Personal recommendations have always been a part of ecommerce, but there has been little innovation since Amazon introduced retail and product personalization 10 years ago. But with the increasing mountains of data at digital retailers’ fingertips, ecommerce is about to get even more personal.
The fact is that right now there is little iteration from personalized ecommerce beyond what is taking place on Amazon. So you’ll see suggestions of what other shoppers who bought a certain item also purchased, or recommendations to similar items to what you have purchased, but there is a whole world of social data, and even more-in-depth purchase data that can be mined by retailers to help increase sales.
Kleiner Perkins partner Aileen Lee agrees with me, “In the future, the best retail sites will know you much better and show you things that are much more relevant.” Lee has helped lead investments in a number of e-commerce companies including Offermatic, One Kings Lane, Plum District, Rent the Runway, and Trendyol and held operating roles at The Gap and North Face.
“We are just at the beginning of a revolution of e-commerce, and existing retailers are going to have to get better at personalizing the experience for consumers,” Lee says.
“Personalization was really important in enabling Amazon to differentiate itself and grow in past ten years,” David Selinger, CEO and co-founder of RichRelevance. Selinger also was Amazon’s Manager, Consumer Behavior Research and helped build some of the site’s personalization features a number of years ago. “Personalization will be the differentiating factor in e-commerce and digital commerce going forward, especially for multichannel retailers and new entrants online.”
Amazon and Netflix represent the first wave of personalization. I believe that we are going to enter into the next wave of a more personalized e-commerce experience as retailers and e-commerce sites move towards mining data to improve sales and conversions.
It’s highly likely that you have helped boost Amazon and Netflix’s conversion rates on movies, books, or other products thanks to personalized suggestions of items that you may like based on your previous purchase data, other consumers’ purchase history and more. In fact, it’s so seamlessly baked into the user experience for both companies, that I tend to not even notice how impactful personalization is on what I purchase.
That’s not to say that Amazon is the only retailer experimenting with personalization. eBay has also been personalizing the marketplace experience with recommendations of similarly viewed or bought items for some time, and is looking to expand personalization efforts with PayPal. And with the recent acquisition of Hunch, we know eBay is going to ramp up data mining.
Recently, I started to receive emails from Gilt Groupe that suggested similar earring to like those those I had added to my wait-list on the e-commerce site. The company also sends personalized email notifications on sales that are tailored to each customer. Gilt, who declined to comment for this piece, seems to realize that personalization is going to be a key product driver for the site in the future. And brick and mortar retailers like Saks Fifth Avenue, and many others are also starting to jump on the personalized email bandwagon.
The best way begin understanding the opportunity of personalization in the future is to realize the immense challenge that retailers face when approaching personalization. As DJ Patil, Data Scientist in Residence at Greylock Partners, explains, “When you go to Nordstrom you have a shopping assistant helps direct you, basically says ‘I’m here to help, what do you need and here’s where to find this.’ No online retailer has quite nailed that,” he explains.
For most retailers, the toughest hurdle is to have enough data on an individual to actually help personalize the experience. For the majority of buyers who purchase from a specific site once every few months, or even less frequently, a retailer may have no real sense of direction on how to present similar products.
Getting these data points is the biggest challenge that retailers face. But retailers do have significant data for the small amount of regular, routine customers for an e-commerce site, including clicks, purchase history, shopping cart information, shares and Likes, and more. Retailers face challenges on how to store and organize this data, and then turn this into personal recommendations
And data comes in various forms. There’s implicit data (which is gained from your everyday actions on a retailer’s site) and explicit data (which you offer to sites via surveys or quizzes). While retailers are doing more with the implicit data (i.e. reminding you when you left items in your shopping cart); no one has yet mastered the art of capturing that precious explicit data.
Google’s Boutiques.com tried its hand at this, as a search engine and fashion site which allowed users to receive personalized clothes and accessory recommendations based on preferences and actions. But Google subsequently shut the portal down last September.
Asking for users to fill out surveys of what they love or like perhaps isn’t the ticket to drawing explicit data, such as brands you love, colors, styles and more. As Patil explains, retailers who ask for this information need to present this as more of a conversation as opposed to replicating the feel of a doctor’s appointment where you are filling out your life history via forms.
Getting these signals from consumers is very difficult from a UI and user experience stand point, he says. His advice to retailers is to find a way to replicate how a store owner or shop keeper would engage you in a conversation when walking into a store and looking for something open-ended, such as a birthday gift. One way to do this is to present a personalized item suggestion but ask the consumer (in a Pandora-like fashion) if the recommendation sucks and how they can make the shopper’s life better “People want to help the system and love to correct things,” Patil says.
And similar to Pandora, people become more invested in a platform that knows their preferences and will be more likely to return.
There’s also the issue of finding the balance between providing serendipity in terms of discovery and personalization. Retailers still want their sites to be this Pandora’s box of discovering items, literally, but personalization can cut down on this discovery process. So retailers need to both anticipate what consumers may want to purchase on the site but also provide items that consumers will be able to feel like they ‘discovered’ on the site.
Patil draws an interesting comparison with how grocery stores have been able to structure their layouts to provide serendipity and useful discovery. “When you go to the supermarket, the stores know you are definitely going to milk aisle, so they often put it in the back of the store, so you can find serendipitous stuff on the way. Online retailers need to replicate that on e-commerce sites.” In the end, the goal is to be able to deliver personalization without being predictable.
At a macro level, retailers also face challenges in finding talent to sort this big data. The difference between doing data personalization well are radical shifts financially for retailers, Selinger explains. The engineers who are able to parse these massive amounts of data are hard to come by, and expensive.
Social data (i.e. the Facebook Likes of products, what products people are recommending on Facebook or Twitter) is going to be a big part of personalization for retailers in the future. Already plenty of retailers are using Facebook social plugins and Connect integrations to leverage Facebook data to show visitors what friends bought or shared, what products relate to their Likes, and which friends they might want to invite.
The problem with this data is that much of it is unstructured, and there is really no one who has effectively nailed social personalization in the commerce arena the way Amazon was able to do with data from purchase behavior. Blippy attempted to socialize purchases, but it failed. Amazon also allows you to connect to Facebook to access your friends’ Likes and recommendations but I find this UI to be clunky, and not very useful.
Selinger thinks that mining social data for ecommerce may lose steam before it takes off, drawing the comparison to email. “In 2007, if you were to walk into VC’s office with an idea about ecommerce and email, you would have been sent out the door,” Selinger says. But he explains that while there is an inherent enterprise value in this social data, it will take a long time to take off, similar to the way it took awhile for personalized email and commerce models to enter the market. “When someone figures out how to do it and do it well, it will grow really quickly,” he maintains.
The challenge for retailers is making sense of the Facebook news feed — i.e. streamlining recommendations, attaching brands and tags to this data and then serving this to shoppers in a useful, personalized format. Basically, your social network can become your Consumer Reports.
The challenge for the data mining community, explains Patil, is actually figuring out the intent in much of the unstructured data that is posted about retail products and brands on Facebook. And it’s important to keep in mind that some of this data from Facebook users is private.
This past week, Facebook partnered up with sixty different startups to add their “stories” to Facebook Timeline, through apps that span different verticals from Food, Fashion to Travel. Part of this involved adding new actions (in addition to ‘like’) to Timeline story options. That includes the verbs ‘bought’ and ‘want.’
There is tremendous potential in developers and retailers being able to mine this data from ‘boughts’ and wants’ as opposed to the open-ended ‘like.’ You can see details on what social shopping mall Payvment is doing with the new protocols here, but basically, the ability to add these targeted buttons could be game-changing for social discovery in e-commerce.
Echoing Lee’s thoughts, Patil is confident that there will be a new wave of personalization and e-commerce. But without data, there is no personalization. So consumers both on Facebook as well as on retail sites will have to be more willing to give up key data like purchase history, Likes and other social actions, and even location in order to get a more personalized shopping experience on retail sites.
The key will be getting consumers to understand that more data will improve their shopping experience, and making the choice of opting-in a no brainer.
Selinger agrees that privacy is going to be an important issue in the next tranche of personalization innovation. “Now more than ever, consumers are more cognizant of what’s happening with their data,” he says. But what retailers have in the favor is a strong foundation of privacy practices, because these companies have had to protect consumer financial and credit card data for time. Selinger believes that retailers will be very thoughtful about privacy and data sharing going forward.
Perhaps sites like Blippy and Boutiques.com were ahead of their time when it came to consumers willingly handing over the keys to their shopping and payment preferences. I envision a day when there will be an app that reads all of your purchase history via your email account and then serves you recommendations based on this data. There are some companies who are already parsing through receipts in your inbox to organize purchases, so why not take this a step further.
And these personalization strategies that are being adopted by retailers are already trickling down to other kinds of sites beyond e-commerce as well. In the same way that ecommerce sites are trying to maximize sales and profits with this data, content sites are also using social and other data to add relevance to their platforms.
So shoppers, be prepared to give up your data. In the coming year, we’re going to see many more retail sites ramping up data-driven discovery. And e-commerce sites who aren’t thinking about how to mine social and other forms of data are probably going to be left in the dust by the Amazons and Netflix’s of the next wave of personalization.