Blog

Podcast | Journey to Retention Growth

Episode 4 Guest: Chris Pook, VP of Retention at Lyst
Podcast
Podcast 4 Pook

In this episode, we’ll sit down with Chris Pook, VP of Retention at a global fashion search company Lyst, and find out what it means to shop for fashion in the era of Google and Spotify. We’ll delve into what’s really shaping the user journey on Lyst – and how Chris is using technological tools and data to grow customer acquisition and retention. And yes, how to make the entire shopping experience smarter and more personalized for mobile users.

Podcast Script:

 

Jean:
"Chris, welcome to the show."

Chris:                    
"Hi, thanks for having me."

Jean:                    
"Why don't you give the listeners a little bit of your background. I cover some in the intro but I'm sure they would love to hear more."

Chris:                    
"My name's Chris Pook. I'm currently VP of Customer Retention at Lyst and this for those people that don't know is a global fashion search platform. It essentially aggregates all the fashion in the world into one site and allows people to still have access and buy it through that site."

Chris:
"My role here is a classically growth and retention role so I touch on the mobiles, CRM, customer care element of the business. We're here to drive growth and retention through adding value to the product. I've had about 15 years' experience so mostly in the marketing/product space. Before Lyst I worked [inaudible 00:00:43] side. I've worked in the travel sector for Carnival and before that I worked in some small start-up, original businesses."

Jean:
"You've got a lot of great background to cover what we are going to talk about today then which is basically how ecommerce platforms are using mobile technology and data to improve customer retention and growth."

Jean:
"Recently we are seeing a lot of changes in how brands are interacting with the consumers especially by fashion, tech or any other tech savvy brands who are learning really fast from CRM and data driven companies such as Amazon and Spotify so that's the first place I want to start with. Last I checked I think it was around 150 million shoppers and close to 4 million skus your platform, Lyst, is handling."

Chris:
"Yeah."

Jean:
"Congratulations by the way."

Chris:
"Thank you."

Jean:
"Given that kind of scale and dynamics you are in in terms of matching buyers with the sellers I expect Lyst to be a heavy user of data on both ends of the interaction. Can you start by describing what is different in your customer journey versus traditional shopping experience say when buying from a department store for example."

Chris:
"The interesting thing about our model is that effectively we're an affiliate site. We're a sort of middle man in terms of that market place aspect. Whilst we support and serve ecommerce you don't directly check out on our site. We'll put you in touch with the retailers in the various other ecommerce sites. We're effectively that middleware and that middleman in that transaction. Now, that's interesting from a two angles for us really in terms of data. A, how we sort of monetize that so currently we generally work on a rev share model with most of our partners. They'll pay us per sale or a percentage per sale and then interesting thing there and the dynamics of the market we're seeing shift particularly in the mobile space actually is around, that's all done on a cookie basis to base. We drop some cookies. Those cookies convert into sales and we get paid via the networks and the retailers."

Chris:
"Now, in a mobile environment the technology there essentially prohibits a lot of those lifetime of those cookies and it makes it very difficult for us to track that aspect really so we're experimenting with actually how we change that model from a mobile point of view and system and that's all sort of engagement. What that then means is that it becomes much more about the final targeting in terms of how we actually play that role."

Chris:
"In terms of the data that we use to profile, and target and those sorts of things. It's all about building up a preference of a user's favorite brands essentially. If we're going to pull you back into the platform at a high level as opposed to trying to get you close to that high intent purchase and push you onto a retailer. If we're going to effectively jump a couple of steps up that funnel and say "Well, actually we want to get you in looking at brands and pulling you into the platform at that earlier stage" we really, really need to know what brands you interact with. What are the value points you look for regularly? Are you a regular sale shopper or is that once or twice a year? Do you check in with certain brands or certain categories a lot? What we tend to do we generally have quite a cold start problem as well in terms of users who generally get to the site. A lot of people that have come on Lyst their first time, when we first get them."

Chris:
"A lot of the early stage of the platform is us building engagement and instruction. We're getting people to look at brands, look at products that we can build this profile and from that within clusters we've got very complex cluster mapping architecture underneath that allows us to say based on interactions with particular brands, particular products to map that out and do some brand adjacency and see where we can say 'If you like brands X, Y and Z you'd probably like these other sets of brands too.'"

