Kunle: My guest on today’s show has implemented personalisation for Amazon, Zappos, Audible.com, Internet Movie Database and over 50 other platforms. He’s going to share tips on how you can effectively harness the power of machine learning and personalisation to upsell and cross-sell products in your online retail store. Do stay tuned.
Hi 2Xers, welcome to the 2X ecommerce podcast show. I’m your host, Kunle Campbell and this is the podcast where I interview ecommerce entrepreneurs and online retail marketing experts who uncover new ecommerce marketing tactics and strategies to help you, my fellow 2xers, grow metrics that matter in your online stores. So if you’re looking to grow metrics such as conversions, average order value, repeat customers, traffic, and ultimately sales, you are in the right place.
On today’s show, I have with me Saurabh Nangia. He is the founder and CEO of TargetingMantra, a SaaS based personalisation and targeting platform that enables small and medium sized ecommerce companies cross-sell and upsell like Amazon. Saurabh is a personalisation and machine learning veteran. He has an MS in machine learning with 5 patents pending in the recommendations domain. He’s a renowned speaker in machine learning personalisation and real time user behaviour. In the last 7 years, he’s worked on personalisation engines for Amazon, IMDB.com, Zappos, Audible and over 50 other companies spanning across ecommerce marketplaces, travel and other online portals. On today’s show, Saurabh will be talking to us about personalisation in ecommerce and its impact on conversions. Without further ado, I’d like to welcome Saurabh to the show. Welcome to the show, Saurabh.
Saurabh Nangia: Thanks a lot, Kunle. It’s a pleasure to be here.
Kunle: Could you take a minute or two to tell our listeners more about you?
Saurabh Nangia: As you mentioned, I used to work within Amazon before, so I was working on personalisation for Amazon and Amazon’s own subsidiaries. It was while I was working for Amazon that I built recommendations for IMDB, Zappos, Audible and a bunch of Amazon subsidiaries at that point. Slowly, what I realised is that a lot of these ecommerce companies don’t have the best personalisation solution out there. That lead me to the idea of starting TargetingMantra where I can create a personalisation platform along with a marketing automation platform for all ecommerce companies out there, so I decided to use my knowledge from my studies, from my masters in machine learning and my background with handling data within Amazon and building recommendations for Amazon subsidiaries to create TargetingMantra. We started with TargetingMantra in October 2013 and this is where we are today.
Kunle: Fantastic. When did you start working with Amazon? What was the period?
Saurabh Nangia: I was at Amazon from 2010 to 2013, for around 3 years. One of the first projects I worked on at Amazon was building recommendations for IMDB and the second one was building for Audible, the third one was building for Zappos, the fourth one was building for LoveFilm.
Kunle: They were all Amazon properties trying to build that personalisation experience for users. Was personalisation quite important? Obviously it sounds like it was quite important to Amazon. My question has to do particularly with the take-aways from Amazon. You applying personalisation to Amazon companies or Amazon websites. What take-aways did you take from that experience in terms of the importance of personalisation to ecommerce as a whole?
Saurabh Nangia: One thing I would say which was kind of awakening for me was the fact that even big ecommerce retailers like Zappos, which are doing around $900 million in revenue back in 2010, there was a scope of improvement in personalisation even in companies out there. When I built the new personalisation engine for them, they saw gains over the existing solution. That was one of the awakenings for me, that there is always that scope of improvement even for the big retailers out there. The other thing that I noticed with personalisation solutions is a lot of times people get lost in focusing only on optimising the machine learning user models and trying to get better algorithms, what they tend to forget is scale. Whenever we are building something for bigger enterprises, you need to keep in mind that not only should you be able to provide relevant results back to the end consumer in the form of personalised widgets, but also you should ensure that you scale well with the number of products out there, with the number of users out there while ensuring that your latencies and your solutions are as real-time as possible. Because what you want is that when a user is browsing on the website, you should be able to give them recommendations right away. It shouldn’t take a day to update your learning models, a week to run all those algorithms in the back and then figure out what each user’s preferences are. It should all happen within milliseconds, where we are able to give a personalised view to every end consumer.
