot of underlying information unless
- 19:11people want to go through the raw individual transactions and
- 19:15use your state where there's a lot of data.
- 19:17Typically people tends to want to move.
- 19:19They've done the test, they've got that finally, they want to
- 19:20move on. And there's a lot of that
- 19:22temptation just to, to, to sort of take it as read and move on.
- 19:27And I think no one does the deep dive and looks at the nuances
- 19:31behind the numbers, especially, especially when there's a
- 19:36negative factor. People just can't.
- 19:39They can't get over it. Yeah, I, I think there's a real
- 19:42emotional side too, because when you test, and especially if you
- 19:45feel like you've got a great idea and that test is negative,
- 19:47there's a real like, oh God. But, but equally, I think the,
- 19:50the, the, the important outtake I get from this is, is you've
- 19:55got to be really crystal clear on what the objective of your
- 19:58test is and what success measure is.
- 20:00Because you can, you can measure so many different metrics and
- 20:03you can have external influences that can change one or more of
- 20:05those. But if if the if there's a key
- 20:08behaviour that you're trying to influence positively, and you
- 20:10can measure that succinctly, then if other KPIs change and
- 20:13you can validate that those are changed through other
- 20:16influences, then your tests can still be.
- 20:17Successful, yeah, I think you've you've really hit on that
- 20:21there's a key behaviour, there's a key behaviour you're trying to
- 20:23influence and you structure the test around that.
- 20:26But there's always this overshadowing backdrop of the
- 20:28business has its KPIs and it's KPIs should drive the business
- 20:35and everything, all the decisions that's made that it's
- 20:37making. There's a there was a really
- 20:39great article, I can't remember who wrote it quite a while ago.
- 20:42It was about, it was about Amazon and the way that they do
- 20:45measurement and they talked about finding what drives what
- 20:50very, very simple kind of concept, what drives what.
- 20:53And you know, being able to explain metrics all the way back
- 20:56to collections of things that influence behave is a new
- 20:59influence enough of the right behaviour.
- 21:00It's in enough direction. We know that they all drive
- 21:02these behaves, which drive these behaviours.
- 21:04And, and that's how they, they design the business.
- 21:07And I, I have found that, you know, a lot of retailers,
- 21:11despite doing some of their best efforts in, in testing, they
- 21:16understand that they should probably try to test what works
- 21:21and test the behaviours and try and influence them, but they
- 21:24can't escape that management level kind of overarching.
- 21:30Or did we make more money or did we not make more money?
- 21:32Which, you know, everything seems to, seems to come down to,
- 21:35you can almost see this, there's a, there's a little bit of a
- 21:37burn the boats culture in the conversion rate automation, in
- 21:41the conversion rate optimization space at the moment.
- 21:43We're trying to pull people away from focusing so, so resolutely
- 21:48on, on at the end of the funnel revenue and trying to actually
- 21:53break that down, work out what's actually influencing the
- 21:57journeys, you know, on the, on the inside.
- 21:59And naturally having this kind of positive inquiry approach
- 22:03where we try to get more all the behaviours that we don't move in
- 22:06the right direction, moving more in the right direction and and
- 22:10working that all the way through the funnel and not allowing
- 22:13skewing factors to to come through.
- 22:16So let's get into the the found it side of things now and
- 22:20understand a bit more out there that platform because you're a
- 22:22user intent that that discussion comes up more and more with
- 22:26e-commerce businesses I work with there.
- 22:29It's done to varying degrees. They've some it's, it's basic
- 22:33segmentation, we change a bit of content for others it's mainly
- 22:36just about the product piece. And obviously user intent can
- 22:39have influence across all user journeys.
- 22:41So I think if we really start, let's start about with your
- 22:46definition, like what, what does it mean when you're saying
- 22:48delivering contextual shopping experience?
- 22:50Like what are you trying to achieve?
- 22:53I I think when everything boils down for at least for us anyway,
- 22:58everything boils down to trying to get ahead of the customer,
- 23:02right? We want to be able to match
- 23:04support and inspire how the customer wants to shop.
- 23:06That's that's the MO, right? Yeah.
- 23:09And that's really whether I exist, intend exists or not,
- 23:12that really should be the, the MO I think of all of all
- 23:15retailers. If the customer comes in and
- 23:17finds that they're going to be able to find the right product
- 23:20quicker amongst all the noise, be able to make a comfort by
- 23:23decision and, and, and become a customer.