Time: 5:30

Jean:
"That's really interesting. I love how you're able to adapt your business model given the fact that a lot of consumer interactions are moving to mobile. Have you seen actual business impact yet? How long has this been?"

Chris: 
"We've seen a decent amount of business impact. I've been in the business 14 and a half years and even in that time mobile was big then but it's huge now. It's a huge element of the fashion space particularly so it's something we haven't been able to ignore. We've always not been able to monetize the app revenue in the same way that we can do with desktop on the online revenue traditionally under the affiliate model but because of the rise in mobile it's something we can't ignore and as a business we have to adapt to the market and to customer behavior. From our point of view it's fundamental to understand how we can actually monetize that model and how we can actually, from a business point of view, add value to the mobile space and those interactions."

Chris:
"We've seen revenue specifically increase. We've been experimenting with a CPC model which is cost per click as opposed to a CPA model as we call it which is a share of the purchase. What that then forces us to do is effectively build a platform and a product essentially that is geared people engaging with it which it sounds like the obvious thing to do anyway but maybe that isn't always the case when we're optimized around revenue and around pushing people basically from high intense searches through to a store effectively. From our point of view it's been a real injection of a pivot and then a change of direction to business really not only from a how do we monetize this better but also how do we actually create a product which is much more sympathetic to how people shop fashion online."

Jean:
"You talked about in the beginning what kind of data you are using to inform your decisions but when you're looking at the entire customer journey and see some of the system domains involved be that previous purchase history, user preferences or identity management do you feel you have enough visibility to make those decisions?"

Chris:
"No, it's a really interesting question actually and I think the general assumption for a business like our, like you mentioned at the top, we're at six million in stock skus, 30 million in total including out of stock products. We got something like 14,000 stores on the platform. You would think with that amount of data we'd get six million new users on a week. You think with that sort of data and those levels of interaction we'd be one of the richest data businesses around. The interesting thing from our point of view is that you mentioned purchase data. As an affiliate, effectively a middleman, we don't get any of that. We get really patchy data back from purchases. We can see that a purchase has happened. We might be able to see a category level but it's very patchy to match back to the [inaudible 00:06:52] that lead to a customer. Now that presents some really obvious challenges around, well actually if we're going to talk to customers about products that they're interested in or they've shown an interest in on the site we don't actually know if we bought them or not. In a classic ecommerce scenario we can, if we were looking at retargeting or abandonment type activity you don't generally stop that when someone's purchased, when someone's bought the thing or something similar."

Chris:
"For us, we don't always know or we don't know certainly within that 48 hours and we may never be able to push that back in. From our point of view we have to base a lot of our decisions using browse data and then making some predictive assumptions on that to allow us to execute, campaign, data in real time but also execute stuff around preferences and stuff like that. Actually from our point of view as a middleman and an affiliate the data landscape is really challenging for us actually."

Jean:
"Seeing how a lot of decisions are becoming more automated by many brands, the landscape, data landscape you are just describing, where are you at in terms of being able to make some of the automated decisions in terms of what content to push to which type of fashion shoppers?"

Chris:
"We put a lot of emphasis into that real time update particularly in terms of our ranking algorithms as we call them. These are the things that decide what products we show you if you're looking at [inaudible 00:08:07] for example. Now, the problem I mentioned at the top around that cold start issue where a lot of people come to the platform we know nothing about them at that point. We have to then, through interactions and through position things in certain ways, [inaudible 00:08:19] data uses profile on the fly essentially to be able to then start to prove to them that we can actually handle their query."

Chris: 
"From a business perspective having all these products is a fantastic thing but it presents a real customer challenge in that no one wants to be presented with six million products because no one is interested in six million products. We have to try and take that six million and break that down to a thousand essentially in most cases. We put a lot of emphasis into the on the fly algorithm updates in terms of how that works but we also then try and store and persist those preferences that when we see a user again, particularly if they're a member but even if not we store that which, again, is a challenge in the mobile space around persistency and in terms of browser and cookie settings to allow us to store their preferences that we may have built in session one but we try and then flow that through and make those automation decisions in terms of what we show you."