Kunle: That’s a very good point. The fact that scale was an issue with companies as big as Zappos. For us to take this into context, could we take a step back, and could you explain how personalisation works both from the machine learning standpoint and the time lapse in terms of how much time the machine actually needs to learn patterns through to scale?
Saurabh Nangia: Personalisation means that you understand your consumers well and you are able to give them a personal experience on the website. In early days, people used to confuse even the smallest steps as personalisation, like a simple greeting on the website where they say things like “Hi, Kunle” and then the normal pages, the fact that they use your name out there, that also is personalisation to some extent, but personalisation goes much beyond that. In terms of products we can choose, let’s say if you’re browsing on any of the ecommerce websites based on what you’re clicking on, what you are adding to the cart, what you are rating, reviewing, purchasing, whatever interaction you have with the website, based on that I can keep on figuring out different preferences about you. Let’s say if you’re clicking on a bunch of different shirts, then I can figure out maybe that you are interested in blue shirts made out of cotton which have strips on them. I can figure out all of those different attributes based on what you have clicked on, added to cart, how you have interacted with the website. Once I learned these preferences, then I should be showing you personalised results in the area that you are interested in, while your colleagues and your friends should see things in different form of shirts which they are interested in. The second level of personalisation was when people started to club users together and they started segmenting users and putting users within buckets. They used to say that “Kunle and Saurabh are both interested in shirts, so we’ll show both of them the bestsellers within shirts”, which is still some level of personalisation, instead of showing you electronic products, I’m showing you shirts. But each user is different and each user deserves having an independent user model, so what we do is we take it to the granular level where we treat each user independently so that the fact that you might be interested in blue striped shirts, I might be interested in let’s say pastel colours, solid patterns. So we’ll see different types of shirts. That’s what personalisation is where you are treating each user differently and giving them a personalised experience. How we do it? We’ll collect all sorts of data on the website, just like you have Google Analytics and all of these analytics solutions out there, which will collect these data points and show them back to you in the form of graphs. We will keep on collecting what they’re rating, adding to the cart, buying, clicking on, how much time they are spending on each page, what are their social likes, there are correlations between things like people like reading books of a certain type or they like to wear certain kind of clothes or they like to buy a certain kind of electronic items. We will collect all of this data for each and every user and we have a proprietary machine learning model on the backend which is an ensemble of five different algorithms and they will automatically keep on getting trained with each and every data point of the user. When you have to compute the recommendations for a given user, we’ll use these five algorithms, figure out what should be the dynamic weightage given to which algorithm based on the user who’s online and then give him personalised results back. To give you a broad idea about how these algorithms work, they can be put into 3 different categories. One is “wisdom of clouds” where you see what happens across the masses on the website. This is somewhat similar to how segmentation used to work early on, the first level of personalisation that people were offering. You’ll see that maybe 1000 users have bought a Samsung TV and they bought a Samsung Blu-Ray player along with it, so there’s some correlation between the two. So if you’re looking at a Samsung TV, I’ll recommend a Blu-Ray player to you. That’s the first level, the second level is where you start taking into account the catalogue attributes. If there are two items, let’s say two different books, and they have a common author, there might be some correlation between the two. Or if there are two books from the same genre, then the two books might be related. The third fact is where you start taking each user’s individual behaviour into account, which is the fact that I was mentioning in terms of the modern personalisation algorithm, you need to ensure that you have the third type of algorithm in place where you’ll not only take into account that “not only is Kunle interested in Samsung TVs, but at this point in time, he’s looking for a 40”+ TV which has an LED screen, with Smart TV support and these three, four features”, while someone else is more interested in the Samsung brand, and he doesn’t care whether it has Smart TV support or not. So even though both of you are looking at the same TV, you guys have different preferences and should see different recommendations.