- 23:26If the harder we make that, the, the more likely they are to
- 23:29defer to easier options to go back to Google, try and looking
- 23:32on looking on Amazon, you know, as, as you said in the piece
- 23:35that you wrote for us, they're always one click away from a
- 23:37bounce, right? It's very, very easy for them
- 23:40to, to, to, to, to, to find the product that they want somewhere
- 23:43else. So we've got to create an
- 23:44experience that matches, supports and inspires how they
- 23:47want to shop. The way that we do that is by
- 23:49trying to understand how and what it is that people are
- 23:52looking for as they're coming in through different top points of
- 23:55the site. And if we can do that, we can
- 23:57create custom experience, the shapes around the customer
- 24:00rather than the catalog. We're trying to get away from
- 24:03this catalog driven, dynamic templated cookie cut vest that
- 24:08is just spat out by plugging A catalog structure into an
- 24:11e-commerce engine, pushing out a load of PLPS in various
- 24:14hierarchies. There's you make an have.
- 24:16You can search it if you want. You'll muzzle through, go nuts,
- 24:20fill your boots. And what we're really trying to
- 24:23do is be able to get ahead of the customer every step of every
- 24:26journey, every touch point, every combination of touch
- 24:29points to go. Was that because you were
- 24:31thinking of fantastic, no problem, they're this way.
- 24:34Let's go and that. Let let me help you get from the
- 24:37noise of 30,000 products down to 10.
- 24:39So you can make that choice at 1:00 and, and get you on your
- 24:42way faster, quicker with this aggro.
- 24:44And that's, that's really the, the, the goal that we're trying
- 24:46to pursue. And to do that, you have to be
- 24:49able to understand at every touch point what combination of
- 24:53them, what is it the customers are trying to find?
- 24:56And that's what, that's how we look at intent, customer intent,
- 24:59which on what that means. Customer intent is what is the
- 25:02user trying to find? What is it they're trying to do,
- 25:04what's the mission, what they're trying to fulfil and basically
- 25:08try and make that as easy as possible for them.
- 25:11Yeah. And that you've got a really
- 25:13impressive brand portfolio. I was like Ukes, M&S, John
- 25:16Lewis, there's, there's some big retailers on there.
- 25:19And I've always, I've had lots of conversation people where I
- 25:23think sometimes the this whole like contextualisation is, is
- 25:26misread as just actually want to show the right products to
- 25:28people. And that's one part of that.
- 25:30But coming back to the point you just made about what is it
- 25:32people want is, is depending on the user or the product or you
- 25:36know, how you purchase and what info you need, your intent to
- 25:40purchase could be really low and you could be in a discovery
- 25:42phase. And actually what you need is
- 25:44beyond just here's the top selling product and you need
- 25:47something else. You might need buying guides,
- 25:49you might need persuasion mission, all these different
- 25:51things. So I'd love to hear like some
- 25:53practical examples from the people you work with.
- 25:55What are some of the the use cases of how people have used
- 25:58like intense signals to make better decisions or change what
- 26:01they're doing? Well, there, there's a few, I
- 26:03think that that kind of stand, I mean there's millions of
- 26:05applications of it, but there's a few that kind of stand out.
- 26:08And we see again and again and again.
- 26:11The the first thing that is probably the most common one
- 26:13that we see because everybody is trend gaps.
- 26:15So this is where customers are looking for something.
- 26:18It's maybe immersed as a new idea, maybe it's immersed as a
- 26:21trend, Maybe something's driven it.
- 26:23Barbie caught was a big one, right?
- 26:25You know, hot pick anything in hot pink.
- 26:26That was a trend, something called Barbie caught after the
- 26:28Barbie and people start looking for it and they start going
- 26:32Google, they start landing on retailers, they start putting
- 26:34into search bars. They're looking for categories,
- 26:36they're looking for it and they don't see it, right.
- 26:38So that's what got the trend gap and all retailers have this.
- 26:43It's because again, coming back to this, customers not
- 26:45catalogues. The catalog has design how you
- 26:48can navigate the site, right? The catalog is dictating how you
- 26:52can find information, the structure of it with it and
- 26:55customers searching myriad of vibes, a lot of which are driven
- 26:59by treks and they arrive on site and now they're looking for
- 27:04something and it doesn't really appear to be there.