Chris:
"On the campaign side of things in terms of our push strategy if you like, so that involves email notifications and those sorts of things. A lot of that is come up with a lot of challenges like we've just described around when we should actually try and push for certain things or launch certain notifications based on not knowing if someone's bought something or not knowing if something's of a significant interest or not. We built some models to allow us to, effectively some propensity markers to allow us to identify on a user level the items they've looked at and how likely they are to buy them and that allows us to set some thresholds to say "If certain products are below certain thresholds they're probably not of sufficient interest to the user for us to warrant push communication" so email notifications or push notifications via the app and those sorts of things. We found that super successful in trying to manage some of the many messages and some of the many push aspects that we can push in front of people to make sure that we're providing information on products that people like."

Time: 12:10

Jean:
"You just mentioned some of the use cases, notifications and things of that nature and given about 150 million shoppers I'm guessing there are preferences when it comes to communication channels they are most responsive to may vary greatly sometimes, some may prefer texting versus push notification, some email, some chat apps. What are you seeing among your customers?"

Chris:
"We actually see that email is really dominant particularly in terms of that push strategy. We combine email with various other what we call push forms communication, so push notifications on the app. We're also experimenting with priority notifications along that aspect as well. We also do push advertising in terms of dedicated Facebook advertising per user but email on pretty much every level from an engagement level, from an incremental value level, from a revenue level. It trumps everything. Email traditionally has always been a big element of the fashion space. It's a key way that a lot of people start their shopping journeys with brand retailers and those sorts of stores but we find it has the same impact for us so we use it as a way to pull people back into the platform essentially. There's various forms of value that we can offer that. We offer a range of tracking tools. Because we've got so much data and so many products we can track when prices drop on a particular product. We can track when things are back in stock if they're out of stock and those sorts of things."

Chris:
"To offer these utility services to customers to make sure that they can get the value out of having so much data that we have but we can also then extend that and look at brand level preferences and let the users know when a certain brand's on sale, when there's new brands they may be interested in and they can track those designer category level searches like Gucci bags, those sorts of things. We also take what's probably a big change over the last year for us in terms of how we approach email is doing it with that incre-mentality then. It's very easy to get sucked into email being a really super effective tool to drive engagement, to drive revenue, to drive all those things and then to hone in on where the biggest opportunities are to drive that engagement. That tends to be where the people are most active so what we tend to do in terms of our email stuff or the changes that we're currently going through there is to make sure that we are looking incrementally."

Chris:
"If people are super active on the platform and getting the value that we want them to get we leave them to it. If they're not or we think actually we can enhance their value experience and we contribute more to them by letting them know about features that they may not be interacting with or if certain things happen around products they like we then view that incrementally and say "Well, actually how can we add to that by our push strategies, or by email via push notifications and draw people in that way?" That allows us to get this balance between frequency and making sure that we're driving people back into the platform to experience the things they need."

Jean:
"Just curious because that is slightly contrary to what I often hear these days meaning a lot of brands are finding that customers are actually increasingly wanting to have a more personalized, real time communication basically the same way they are communicating and interacting with their friends, two-way messaging and things in that nature rather than emails which is a time delay and mostly one-way communication. Do you see any of that trend catching up with your user base?"

Chris:
"I think it depends on the value that we're offering. I think in terms of that real time, almost like chat type interface. We're trying a new adjacent product actually which offers real time, personal shopping assistance effectively but that only really serves a particular type use case which is "I want to buy something now." I've got a particular problem that I can't solve through your traditional UI, through your search capabilities. I'm stuck effectively. Somebody help me so there's some really cool stuff that we're trying to do with that to allow us to help people better when they get stuck, when there's a bump in the road that we can't help them with but moreover the value that we offer is through tracking, is through those things. It's less real time in terms of the customer's experience."

Chris:
"If a customer likes a particular Gucci bag for example, wants to track that item to see if it goes on sale that inherently has a lag to it. It's not on sale right now but it might be in the next week to two weeks, a month, whatever. The trick for us is to make sure that we don't let that relationship go cold. Someone says "Hey, I want this Gucci bag. I'm prepared to wait for it to go on sale. Get back in touch when it is." The potential for us is that doesn't come back in for three or four weeks so how do we keep that user engaged and looking at things that they like and making sure they're seeing the value in the platform all the while we're searching for that in the background? From our point of view the use cases are many fold. In terms of when someone's actively looking to make a purchase there and then that real time help and that real time assistance is crucial I think. I think people are ever more demanding there in terms of if your search capabilities, if your browsing capabilities aren't up to scratch you need to augment that through chat and through those sorts of things."