Kunle: Wow, that’s a lot of data info. With some of the crowds related data or related correlation, matched correlations and individual data on a case to case basis. I never really thought about looking at actions as a data point in the personalisation experience, which is very interesting. What key data points are brands harnessing to effectively deliver a personalised shopping experience at level 3?
Saurabh Nangia: The strongest data point is definitely purchase. If someone is buying a certain product, that has the strongest signal in terms of their liking to that particular product. The second one comes in the form of ratings, what we are rating, what we are reviewing. That means you have a certain sense of likelihood towards that product. Then are the other ones, in this order. Add to cart, then it’s the clicks and then are the social likes. In this order those are the behaviour points that we take into account. In terms of the catalogue data points which we have, the catalogue attributes which we take into account, which could be things like, let’s say two guys are looking at a laptop, there are different catalogue attributes which can be taken into account, what’s the processor type he needs, how much RAM do they have, what’s the hard disk, and all those different preferences. In terms of that, the weightage which is given to the brand or to the material or to the gender or to the product description or whatever, that depends on company to company, that depends on user to user. There is no definite tool. You should always be giving more preferences to two products of the same category or two products from the same brand or two products with the same material or colour.
Kunle: That’s very interesting. You mentioned widgets. The fact that a lot of the personalisation is being executed through widgets. My question has to do with how email marketing actually fits into this jigsaw. There’s been a lot of talk about personalised email automation, not just talk, I’m actually seeing in the frontline businesses harnessing the power of personalisation through email to nudge their customers through to getting more sales off the back of their customers just by delivering marketing funnels through email. How does external personalisation through widgets actually merge with email marketing or are they, at the moment, still exclusive channels to tap into?
Saurabh Nangia: Definitely, both of the two things are related. The personalisation widget fits very well within email marketing. Talking a little bit from TargetingMantra’s perspective, we divide email marketing primarily into two parts. One is targeted and one is triggered. Targeting is something where a marketing manager or a category manager decides that he wants to target certain consumers based on whatever rules he has in mind. It could be things like people who signed up last time and have spent more than $1000 on the platform since then. People who have viewed at least 10 products on the website, but have bought 0 products. So they create all of these different campaign rules and we will then dynamically choose what users fall into that bucket. Now comes in the part of personalisation widgets. Once he has created these segments that he wants to target, he’ll take each of these users individually and see what their preferences are. So instead of sending the same bestsellers to all the users out there, what we will do is see what is the likelihood that a certain consumer will be interested in what bestseller. So I won’t send books bestsellers to everyone or I won’t send electronics bestsellers to everyone or I won’t even send Nikon’s same lenses to everyone. I’ll see that this guy is more likely to buy a Canon lens for his camera while someone else is more likely to buy a Nikon lens. So each one will get these bestsellers also in their email which are personalised to their taste. The personalisation doesn’t just stop at product delivery but it also goes to the extent of what time the email is being delivered at. Simple facts like, let’s say I’m a morning person, I tend to wake up at 6am in the morning, it will send me an email around 5.45-6.00am when I’m more likely to open the email, while a colleague who tends to check his emails after lunch every day, he might get an email around 2pm or 1pm, right after lunch when he’s more likely to check it. So we have not only personalised the content inside the email, as in the products which are present in the email campaign, but we also choose what time we should deliver the email.
Kunle: I reckon you choose the time based on previous open rates? Times they’ve opened the emails in the past?
Saurabh Nangia: Yes. So there are two different factors out there. One factor is “what time have they opened the email?” and the second factor is “what time did they visit the website and buy something?”. To give you an example, let’s say you might tend to open an email in the office right now from an ecommerce company, but you want to consult your partner, your spouse, your girlfriend once you are back home whether you want to buy that particular TV or not. So we’ll then send you an email later in the night when you’re more likely to open the email as well as visit the website within a certain timespan. So our algorithms will figure out that optimal time using machine learning to choose for each user independently. So no more need to create those graphs and figuring out “the majority of my customers open emails at 3pm, so I’ll send it to everyone at 3pm”. What we are trying to advocate in the market, what we are trying to tell everyone is that each user is different, so even though, let’s say 40% of your audience was opening an email at 3pm, maybe the remaining 60% were opening them at different times. So we’ll send that 40% of your audience still at 3pm, maybe we’ll send 10% at 7am, 15% at 11am, and at different times when they are more likely to open it.