- 27:05It is there, but the retailer hasn't really named it,
- 27:10described it, attributed it, organised it into those
- 27:14collections. So I mean, put on there on
- 27:16youths water. We had lug sole boots and
- 27:20milkmaid dresses coming out of the woodworks and there's like,
- 27:23well, what's a, what's a milkmaid dress?
- 27:26What's a lug sole boot? It's like, well, actually it's
- 27:29a, we know, we know, we know what that is.
- 27:31We can identify that in the, in the product information, but
- 27:34the, the, the retailer hasn't got that information to organize
- 27:38their products or to tag their products.
- 27:39And so customers are coming in and they're missing it.
- 27:41And of course, once we've spotted that, that can all free
- 27:44straight into the system. Data can be enriched through
- 27:46categories can get created attributes and facets journeys
- 27:50can be supported. And now the customers can find
- 27:52the sort of trend that that they, they can sort of really
- 27:56want to look on to. I guess another aspect of, of
- 28:01using intent signals in improving the experience is to
- 28:06plan and organise the category better.
- 28:09So we were working with the DIY site who I was on one particular
- 28:15category. I think it was back boxes.
- 28:18I don't know if you'd like a back box.
- 28:19Maybe viewers know what about if you ever done any wire in the
- 28:22house? The back box is what you put in
- 28:24the wall. Your wires come into it, you put
- 28:26the plug plate onto it. And that's what gives you your
- 28:28plug socket in the house all goes behind and you know, we
- 28:33were navigating through the the back boxes section.
- 28:36There's 450 odd of these back boxes.
- 28:38That's a lot of back boxes. How are we going to break this
- 28:41collection down? How are we going to enable the
- 28:43customer to filter it? And what do you think the first
- 28:47filter menu was on back boxes? On the on the site, the current
- 28:53site, yeah, on the site would have been something probably
- 28:56weird like material colour or something like that.
- 29:00It was colour there we go be because I often I worry about my
- 29:04back box colour massively it's. Stressful, of course you do
- 29:07what's it's in the wall there behind the plaster work and
- 29:11yeah, it's a really important what colour it is, right.
- 29:14And then I think it was brand I I could tell you a brand of back
- 29:18box from a hole in the wall, no pun intended.
- 29:21And it is like, Casey, this is not how.
- 29:23And again, come back to this match support and inspire how
- 29:26the customer wants to shop. That's that's the MO for any
- 29:29given. So how do we use the intense
- 29:31signals to to work that out? And what we can see is that, you
- 29:34know, the movement products, we can see how people arrive in the
- 29:37category, what they search for, what they search for after they
- 29:40move out of the category. And we can see that actually,
- 29:43what is it that you really want when you're buying a back box?
- 29:46Not most exciting product collection.
- 29:48So forgive me, but what is it you really want when you're
- 29:50buying a back box? Well, maybe it's the number of
- 29:52plugs that I want on the back box.
- 29:55OK, great. And the great thing is they've
- 29:57got this data, but they're not using it to expose the filters
- 30:01that they present to the customer.
- 30:03Then it's going to come down to, well, what colour is it?
- 30:06Well, is it, it was a metal or it's a plastic mug.
- 30:08What's trending right now? Actually, there's USB ones where
- 30:11they support being able to plug your USB directly into a plug
- 30:14sock. Those are the other things that
- 30:15customers are looking for. So by understanding this
- 30:17information, you can go right. We don't have our taxonomy
- 30:21organised properly. We haven't got our product
- 30:24attributes defined correctly. Our data is not rich thoroughly.
- 30:28We go back to the products in the category and redesign that
- 30:32system. So that's kind of the planning
- 30:33prioritisation side of things. Again, trying to make sure that
- 30:37when a customer comes in, the browsing experience, matches,
- 30:41supports and inspires how they want to shop for that particular
- 30:44group of products, OK, then you'll have something that's
- 30:48more expert, expertise driven. And so I think or educatively
- 30:53driven, we're trying to educate the customer.