Chris:
"I think for a lot of brands still that the inherent value they offer isn't in the moment, I think, the inherent value and that's certainly the case with us. Inherent value that users and members will come back to on a regular basis is periodic. It's checking in on new items from brands and stores you like. It's tracking sale, promos and price drops on items that you like. It's us letting you know when there's new collections and new brands to discover. For us it's about trying to make those relevant and trying to keep those relationships alive over the long-term which is really a key challenge."

Jean:
"As a consumer I consider myself as a shopper who prizes convenience but fashion is one thing I still buy from, well I can't even count how many sources, a lot less now than a few years ago but still which makes me really wonder if buying fashion is really a search problem or something that's intrinsically different about it versus shopping for other things. What does your data say?"

Chris:
"I think it's a really key insight actually and I think it's an insight that's really at the heart of how we're driving products and growth development at Lyst actually. I think you're absolutely right. I think that the fashion landscape is quite complex actually in terms of how people shop for fashion. I think there are various components to it in terms of where people go for inspiration and style advice and we see that at the likes of Instagram and Pinterest and still you're classical publications like the Vogue's and things like that are leading there. People are getting their style tips in that way."

Chris:
"Retailers have always historically done a good job there so through their edit and various editorial content to try and stimulate ideas and new trends and then it's a case of how we try and take that and how much of that is repeatable and regular for us to try and take and build into a platform. Actually consumers generally quite like going to look at nice fashion imagery, something they like and find pleasure in on a commute or actually just like spending a couple of hours scrolling through lovely items. We've got to figure out as a product and as a thing that we can offer in terms of value to the customer where we sit within that journey. Where's the jumping off point. What if someone says "Right, hey, now I need this." Generally speaking we say "Leave the inspiration to the guys that do it best." We're here to basically keep a track on the things that you may like historically. We roughly look at it across. There's 50% of predictable searches that you make regularly."

Chris:
"These are the things like you go to the what's new page on your favorite store or you might search for Gucci bags or you might search for a particular brand on a semi-regular basis, maybe once every week, or two weeks, or once a month. Then you've got the unpredictable searches which are "Hey, right, I've got a hot date next week. I need a new dress" or "I'm going on holiday in two weeks, find me some Hawaiian shirts or what's the style going to Ibiza" or those sorts of things. We treat those two things differently and separately. I think the second half of that in terms of that search stuff is where we're trying to tackle it slightly differently to try and bring together all of those different sources to allow people to query that large inventory in different ways, so searching by occasion which isn't a common way to search today, searching by a particular style of [inaudible 00:17:36], those sorts of things. Then on the other side of that which is the retention, growth angle, it's understanding those underlying preferences as customers."

Chris:
"Generally speaking a lot of customers today start and finish their journey on a retail store. They might go to particular department store or a particular boutique or whatever it might be but they generally say "I need some jeans" or "I'm interested in looking at what's new. I'm going to go there, start my search there and probably end up finishing and buying something there over a period of time." From our point of view we then have to tap into that convenience and that ease thing but we're also being able to put together all the things that a retailer potentially can do for you there and try to make that easier for the customer, more valuable for the customer for them to come back and do that through us on a regular basis."

Time: 21:53

Jean: 
"Now, shifting a little bit to the perspective of what's happening on the operational level. Can you walk us through your toolkit a little bit and help us imagine how you're using what to achieve what kind of goals?"

Chris:
"Sure, we're lucky enough that we've got a decent engineering setup here. We tend to build over buy in most cases but we generally take the view of whatever gets us to where we need to get to quicker. That's how we get there. In terms of the analytical stack we have just implemented a first level cross device tracking system that allows us, we call it a consolidated view of the customer and that allows us to basically bind together all the device ID interactions to a unique user ID. Historically before then we'd always needed a user ID via a membership so someone to sign up or login to generate user ID that we track against but a large, large portion of our traffic is either new to the site or doesn't have a login or a membership so we've now created a bunch of user IDs for every single device that we see."