Kunle: Ok, very interesting. I wanted to ask you, given the fact that you are running TargetingMantra, you’ve got loads of clients, you have a heavy presence in India and you’re growing your presence out in the US and the UK. What industries, particularly in ecommerce are you seeing benefiting from personalisation? Both from browsing a website through to personalisation through on-site widgets? What industries are you seeing tremendous or significant value off the back of recommending personalisation and recommendations?
Saurabh Nangia: Personalisation is not something which is limited only to a certain category of ecommerce companies or only to a certain monticule. To give you a broad idea, we have clients across whole marketplaces who sell all sorts of items, like travel, jewellery or apparel, baby products, and all of them have benefited from personalisation. On average personalisation drives anywhere between 10-40% of the overall site wide revenue. That being said, companies who benefit the most out of personalisation are companies who generally have a long tail of catalogue. If you are anywhere selling, let’s say 20 items on the website right now, then personalisation won’t help a whole lot. One of the key uses of personalisation is saving your customers from the tyranny of choice. There was a very famous jam experiment conducted in the 1990s by a professor in Columbia where all the consumers were shown a bunch of different jams on the shelf. In one case they were shown six jams on the shelf, in the other one they were shown 30-50 jams on the shelf. They saw that when people were given six options, they were more likely to buy the product as opposed to people who were given 30 or 50. That’s here personalisation comes in to picture.
Kunle: Very interesting because I’m reading a book called “Nudge” at the moment, it’s by a gentleman by the name of Richard Thaler and he’s talking about the concept of being choice architects and how to optimise that experience of being a choice architect and nudge people, not exactly control them, but nudge them towards doing the right thing. Obviously, the skill could be used for good and evil, but I could see this actually used for good in general and to get them to narrow down exactly choice and give them the best available options.
Saurabh Nangia: Exactly. That’s a very key thing. A lot of people get stuck in the paralysis and analysis of things and they don’t end up deciding anything and even if they decide there was a follow up of this jam experiment, I think the professor’s name was Professor Lyengar, and then someone else did a follow up experiment on this that even though when people chose jam from the wide variety as in from the 30 or 50 jams, they still doubted their choice, they had that buyer’s remorse. They were still wondering whether the other jams tasted better than the ones which they bought. So not only do people face more trouble in making the choice, but if they are given a lot of choice, even after making the choice, they aren’t that happy about the choice because they keep on thinking what are the other hundreds of thousands of other options which were out there.
Kunle: So it’s less about a particular vertical in ecommerce and it’s more about the size of your catalogue and you alluded it to being more about long tail. What size of catalogue, from an SKU standpoint, would you start to see personalisation or personalised widgets starting to make an impact?
Saurabh Nangia: Personalisation widgets usually make sense if there are more than 400 SKUs on your website, as in you have more than 400 products. Also, if you are a m-commerce company, if you have a mobile app who is targeting the ecommerce space, then personalisation also makes a lot of sense because you have limited real estate in terms of limited space that your end consumers can see on the phone, so you want to choose which are the most alluring products and show them to them. Where you could have possibly fit 30 items on a category page on a webpage, you can only fit maybe 8 or 6 on a phone. Therefore personalisation becomes a lot more important out there.
Kunle: Besides Amazon and all their companies that you have worked on, what US or UK ecommerce players are you seeing executing personalisation really well?