- 30:56I think as retailers, it's very easy to get caught up in the
- 31:00expertise that we have on our own products, but not realising
- 31:04that the customer isn't an expert in buying that particular
- 31:07product. So if I go into being cute, I'm
- 31:10looking for taps, I might just go to the search engine just
- 31:14like taps, you know, OK, great, there's 2000 taps, but now what
- 31:18do we do? Or do you want to lead me to
- 31:20browse and fascinate through 2000 taps?
- 31:22I was like, well, no, let's start with how customers are
- 31:25going to break this down. So you're probably going to is
- 31:28it kitchen, outdoor, indoor. It's in the kitchen.
- 31:30Brilliant. So then you go to the kitchen's
- 31:32page, you go and all I know is that I want one of those taps
- 31:35that kind of goes over like that.
- 31:36And it's got the pulley out. You bit that you could whoosh
- 31:38around the sink, but I don't know, I've got the words, I
- 31:41haven't got the, the, the right language to describe that yet.
- 31:44Haven't worked that out. It's been like when you you ever
- 31:47go into HMV as a kid and try to yeah, have you got that song and
- 31:50embarrassingly sing the songs of the guy on the counter him tell
- 31:55you what that song is? Yes, that that song.
- 31:58If you got that song, it's kind of like doing that, but then you
- 32:03know you can present it. Well, okay, you want to pull out
- 32:05a pull out that I'll pull out. That was called a pull out tab.
- 32:09Yes, I'd like a pull out tab. Great, here's here's the 10 pull
- 32:12out tab. So here's the 50 pull out.
- 32:14So now we've got the customer down to selection.
- 32:17Even though they didn't know how to describe what it was that
- 32:20they actually wanted, we've helped guide them through the
- 32:23options that actually have led them to the right product.
- 32:26And then I think one of the most exciting ones that we've been
- 32:29working on in terms of using intense signals is working with
- 32:32M&S on creating inspirational architecture.
- 32:37And one of the briefs that they gave us was to create a more
- 32:41gaugey experience. MNS are maniacally customer
- 32:44focused, you know, they want to get everything right for the
- 32:47customer and they're trying to find new ways to engage,
- 32:51especially that the, the younger groups of customers as well.
- 32:54And they want just to improve their holiday shop.
- 32:58So the holiday shop is maybe 800 or 1000 products in it.
- 33:02And it's very, it's organized very logically.
- 33:05And I think all, all retailers, not their fault, all retail
- 33:07websites are organized very logically, very functionally.
- 33:11You know, do you want a dress? Yes.
- 33:12Do you want a black one? Do you want a white one?
- 33:14Do you want a long one? But actually the customer's not
- 33:16thinking in terms of these functions yet.
- 33:19They're thinking in terms of their mission.
- 33:21They're thinking in terms of the, the aspirations, the
- 33:24lifestyle things that they hope to, to want to deal with
- 33:27problems and use cases that they're trying to solve.
- 33:29And they wanted us to build that into the navigation.
- 33:32And so on the holiday shot, rather than it being there are
- 33:35swimsuits and there are tops and there are dresses and there are
- 33:38trousers and there are, you know, all of these different
- 33:41kind of clothing categories. It was organized more around the
- 33:44mission, which is so you're probably going to the beach.
- 33:48OK, well, what's going to the beach?
- 33:49OK, well, that's a straw hat. It's a tote bag.
- 33:51It's some water bottles and some towels and being able to create
- 33:54these products assortments that fulfilled the missions and the,
- 33:59and the lifestyle aspirations that we knew that customers had
- 34:02and wanted to achieve. And so we created a navigation
- 34:06the actually and the assortments that supported that kind of
- 34:10experience and interrogation of the category.
- 34:12So you could go into a holiday shop and it would be, well, you
- 34:14probably want to do comfortable travel or you're doing a city
- 34:18walk, or you're heading to the beach or you step in out at
- 34:20night, want to look glam. And so basically being able to
- 34:23assemble all of the products and create these assortments and
- 34:28these these these journeys for customers that drove a lot more
- 34:31engagement in the category and meant that a lot more people
- 34:35found a lot more product and a lot more product got viewed as a
- 34:38result. So again, it all comes back to
- 34:41if you understand what it is the customer's trying to do, you
- 34:44understand what it is they're most likely to want.