Chris:
"We can basically join and stitch together those devices across, those many devices to that single customer and that allows us to do a few things in terms of being able to persist preferences when we see a user pop up in a different device. It also allows us to model much better so we can take clusters of users that see similar things and manage to join the data that way but it also allows us to create audiences to target with through ads, so through paid media, and social media, and those sorts of things. We can now target a particular user more effectively and get bigger reach that way by knowing it's a particular customer."

Chris:
"In terms of the website, the app side of things we've got two sides to it. We've got the algorithm setup if you like which allows us to basically put the products on the site and serve in a way that it should hopefully make sense to customers and there's sort of two real processes to that effectively. One is what we call deduplication. As you can imagine with six million product or 30 million if you include the out of stock products, some of those are in there from different retailers, same product so we need to what we call dedup or merge these products together to allow us to say "Right, this Gucci bag exists at four different sites and we put it together as one product to understand that." We do that in a couple of ways. We get some data from our retailers. Unfortunately some of that is inconsistent. Not every retailer gives us the same level of information to allow us to track so we're looking at how we can reach things like GTINs and that sort of stuff to make that process easier but we also do what we call internalized perception which is like image recognition."

Chris:
"That allows to do two different things. It allows us to basically look at products that look similar and merge them. It also allows us to use the product imagery to generate tags and generate descriptions effectively which we tend to find actually is a bit more reliable than relying on feeds and metadata that we get from retailers. When we do that because of the sensitivity around that we have a human layer that signs that off effectively or approves that. We'll use our algorithms to generate potential merges and potential duplicates and then we'll have a human. We use a crowd sourcing platform that effectively has an eye cast over that to agree whether that's being merged correctly or not."

Chris:
"Then on the website of things we try and take all of that. We've only got the IOS app at the moment because vast majorities of our user base actually about 85% use IOS so we haven't actually extended that beyond [inaudible 00:21:23] just yet but that's coming very soon actually. Then we take that and we've got a, well we're just trying to move over to a [inaudible 00:21:30] stack to allow us to create a better, different experience of people on [inaudible 00:21:34] so they can actually query that inventory in much more detail."

Jean:
"In terms of ROI among the things that you were talking about as to what worked best for your growth strategy? Can you give us a specific example?"

Chris:
"Something we always come back to is delivering value regularly. In terms of growth that's fundamental to it. There's many different channel strategies, many different ways that we can execute on that but a fundamental level if users don't get value out of the platform there's a very slim chance they're going to come back. From our perspective the biggest ROI drive is, or the bigger value that we see in terms of that LTD perspective that we tend to look at the best is around delivering value. The quickest way generally for us to do that for a customer is through a [inaudible 00:22:16] and those sorts of things. That tends to be the thing that we can deliver on more regularly, more quickly. If a user wants a price drop on a particular item we've got a much higher chance because of the nature of discounts that are available online and being able to deliver that to a customer than a back in stock which may not ever come back in stock based on the availability of the item, how long it's been out of stock."

Chris:
"We tend to focus heavily around what we call activation strategies around trying to get people to see the value quickly. As I say, there's two ways that we can prove that really quickly. One is to get people to see that they can save money through Lyst by tracking items and us delivering the [inaudible 00:22:50]."

Chris:
"The other aspect to that is getting them to set up what we call lists of all their searches. This is getting us to tell you the sorts of things that you do regularly. Like you mentioned just before around that predictable search element, what brands do you track regularly? Are you interested in new items or actually didn't care too much? Are you interested in specific designer category combinations that we can tap into? We basically ask users to setup a particular [inaudible 00:23:16], a list within that that we can then play back to them and to deliver value. That also allows us to underpin some of those profile and analytical challenges that we can play into recommendations. As we sort of mentioned before, the biggest driver of that actually today is email and doing that intelligently but we see the biggest long-term engagements. We take a lifecycle approach. We generally tend to acquire people by [inaudible 00:23:39] search. We tend to then acquire membership and deliver value to them through email. Then we try and migrate them onto the app which has a much higher, we do much better at the pull aspect of that attention, [inaudible 00:23:51] sort of model."