Saurabh Nangia: One of the key things in personalisation which I also want to stress on is the fact that personalisation should be real time. A lot of these ecommerce companies within the US and within the UK too, they have good personalisation results out there, but they aren’t real time. They won’t take into account my behaviour on the fly. Let’s say I was buying a t-shirt last time I was there on the website because it was summer, but now winter has started so I’m browsing through jackets, it will still remember the fact that I was browsing through t-shirts and will keep on showing me a lot of t-shirt similarities even though I started clicking on jackets. That is a key thing where a bunch of different websites have a huge scope of improvement. That being said, companies which have really good personalisation right now, I haven’t really shopped a lot within the UK, I’ve never lived in the UK, but I would say within the UK Swim Shop are good because they are one of our clients. Also within the US, Etsy had good recommendations, Ebay definitely invested a lot in their personalisation, they have good results. They have another problem which is called deduplication of items. With marketplaces, one thing that becomes important is that there are 10 people who might list an Iphone out there so you cannot show these 10 different Iphones as similar to each other because they are all essentially the same product. So you want to group all these SKUs into a parent SKU where you will have the 10 different varieties of Iphones listed out there. So bigger marketplaces have done that quite well. Ebay is definitely a leader out there. Those are the names that come to my mind at this point.
Kunle: Very good points. What I’m gathering here is TargetingMantra is bringing enterprise level personalisation which typically was technology for marketplaces and really high traffic enterprise websites to the mid-tier.
Saurabh Nangia: Yeah, that’s right.
Kunle: Ok, that’s very interesting. We’re going to move into the final sets of questions segment and I’d like to ask you about your predictions on the future of personalisation in the space of ecommerce?
Saurabh Nangia: I believe personalisation will keep on evolving a lot with time and it will become more and more important. Specifically, over the internet, there are over 3ZB of data being created every year. That’s like 1023. That amount of data is huge, but if there’s so much data out there, personalisation and personalising algorithms will become all the more important because you’ll have to choose what is right to show to what consumer. No one wants to know about all the news articles out there or all the movies out there or all the products out there. Everyone is interested in figuring out things that are of their interest. That’s a key area, not only in ecommerce, but in any of the listing websites too.
Kunle: Machine learning obviously seems to be a dominant theme or technology for this to occur. What are your thoughts on privacy? It’s almost a give or take. You can’t eat your cake and have it here. You want a much more personalised experience, a much more tailored experience while you’re browsing the internet or your favourite internet platforms, however you have to give a bit of knowledge about yourself to get that personalised experience. Where do you see the lines blurring? Or where do you see a straight cut line as to personalisation and privacy?
Saurabh Nangia: So far, our end consumers, speaking from the point of view of end consumers, definitely people don’t like the fact that they are getting tracked somewhere. What I started observing in terms of end consumer pattern is that they are open to the fact that you are making their lives easier whether it’s in the form of movie recommendations on Netflix or product recommendations on Amazon or news recommendations on a news website. They are open to that fact, but they don’t like the fact that they are being shown ads in different places. They don’t like the fact that “I was looking to buy a dress as a surprise for my wife and I am on this news article website and it’s showing me the ad out there to go ahead and buy this dress”, and then my wife is sitting right next to me and she’s like “Are you cheating on me? Why are you looking at this dress?”, so definitely that’s not a good thing out there. I would say that people would still stay open to personalisation on a website but ads, seeing the amount of installations in terms of AdBlock Plus, it will evolve and change a lot. I don’t know how it will change or what will happen in the future, but there are over 50 million AdBlock Plus installations on different browsers out there. That means people don’t want to see a lot of these intrusive ads.
Kunle: I think that it would impact on Google’s revenue, but I’m not quite sure how that’s going to impact on Facebook, especially the inline sponsored post, but that would be interesting to see in terms of its impact on personalisation and advertising. What one marketing channel, given that the future is clearly going to be a personalised one on the internet, what marketing channel would you advise ecommerce entrepreneurs to take seriously and prep themselves up for the age of personalisation?
Saurabh Nangia: Having spent a lot of time in India and the US itself, I would say one of the key areas that people aren’t targeting at this point is mobile notification. Reaching out to your consumers through mobile apps or notifications on the phone. Within India itself, leading ecommerce companies have more than 50% of their sales happening on mobile platform in today’s date. Overall the countrywide average is around 35% at this point. Within the US, right now, it’s at around 8-9%. So it’s increasing, I think it was 5% a year before, but it will take a little while before it gets popular within the US. The companies who realise that mobile is going to be a big channel, that consumers are going to browse and purchase things from, they will be at an advantage if they start targeting their end users on a mobile platform early on.
Kunle: I have a question in regards to mobile notifications. How do websites, especially in the mid-tier, tap into notifications when they don’t have mobile apps?
Saurabh Nangia: Thankfully Chrome has created a thing right now which sends web notifications on your mobile browser even if you don’t have mobile apps at this point, they can send notifications to consumers who have their websites open on the mobile browser. That’s something which just launched a couple of weeks ago. I haven’t seen anyone send any notifications on it.
Kunle: I have. There’s a company called Roost and they were doing iOS notifications for desktop notifications and they started to implement something in Chrome notifications just as you said, which will be quite interesting. I’m hoping Safari will join the party, really, especially with the watch, it would just make more sense if Safari joins this. There are more iOS devices in the US and the UK as compared to Android. I just want to take us back to what you said in terms of one of the data points you look into. You mentioned that fact that you look into time and time is almost like a Z dimension, but what about location and device? Because if someone’s on their tablet and they are on a website at 9pm on a tablet, I would believe they are very likely going to purchase at that point in time, 9 o’clock, evening, relaxed, I’m in my home environment, I’m shopping probably in front of my television and I’m with my Ipad. Is your platform and your technology taking these other factors, in terms of understanding the location of the user, are you able to also figure out, if they’ve put in their delivery address, would you be able to get the coordinates, this is now sounding creepy, and understand their device and give them a much more tailored experience?
Saurabh Nangia: Yes, actually we do that, as creepy as it sounds. We figure out the location of the user. If you’re browsing from let’s say Seattle at this point, I would show you something related to rain, you might be likely to buy rain boots or something like that. If you are shopping from California, I will probably show you something which goes on fashion in the West Coast. If you are shopping from New-York, I will show you dress shoes. If you’re from Hawaii, I’ll show you flip-flops. We do take location into account with regards to what’s the weather and what the overall location behaviour is like. What you are saying is something that we are building at this point which we call contextual recommendations where we try to take into account the context as in the external context of the consumer in which he’s browsing. This could be things like if you’re looking to watch a movie on Friday night and we know that on Friday night you usually tend to watch a movie with a date, we’ll recommend you romantic movies. While if you’re looking to watch a movie on a weekday around Tuesday or Wednesday, we know that you are going out with your friends or something so we’ll recommend a comedy or action movies depending on your taste. That is a different level of contextual recommendation where we are trying to figure out, at this point in time, these users tend to be in so and so mood and they will be spending in this way.
Kunle: The fact that you mention mobile being the future, I think that location is really going to add into context. One other question I had was in regards to could it be beneficial on some occasions for brands to utilise the space for personalisation widgets to drive in a message like an offer for instance or a coupon code rather than other products? So if there are 10 items in a basket and you want to nudge them to checkout, is that an option or do you think other means like an exit pop-up would be more beneficial at that point in time?
Saurabh Nangia: I love your questions. You are mentioning all the points that we guys are offering at this point. To answer that, yes. I’m not a big fan of pop-ups, I usually advise all my clients to stay away from pop-ups because it’s kind of intrusive to the end consumer and they don’t like it. In terms of personalising offers and showing the right coupon code or the right coupon to the users, the way we do it at this point is we do it in the form of a hello bar at the top. So right at the top of the page, we’ll show a very thin bar and that will contain a coupon code or something which is most applicable to you. So if you are in the market to buy sunglasses, we’ll show you a coupon code for sunglasses and in the price range that you are likely to buy at. The other form of intelligence we are putting out there at this point is figuring out how much of a nudge does the end consumer need in order to buy at that point. That’s something which we are still working on, we haven’t launched yet, where we’ll figure out “does it make sense to show this coupon code at this point or should we avoid showing the coupon code?” “Whether he’s most likely to abandon the cart or whether he’s most likely to not make a purchase”.
