Thoughts on Merchandising, Social Recommendations, and Content Curation

I was following the twitter of Oyster CEO Eric Stromberg’s twitter when I came across this article on curated content vs. recommendation engines. Oyster is a new startup in that hopes to become Netflix for books. As you can imagine, given my background at two different startups that aspired to be “Netflix for Textbooks” you can image I was quite interested and had some great subsequent discussions on the right mix between different types of recommendations and this area of “product discovery.” 
 
I looked through Fab every day for a month and found Fab pretty useless. Perhaps it was the wrong curator for me. I didn’t really like their taste and I felt like most items would just be apartment clutter. I think I might be up for trying Trunk Club if I had the money, but I’m skeptical that anyone else is going to get my “fit” with clothes. When I was redesigning the apartment I spent hundreds of hours on Pinterest but too much of it is dreaming and not enough is real pictures of real apartments executing stuff that is within reach. Same with Houzz. So discovery for me has remained very much a manual process.
The key insight from that article was that users want to understand the rationale behind an algorithmic recommendation – the “why.” The other good thought was that customers want a point-of-view – hence the rise of curation.
 
Branding 101: When I go to Macys.com – I judge the store on how it meets my needs and whether I like the clothes. I judge it critically and rationally. But when go to Jcrew.com – they have a clear point-of-view built into the entire experience – the advertising, the homepage, the stylized emails, the fit, the color palette- and if I don’t like something – like Denim Jackets – I give J.Crew the benefit of the doubt and surmise that they are reaching toward a specific look, a specific point-of-view, that simply doesn’t align with my own. It’s like the difference between disagreeing with someone on the facts – which I can’t get past – vs. agreeing to disagree on our opinions, and still remaining friends. 
 
Perhaps that was rambling. A more concise attempt:
 
1) Insight from article: users want meta-information on why they are being served up this recommendation.
2) What we’ve always known from branding 101: we can and should package recommendations as a lifestyle idea or a point of view. We must always aspire to build an emotional connection with our customers – this happens on top of a point-of-view.
3) Example: Rather than Macy’s which tries to be everything-to-everyone with one brand, one set of advertising, one store experience, but little sections of clothes from different sub-brands, success will be had by smaller brands that instantly evoke a particular point-of-view.
4) How this impacts curation / social recommendation: We’ve taken the “infinite shelf space” idea to it’s natural limit. Every store now stocks every item online to the point where they start to lose their identity. Amazon.com and Macys.com, for example, have no identity. The natural result now is trying to make the shelves seem personalized to you, so that you continue to shop at this giant big box online store. 
5) An alternate path: Perhaps the next big wave of ecommerce sites will be online boutiques with highly stylized brands and curated content with Fulfilled-by-Amazon (or eBay or Google) infrastructure on the backend. Perhaps curated content and recommendations is a reaction of online-big-box retailers to the need for boutique experiences and that – as front-end web-design gets commoditized and backend drop-shipment infrastructure also gets commoditized, small boutique stores on top of big platforms will become the norm. 
Advertisements

2 thoughts on “Thoughts on Merchandising, Social Recommendations, and Content Curation

  1. Recommendation engines work well for amazon, netflix, facebook, spotify etc because they have a ton of data to extract meaningful signals. By ‘ton’ of data, I mean they have both ‘items’ (content such as movies, movies, books, photos, friends) and users. With the combination of both, recommendation engines such as collaborative filtering can be very powerful in surfacing content that your friends listened to based on similarity scores between you, them, and the items.

    Adding the ‘editor’, personally-curated information is actually another way for these sites to capture signals about you. For example, when you first come to fab.com, them sending curated content and your interaction (esp. the purchase decision) will actually enhance that curated content down the line.

    So basically, the human – touch combined with your interaction with content can actually enhance and improve the knowledge these engines have about you. Eventually, this can imply that the human touch might become less necessary and scalable with enough relevant data.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s