Chris:
"When people get to a certain point we push the app onto them and that's a really great way for people to see easy access brands, those searches and things in a much more sympathetic pace. It's less about us pushing those notifications and that value onto you and much more about you coming, and opening the app and using it in that way to get that value."

Jean:
"Curious, what would you say the percentage of user base actually interacting through your app on mobile versus their computer?"

Chris:
"It's interesting. This is probably a lot of brands, a lot of products. In terms of the engagement back to what we mentioned before it's up to about 70%, those devices. If we take that consolidated device view about 70% of those devices are mobile devices which is absolutely enormous when you scale [inaudible 00:24:38]. In terms of a revenue output only about, I think it's about 16% of those devices convert in a mobile device and it's that data and that level that's prompted us to look at that revenue model and say 'Well, actually from a revenue standpoint people still aren't purchasing on mobile' and some of that's the technical challenges around cookies. Some of that's still a fundamental shift around actually they did a lot of research, a lot of browsing, a lot of checking in on stuff they like on mobile but actually still pull the trigger on the desktop device."

Jean: 
"I understand your users are not actually making purchase transactions on your platform but are you performing any kind of security measures to make sure your users are genuine human beings or anything like that?"

Chris:
"The way we tend to that is actually when users sign up and become members. We then go through and do a validation on that and that's a two-step process essentially. Effectively, upfront before that we tend to, in terms of the background we obviously keep our data secure in terms of that device matching element. We don't do any specific other than the standard safety measure that any website or app still puts in place. We don't do anything specific on top of that apart from say growing our data on the backend obviously."

Time: 30:55

Jean:
"In some ways I find shopping is local, culturally and otherwise including privacy and security regulations. Speaking from my own experience I find buying something online here in Germany can be very different from what I was used to back in the States. Were there any specific challenges you experienced in any specific market either in the front or in the backend?"

Chris:
"That's a really interesting point actually. I think it's a really good observation and I think it's one that's absolutely true in fashion. I think there's two aspects to it. I think fashion inherently is quite a tactile thing. People buy stuff to wear that they like and some people, the online experience would never replace that ability to go into the shop and maybe be treated well and try on a bunch of stuff."

Chris:
"We don't try and tend to cater for that. I think the cultural difference is an interesting one. We're actually just going through a localization, expansion route at the moment. We're focusing on western Europe with a view to then extend into Asia and beyond and a lot of that comes down to assortment. The retailers having a local slant and also that there's products which trend slightly differently across different regions because there are different shipping considerations and so you're back in, [inaudible 00:27:02] they're back into some of the more early days of the platform."

Chris:
"We have a fantastic amount of inventory, fantastic amount of products. We connect someone with a product they like but then they get stung with say £20 or $20.00 duty because it'd be coming from a different country that isn't supported within the [inaudible 00:27:17] and it's hard for us to be able to communicate that to the user. A part of that relationship that they would put onto us and quite rightly, broke down so they'd find a thing they love which is what we're here to do essentially but actually part of that process breaks down because they get slapped with a big, juicy fee at the end of that which clearly breaks down that value that we're trying to offer the user."

Chris:
"We've had to put an awful lot of effort into trying to match the right retailers, the right shipping requirements with the right locations, right domains, the right areas of the world. The other aspect as you mentioned in terms of service element as well is that quite often with a middleman and quite a few users don't see us like that. They see us intrinsically as part of that buying process. When someone finds their product through Lyst it goes through to the retailer. They then are in the retailer's hands and they'll check out with the retailer and it's up to them to fulfill and deliver that item."

Chris:
"Now, if something goes wrong there, the item gets lost, or delayed, or delivered or there's an issue with the terms or whatever, quite often we're then brought into that process. The people think that obviously we're culpable in some way or we're actually part of that process where actually it's just the retailer entirely. Part of our job actually is to try and be as clear about where the boundaries in terms of the customer journey we sit and make sure that people are clear that we're now passing off to retailer. It's their issue but quite often we have to then get in and try and smooth over some of those times when it goes wrong. If you ever look at reviews of Lyst, a large portion of them are around just actually logistically things going wrong as opposed to just the experience of the site and the app itself in terms of how well it can make you find fashion that you like so we have to step in and we've got a team of people that work with the various people across the world to then and try and help manage those relationships as well."