Kunle: As much as you like my questions, I would slightly disagree with you in regards to exit intent pop-ups, but I would say that the technology available for exit intent pop-ups, I think is a bit one-dimensional and I think with this deep machine learning platform such as TargetingMantra and whatever alternatives there are in the market, there is a lot of opportunity to use all that data and harness it through different technologies, be it the widgets, be it email marketing or be it exit intent, either through the platforms themselves launching products within that ecosystem or providing APIs to third party systems that do it really well. I don’t know what your thought are on that.
Saurabh Nangia: To me, as an end consumer, I haven’t been a huge fan of pop-ups. I do get your point of view that if personalisation is done in the right way, then there won’t be any irrelevant pop-up widgets being shown to the end consumer. So if it’s a relevant thing and it’s coming across the page, then yes, it will be useful to the consumer. But it’s something which we will have to maybe test and figure out what the majority of the people want to see.
Kunle: It is subjective and I guess wisdom of the crowds would nudge us in the right direction. With regards to tools, books and resources, anything, can you make a recommendation to our listeners? They are ambitious online retailers looking to drive growth. Any resource, any tools you think you would suggest that they grab, they learn, they tap into?
Saurabh Nangia: One of the things that comes to my mind at this point is a book I’m reading by Peter Thiel. It’s called Zero to One. That goes around how you should build your company, how you should take your company from Zero to One, how you accelerate your company. Definitely, that’s one of the things which I would recommend people to read. Something which is specific to ecommerce, I remember there was this blog by Chris Anderson on Long Tail. What happens is that the majority of the people are still coming from the mindset of the retail stores and what happens in retail stores is because of the limited shelf space. They used to keep the bestseller items close to the counter or at eye level because people tend to pick those up and buy them right away. Coming from that same mindset a lot of the time ecommerce companies are focussing only on keeping promoting their bestsellers. What they don’t realise is that if they’d rather demand craft across a product, then the long tail for the majority of the products might sum up to be much more than the sale of bestsellers. Chris Anderson wrote this article in the mid-90s I would say or early 2000s. It’s a very famous long tail story by him.
Kunle: I have read articles. It is a fantastic insight into the impact of the long term and Zero to One is amazing. I like it because it embraces monopoly. I’m seeing that success in ecommerce as coming from the more innovative companies that have not done what anybody else has done before and they just take the lead in that segment and then they hold it, which is interesting, thank you. Before you say your goodbye, could you share the best way listeners could get in touch with you? Or prior to that, could you just give on parting piece of advice to online retailers looking to grow their businesses?
Saurabh Nangia: I would say always remember that your consumers are right and you’re supposed to make them happy. Don’t get carried away by sending too many campaigns to your end consumers or just trying to provide them with a bad experience. Always keep a consumer centric approach. Amazon and Zappos are the leaders in terms of consumer centric ecommerce companies. Do keep them as an ideal when you are building your ecommerce company. Do ensure that your customers get the best treatment out there.
Kunle: Fantastic. Do you have any plans of coming into the US and the UK market?
Saurabh Nangia: Yes. We are rolling out across a bunch of different companies in the US and the UK market. You also have an offer in the US market and the bay area in Mountain View. In fact I’ve spent 7-8 months in the US itself. Within the UK, we aren’t setting up an office at this point, but yes, we do keep on travelling to the UK quite often and we do have a customer base out there too.
Kunle: I’ll be sure to share all of your contact details in the show notes. Finally, what’s the best way for listeners who want to get in touch with you?
Saurabh Nangia: I would say the best way to reach me is through email because I am travelling quite often and my phone might be off or I might be in a different time zone so I might not be able to receive your calls. My email is Saurabh@TargetingMantra.com.
Kunle: Fantastic, I’ll share it. It’s been an absolute pleasure having you on the show Saurabh, and thank you for sharing your insights on personalisation and ecommerce and the future.
Saurabh Nangia: It was a pleasure talking with you too, Kunle. Have a great day.
Kunle: Cheers.