- 34:46You can formulate that into the collections, the journeys, the
- 34:51options that match, support and inspire how they want to shot
- 34:55the category rather than relying on them actually knowing it,
- 34:58being able to model through the sides to find it.
- 35:01And in terms of like intense signals, how many, how many do
- 35:05you capture because there are there are so many behavioural
- 35:07event based signals that you can get out of a browsing session.
- 35:13Yeah, I, I, I mean, without wanting to bamboozle anymore or
- 35:17going to, to, to the algorithm in too much debt, essentially
- 35:21what we're trying to work out is for any page of the site, you
- 35:24could take any page you want. It could be a search page, it
- 35:26could be a product page, maybe it's APLP, even the home page.
- 35:29We're trying to understand what the next step, what the likely
- 35:33next step of the customer is going to be OK or the likely
- 35:37next request for information is going to be.
- 35:40And in order to do that, we're trying to understand all of the
- 35:43people that flow through a particular point and flow out of
- 35:47that particular point. So it's like a big, a massive
- 35:50two way funnel. All the people are going in, all
- 35:52the people are going out and trying to work out all the
- 35:54different things that they came in with and all the things that
- 35:56they moved out to. And what are those particular
- 35:59pursuits? What are they trying to find
- 36:01when they're doing that? And that's the information that
- 36:03we crunch. And, you know, that can mostly
- 36:06come from analytics and that's going to, you know, we're going
- 36:10to be pulling hundreds of millions of rows of information
- 36:12at night for for any retailer. So it's, it's pretty big data.
- 36:16It's pretty big. And the reason we we pull so
- 36:18much information, and this is a particularly different take from
- 36:22the personalization camp because there's personalization has got
- 36:26some phenomenal applications. But what it can't tell you is
- 36:29how everyone is doing something or but it can tell you what one
- 36:33person's doing doing, but there's often not enough
- 36:35information there to support what it is that they're actually
- 36:38trying to find. So you have to pull all of it
- 36:40together to go, OK, lots of people seem to be going this way
- 36:44and what is that? Oh, well, that's the subcategory
- 36:47of kitchen tabs. OK, so kitchen probably a big
- 36:49thing. What's the big thing in kitchen?
- 36:51Well, they do this and search for that, these things.
- 36:54And OK, so this feels like a modern, modern style.
- 36:58So maybe that's an intense signal that we need to roll up
- 37:01inside kitchen. Does that make a lot of money?
- 37:02Does that convert very well? Do we have those products?
- 37:05There's all this decision making, You know, in another
- 37:07words, you might just call this merchandising, but we're doing
- 37:11it at such a data-driven scale. And that's really the benefit of
- 37:16AI that we've got the algorithms that we use, but it's
- 37:20essentially a data-driven picture of how customers want to
- 37:24navigate. And turning that into either
- 37:27navigation, product enrichments, attributes, tags, collections,
- 37:32assortments, just to make sure that we can support every single
- 37:37journey that the customer wants to take that was to.
- 37:41Answer the question in it does it, it does.
- 37:46I mean, I guess it's, I guess the fact is it's across all
- 37:49journeys, isn't it? And this is a reality of, of
- 37:51what intent impacts every part of the site.
- 37:53It's not just a product page or a category page.
- 37:56It's search, it's the ending pages.
- 37:58It's, and it's also then how that intent changes over time
- 38:02because I could, I could be on a site now and I, I could be low
- 38:05intent for a purchase and then I can refine my intent in the next
- 38:10two or three sessions. And actually my product set that
- 38:12I'm looking at has shifted. I could be in the same category
- 38:15but in a totally different brand or price range or style because
- 38:20I've refined and that and the algorithm has to learn and shift
- 38:23accordingly. I think it's interesting you
- 38:25talk about that, that low intent part, because I think regardless
- 38:30of whether you have intent to buy or you don't have intent to
- 38:32buy, you still have a desire to find something, whether it is or
- 38:37not that you actually tend to buy it today.
- 38:39And you might be at different level of clarity on what that
- 38:41item is. And so that's a bit that sits
- 38:43actually above the funnel, you know, in that intent piece to
- 38:46try and work out what what is it that James is trying to find
- 38:49that maybe he's not ready to purchase yet.
- 38:51But if I don't help you find it, what we're on a hiding to
- 38:55nothing really, aren't we? So, so that's why we put that
- 38:58this, this, this idea of prioritizing this journey
- 39:01optimization and find ability at, at such a high level.
- 39:04Because if, if everyone finds the thing that they're looking
- 39:07for, eventually everyone will be able to purchase.
- 39:09But you know, if, if we're only sort of working on the high
- 39:12intent, the people that have got that, we know they're ready to
- 39:14purchase. Firstly, there isn't the data
- 39:16there to, to clarify. And secondly, just fishing in
- 39:19the margins. So we want to make sure that
- 39:20every customer can find the thing that they're looking for,
- 39:23that every product is findable and, and then we'll be able to
- 39:26move the retailer for wholesale. It's really, it's kind of a,
- 39:31there's a different question that we ask, I think.
- 39:33And then that's kind of our starting point for everything
- 39:35that we do, which is the, what if we knew all customers wanted,
- 39:40what would we do differently? And that's the, that's the, the,
- 39:43the big questions. What, what, what would we do if,
- 39:47if we knew, what would we do? And, and that might be a
- 39:51collection of different things, of course, different parts of
- 39:53the operation, different parts of the stack, different parts of
- 39:55the process. But again, it comes back to this
- 39:59organizing the site around the customer rather than the catalog
- 40:02and going, well, in order to do that, you need to know what a
- 40:04customer wants at every touch point that they're going.
- 40:08And that's that's really the objective.
- 40:11Yeah. And I saw it, there was a
- 40:12testimonial on the site from M&S, the digital strategy
- 40:15manager said about how founders help them stay ahead in SEO.
- 40:19And I was interested to probe this more because that's quite
- 40:21generic statement. I'd love to know what that
- 40:23really means because is this about, is this about because you
- 40:26talk about obviously your founder is able to help with
- 40:28like product enrichment, attribute enrichment, which with
- 40:30richer data you've got stronger on page content.
- 40:33Is that is that it? Is it, is that how you happen or
- 40:35are there other specific areas in which the platform is
- 40:37supporting SEO strategies? So, so this is a interesting
- 40:41angle. It's almost, you've got to flip
- 40:44it all on its head really to, to understand the SEO of angle.
- 40:48The majority of retailers SEO strategy relies on information
- 40:52architecture. How well are the categories
- 40:55organized? How well are they orchestrated?
- 40:57How well do they break down around keyword topics customers
- 41:01actually are talking to, right? That, that, that's essentially
- 41:05the, the main component of of the SEO strategy.
- 41:09You got to have the content organized in the right kind of
- 41:12architecture. And So what what you find is
- 41:15that when you listen actually to the intense signals, when you
- 41:17realize that everyone in kitchen taps wants a pull out tap and
- 41:21everyone that wants a pull out tap wants a, a vintage one or,
- 41:25or a Chrome one or something like that.
- 41:26You actually design the information architecture around
- 41:31the search turns that customers are actually using around the
- 41:34topics that they're actually going to Google with.
- 41:36So you have this kind of synchronicity between how people
- 41:39are searching outside of the website and the structure and
- 41:42the organization of the website itself.
- 41:46And when you have that, your SEO flies.
- 41:50And it's probably an interesting side that found it started out
- 41:55as a search engine optimization technology.
- 41:59What we were doing was automating SEO for our
- 42:02customers. And so we were working out.
- 42:04I remember it's one real stand out example where you know, it
- 42:10was a cushions page. They had about 708 hundred
- 42:12cushions and we were redesigning the facets and the navigation so
- 42:19people could find the right cushion it more more easily.
- 42:22And I think the most popular thing was time was Grey and
- 42:25Scandi. And of course their rankings
- 42:27went up for that as well on Google.
- 42:31Because if if you were going to Google and you were looking for
- 42:33a cushion, the most likely things that you were probably
- 42:36going to be looking for was a great cushion or a Scandy
- 42:38cushion. But also the the side is just a
- 42:40microcosm of that. So obviously where we improved
- 42:44the experience on the site and made it easier to find a great
- 42:47cushion and easy to find the Scandi cushion, we were giving
- 42:49Google all the right signals that it needed to be able to go.
- 42:52This is the site I've already ranked for those terms as well.
- 42:55So when you understand how customers are looking for things
- 42:59and you're designing the site around that, you're giving
- 43:01Google everything it needs to to to rank the site stronger
- 43:06because you're creating the information architecture and the
- 43:08cross links and the depth and the breadth and the the
- 43:11hierarchies around all of the topics that customers are typing
- 43:14in. Yeah, and this is, we just did
- 43:17an episode around the AIO views and Gen.
- 43:20AI and AI agents in search. And actually it feels like we've
- 43:24gone full circle in technical ICA because structured data and
- 43:28good semantic data has always been really, really important
- 43:31for you have indexation and for search engines understanding.
- 43:35And also really good structured and well organised data is good
- 43:39for people with assisted technology.
- 43:41So you know the things that you want and should be doing.
- 43:44But now as people get obsessed about AI driven search, it's
- 43:48renewed the interest in getting your data right.
- 43:52You've got to get the data right.
- 43:53I think that's, that's always, it's funny because whenever we
- 43:55were doing e-commerce re platforms, the data was always
- 44:00the last thing that everybody thought about.
- 44:02You spent all this time getting all the things out-of-the-box,
- 44:04building out all the features, designing all the structures.
- 44:07And then everyone would go right now we've got to get the product
- 44:10data from the, we've got to lift and shift the product they for
- 44:12the old platform and get it into the new platform and it's not
- 44:16fit for purpose. And I think that's what a lot of
- 44:19retail and I've been on like this for almost for almost for
- 44:24most of my career now, even as a as a digital marketer, when we,
- 44:29when Google started doing shopping feeds, we had to get
- 44:32people's product feeds out from their site into a structure that
- 44:35Google could work with, which been a huge amount of time
- 44:37optimizing product feeds. And I've still never seen any
- 44:41retailer conquer or any any group retails conquer the great
- 44:46product feed with all of the rich data, all of the accurately
- 44:50attributed data, getting that data rich and relevant and
- 44:54accurate for every single product make, make, make to make
- 44:58to help make it more findable in your system in Google system is
- 45:01also going to make it findable for every future system that's
- 45:04going to be based on AI as well. So getting your product data,
- 45:09and I often say it to our customers, start with a product
- 45:11vector. Everything else follows because
- 45:14once all the product data can be sliced, grouped and matched to
- 45:18any conceivable customer requirement, any system that you
- 45:24don't choose to work with will be able to create an experience.
- 45:28I keep saying the matches supports and inspires how a
- 45:31customer wants to shop. That, that's the kick really is.
- 45:34You get that 100%. That could open.
- 45:38I mean to be honest, the whole like structure, data,
- 45:40atomization data owes up an entire new conversation.
- 45:43Probably spent hours talking about that alone.
- 45:46But we would we would be alone. Yeah, yeah, yeah.
- 45:50I really, I really appreciate you coming on.
- 45:53I've really enjoyed that action. I'm sure that this has got a
- 45:55load out of it. I love the starting on the like
- 45:58the headline, grabbing what Elton John bought to test and
- 46:02then getting into like the the contextual platform piece.
- 46:06It's really interesting for those listening, if you've got
- 46:08more questions or you want to push Warren and the team that
- 46:11found it on on what he's discussed or get a better a
- 46:14better understanding of how it could support you with your on
- 46:17site plans, then the contact details are on the landing page.
- 46:21Do reach out or I'm sure I'll be happy for you to, to LinkedIn
- 46:25message him just like I did. Please, please stick Warren.
- 46:29Thanks so much for coming on. Thank you, Chase.
- 46:31I've really enjoyed talking. Excellent.
- 46:33Take care everyone, until next time you're listening to the
- 46:37Insidecommerce podcast. Hosted by experienced
- 46:39consultants James Gerd and Paul Rogers, Insidecommerce is your
- 46:43fast track to digital knowledge, featuring interviews with
- 46:46leading e-commerce retailers, industry Thoughtleys and the
- 46:48C-Suite, key tech companies to explore the latest digital
- 46:51strategy, customer experience and tech.
- 46:54Discover what the latest trends really mean and get practical
- 46:57advice based on experience to help fine tune your E Comm
- 46:59strategy and tech stack. Follow us on LinkedIn and reach
- 47:03out to James and Paul if you'd like to discuss an e-commerce
- 47:05project. We hope you enjoy the latest
- 47:07episode.