Jean:
"Here is my last question. Looking at what you're experimenting right now I'm assuming you're looking, trying different things, any thoughts or plans on changes you want to bring about and some of the concerns you might have?"

Chris:
"I think the big shift we talked about a little bit so far and it's the shift in terms of revenue model. We're testing out CPC in certain markets and seeing if that's a better way to monetize higher, upper tunnel traffic, particularly mobile so that's a big thing that we're gravitating towards. The other, extending our value through this push strategy, these notifications and all the different forms that takes. As I mentioned before, email's a dominant that we see working really well but looking at how we extend not only the functionality within that but also different types of points and channels that we can offer, so browser notifications, different elements like that but really for us as well there's a real massive challenge around merging, deduplication, getting accurate price information so with such a big inventory it means that we're always fighting hard to make sure that the products that we have on site are accurate in terms of the price, accurate in terms of the stock and also that we can merge and de-duplicate those products to make sure that we give a good experience to customers and that is getting every harder actually with the amount of retailers that we're putting on, the amount of different technologies that retailers adopt."

Chris:
"Not everyone has a common ecommerce platform, or common data structure, or common taxonomy and that's getting ever more fragmented actually with the endless selections out there for different retailers. We're working towards how we can try and standardize and systemize that and codify that for our retailers to make their lives and our lives much easier and make sure that we can really deliver that experience that is truly bringing together all of those items across the web into one thing."

Jean:
"Actually I have another question in terms of I remember a couple of weeks ago I was talking with a fashion brand here in Munich and we are talking about this possibility of using this two-way messaging and dialogue basically to get better insight as to what the user is liking in terms of colors and those things. It's basically using communication as a data application to gain those insights about the users rather than just a channel for communication. Is any of that something that you are thinking about as well?"

Chris:
"Yeah, for sure, I think there's two aspects to how we look at that data application piece. I think one aspect is that direct interaction. I think looking at ways that we can weave that direct interaction into the product value that we offer and that's through helping people with particular queries. I'd say with this personal shopping element. If we can't solve it through the search or through the browse, or general electives that we have how can we offer something on top of that and how can we then [inaudible 00:31:40] some of those real key insights within that."

Chris:
"The other aspect is how we can join together offline information and that's got two purposes really. One is to make sure that we can understand what customers are buying actually offline because we know that the vast majority of the research element of fashion or the browsing element of fashion is done online. There's still a decent chunk is actually offline or the majority of is actually a bit still offline so how we can join together that information from the customer as to what their actually purchasing so there's real leakage there in terms of our revenue model. If we're being used as an engagement platform which again suits this CPC model a little bit more but there's a real leakage there in terms of if people are using the platform without to look at a bunch of stuff, find some things, find the sources available then just going down to their local store to buy that, that's great value that we've added but there's leakage there in terms of our revenue model."

Chris:
"How we can join that up but also how we can try and stitch that offline bit back into this online journey. Actually if the product's out of stock can we try and join together offline and venture into that if there is actually a better price available in-store or better size availability can we connect directly to that store within the platform? A lot of that then is surfacing. [inaudible 00:32:43] find inventory and again we've got challenges with the amount of online inventory that we have. We're looking at then ways at how we can weave that information in to offer that as a better service but then also what are those opportunities where a personal touch really matters. For high spending VIP customers, for people that we can't meet that have a very specific search query that we can't meet through us or traditional tools how can we offer services and things on top of that that really heighten the experience but also means that we can learn a lot more about that customer."

Jean:
"It sounds like a very fast moving industry that you are in. Is there any resources you want our audience to check out to learn more about the topic and what you are doing?"

Chris:
"In terms of from a Lyst standpoint we've got internal engineering blog and some medium stuff that we do to search for that online made by Lyst and we also look to some of the [inaudible 00:33:31]. We ship stuff from the guys at Reforged, Brian Balfour and Andrew Chen and those sort of guys do from a value and a growth perspective and then we look at ways that we can utilize mobile technology and things like that within that great framework."

Jean:
"With that I would like to thank you Chris very much for sharing your thoughts with us and that was a lot of topics we covered."

Chris:
"Thanks for having me Jean, absolute pleasure."

 

Profile picture for user Jean By Jean Shin
Share: