Are online retail subscription models a fad?

At this year’s Cyberposium conference I attended a panel that questioned if subscription models were a fad. One of the panelists was the former CMO of Shoedazzle, a subscription-based e-commerce company that each month sends members a selection of shoes, jewelry and handbags from which they select one item to purchase for $39.95 / month. However, I was surprised to hear that Shoedazzle recently scrapped its subscription model and is now focusing on providing its 10 million members with personalized recommendations and is expanding into other verticals (lingerie and dresses) (1).

So, are subscription models a fad? We all know that companies, and especially VCs, love their recurring revenue stream but the question is does the end customer enjoy and value a continuous influx of non-discretionary products? The founder of Shoedazzle admitted that this model didn’t quite align with its customers’ actual purchasing behavior and preferences as “they’d like to buy two or three times in one month and then maybe not for a few more months” (2).

When the subscription model doesn’t work…

Beyond the recurring revenue stream and strong inventory management provided by subscription models, one of the reasons why companies use this approach is to build long-term relationships with their customers (3). However, I think that applying the subscription model to certain categories, namely non-discretionary goods, can actually destroy customer lifetime value. Discretionary and non-discretionary items are purchased differently. If I am buying dog food (discretionary) for my pet, I know that I will likely need to buy it again at some future point and will not contemplate this decision, so I will likely sign up for a subscription. However, when I am shopping for jewelry or shoes (non-discretionary), I enjoy the serendipitous aspect of the shopping experience – I like being delighted with a selection of items and knowing that I can either purchase or walk away from an item.

Therefore, I disagree with Shoedazzle’s founder’s argument that “subscription models work best (…) where the product is a necessity or when it’s an absolute passion” (3). The subscription model compromises the serendipitous nature that people enjoy when shopping for non-commodity items and can thus undermine its relationship with customers. Instead of being delighted, shoppers may “reach a saturation point” , no longer enjoy the experience and get frustrated by having to “stuff another pair of cheap platform pumps into their closets” which ultimately leads to high churn rates (1). As Sucharita Mulpuru, an e-commerce analyst at Forrester, succinctly noted, “Shoppers might find it harder to justify a recurring credit-card charge for colorful suede booties” (4).

When the subscription model works….

To contradict the founder of Shoedazzle’s view on when subscription models work, one of the panelists at the Cyberposium noted that the approach works best when the delivery of the product is also long-term. Therefore, to answer my initial question, I think subscription models are a fad when implemented for faddish items – shoes, make-up, jewelry, etc. – and should only be employed for items of necessity that have clear, recurring value for customers.  As Alex Zhardanovsky, co-founder of PetFlow, a montly subscription service for pet food, explained why his business works, “’Dogs never stop needing to eat” (3).

From subscriptions to curation…

Shoedazzle is now focusing on curating items for its members (1). I think many fashion/beauty e-commerce subscription models should also shift towards curation since I believe is the best way to reach customers as it taps into the serendipitous nature of shopping for faddish, non-discretionary items. Furthermore, the companies can still deliver customer lifetime value devoid of a recurring revenue stream by using curation to enhance user engagement and increase customer loyalty/stickiness to a site.

 

Sources

  1. http://nymag.com/thecut/2012/03/kim-kardashian-shoedazzle-monthly-model-lingerie.html
  2. http://www.fastcompany.com/1826421/shoedazzle-ditches-monthly-subscriptions-boutique-style-pampering
  3. http://www.nytimes.com/2012/03/08/business/smallbusiness/selling-online-products-by-subscription-is-all-the-rage.html
  4. http://www.nytimes.com/2012/11/26/technology/building-start-ups-using-stars-ties-to-fans.html


 


read more

A glance at any news source will reveal an article or blog post on the tension between e-commerce and brick-and-mortar retail. Yes, it’s true that e-commerce is quickly encroaching on the historically offline retailers.  E-commerce sales in the U.S. are expected to reach $224 billion by the end of 2012, up 15% since 2011. [1] Yes, brick-and-mortar retailers still have an advantage, especially in high fashion and apparel more broadly. Customers that want to feel the material and touch the product, enter and browse stores for inspiration, or need an item immediately will still opt for the offline experience. And yes, retailers that master both online and offline will presumably fare the best.

 However, I believe that there are a set of issues that both online ecommerce and offline brick-and-mortar retailers alike are facing – and depending on how you cope with each will ultimately determine how and whether you come out on the other end.

 Increased retail commoditization puts pressure on online and offline retailers

 Retailers and brands have seen their products commoditized over the years. This first happened offline, but worsened as online retailers came into the mix, elevating the importance of price. For example, Intermix has always competed with similarly priced fashion chains, like Scoop, and department stores, like Saks and Bloomingdales, but as brands have opened their own boutiques, increased competition has come from the very designers that Intermix sells. [2] Of course, this commoditization is only exacerbated by the emergence of sites such as Net-a-Porter and Shopbop. While some retailers have tried to combat this issue by partnering with brands to develop exclusive lines, at the end of the day, brands often don’t want to be limited by a single retailer or take the risk of a failed partnership.

 Different shopping habits of millennials forces retailers to innovate

 There are currently 11.8 million millennials age 18-30 living in U.S. households with annual incomes exceeding $100,000. Not only is this group larger than many expect, but it will take over as the largest generational segment in the luxury consumer market around 2018-2020. [3] Furthermore, this group is looking for a unique in-store experience and online social validation. Retailers have spent enormous amounts of energy and money to better understand and cater to this emerging demographic. Nordstrom’s investment in HauteLook and Bonobos, partnership with TopShop, and significant IT capex ($100 million in 2011) on rolling out in-store touchscreens and mobile devices for store associates highlight its intent to be relevant to the millennial shopper and break the mold of being “your mother’s department store.” [4] Online retailers spend marketing dollars to align with influencers within the segment.

 Personalization & curation is difficult

 It’s a common misconception that ecommerce retailers have personalization and curation figured out. In fact, many are still at the initial stages of trying to integrate social sentiment tools, analytics and big data insights in order to better recommend products to customers in a scientific way. [5] Many online retailers tend to have algorithms with cookie-cutter product recommendations based on aggregate site visitor behavior rather than individual customer tastes and preferences.

 At the physical store-level, retailers rely on sales associates to deliver a personalized experience to customers; however, this commonly misses the mark for a variety of reasons. First, sales associates are under-utilized in stores – typically spending much of their time on low-value tasks. Second, given the traditionally high turnover at the shop floor level, retailers are hesitant to spend too much on training and therefore sales associates often miss on easy opportunities to up-sell and cross-sell. Finally, sales associates are not given the necessary data or information about a customer’s purchasing history, tastes, and brand preferences to improve customer service and efficiency.

 Human element is key to differentiation

 At the end of the day, the retailer that brings the best overall experience wins. Part of this is getting mobile POS incorporated into the stores, which most brick-and-mortar stores are already in the process of figuring out. More importantly, though, is how sales associates – online or offline – play into the sales process. I am in complete agreement with Hil Davis’ assertion that personalization and curation will still require a human element that is the sales associate or stylist (with the help of technology and data). [6] A well-trained sales associate with the right data-informed tools will be able to provide the differentiated experience relevant to millennials and the broader tech-savvy audience. This will lead to increased customer loyalty and consequently increased sales. Neiman Marcus has found that customers who shop with the same associate three times spend almost 10 times more than those who go to a random sales clerk. [7] It’s an art and a science – a delicate balance of using data and relying on the personal relationships that develop between sales associates and customers.

 

[1] http://www.emarketer.com/newsroom/index.php/apparel-drives-retail-ecommerce-sales-growth/ 

 [2] http://www.crainsnewyork.com/article/20101017/SUB/310179968

 [3] http://www.forbes.com/sites/larissafaw/2012/10/02/meet-the-millennial-1-young-rich-and-redefining-luxury/

 [4] http://blog.shop.org/2012/09/12/customer-service-is-changing-and-so-is-nordstrom/

 [5] http://allthingsd.com/20120604/e-commerce-accelerating-due-to-personalization-pinterest-and-ipad/

 [6] http://www.forbes.com/sites/ciocentral/2012/10/25/the-future-of-e-commerce-bridging-the-onlineoffline-gap/

 [7] http://www.retailwire.com/discussion/15855/app-lets-neiman-marcus-know-when-best-customers-arrive


read more

Paperless Post – “Vanity, Definitely My Favorite Sin”

Last summer, I read an interesting blog post titled Building successful consumer apps in The Age of the Humblebrag (1)(2). In his post, William Peng describes the three main building blocks of building a successful consumer app:

1) Playing to inherent human nature / behavior / psychology

2) Driving interaction in an app

3) Productizing around the behavior – gamification / network effects

Peng does a great job going through some concrete examples of how Facebook, Twitter, Instagram and Foursquare have done exactly this. I wanted to apply this same logic to another company that is seeing consumer traction: Paperless Post. Paperless Post is an online invitations company that is driven by their focus on design. Senders are able to pick and choose between different designs and customize for their own unique aesthetic. Paperless Post currently has a $10 million revenue run rate (3) – how did they do it?

Playing to inherent human nature / behavior / psychology

Co-founder Alexa Hirshfeld visited campus this past weekend and quoted Al Pacino from The Devil’s Advocate in describing her users: “Vanity, definitely my favorite sin.” (4) In invitations, people like to show off how capable they are in designing something beautiful. This is why Paperless Post has stolen customers from services like Evite, which allows very generic customization. The average number of recipients for each Paperless Post is roughly forty, a great captive audience that has a vested interest in your invitation.

Driving interaction in an app

With Paperless Post, you are able to create distinctive cards through a combination of great content and customization. And, even though there are an infinite number of combinations, each one is a reflection of its sender. Each card is distinct and reminiscent of fine stationary, thus giving the impression that the sender spent time designing a beautiful invitation just for this event, when in fact, it was more than likely just a few small tweaks to an existing template.

Productizing around the behavior – gamification / network effects

The initial sender may be Paperless Post’s first customer, but each recipient also becomes a customer as soon as they receive their invitation. Because the recipient needs to interact with the invitation to RSVP, they are then exposed to the platform. As long as their initial impression of the product was positive, it is a great built-in marketing tool to spread the Paperless Post name. Additionally, Paperless Post utilizes an interesting pricing plan, which includes coins instead of real dollars. And, although you can purchase coins (average $0.08 per coin), it is also possible to earn coins by connecting with Paperless Post on social media or by inviting your friends. With these coins, you can add small details such as envelope liners or premium content furthering your ability to customize and impress your invitees.

It seems clear to me that both aesthetic design and market design has played a critical role in the success of Paperless Post. Now, with 1.5 million registered users, 30 million plus invites sent and over $12 million in funding (4), I believe Paperless Post has developed an interesting platform for the future of “printed” digital correspondence.

 

(1) http://williampeng.com/post/24689752914/building-successful-consumer-apps-in-the-age-of-the

(2) http://www.urbandictionary.com/define.php?term=humblebrag

(3) http://online.wsj.com/article/SB10000872396390443624204578060573322740436.html

(4) http://www.youtube.com/watch?v=3M68wcB6L0s

(5) http://www.paperlesspost.com/info/about/about_us/about


read more

It’s a common scenario.  The weekend is here and you’re looking to try a new restaurant.  The problem is that you don’t know which restaurant to try, and you certainly don’t want to pick the wrong one.  Why we attach so much buyer’s remorse to this indisputably low risk decision is a good question, yet it can’t be denied that we do.

Driven by the quest for a guilt free restaurant choice, you head for today’s natural starting point – a quick review of Yelp.  Maybe you also run a quick Google search for “best restaurants in MyCity,” but that search likely brings you back to Yelp anyways.  So, you dig in.

What do you find?  Well, a search for restaurants in Cambridge returns over 40 restaurants with at least a 4-star rating.  Not much distinction there, so you begin to click through a few.  Lots of “I love it” and “best ever’s” are scattered across the occasional “I hate it.”  Still, tough to glean much from that.  Absent the restaurant name on top, the reviews are all essentially interchangeable and are all either incredibly positive or incredibly negative.  Maybe if you focus on Sam L., you’ll find better advice.  Sam L. kind of looks like you in his ¼ inch x ¼ inch photo, and he may even be wearing the same shirt that you own.  He must have similar dining tastes.  Unfortunately though, Sam L. seems to base his star rating more on the perceived attitude the hostess gave him rather than the actual quality of the meal.  No help there.  (click on the “Real Actors Read Yelp” link below for more brilliance from Yelp reviews)

Eventually, you realize that you’re not getting what you’re looking for and leave frustrated.

This frustration is something that I’ve experienced often, and I believe it’s experienced by many as online reviews begin to inundate consumer web-based research.  Without claiming to have the full solution, I’d suggest that there are 3 areas of improvement that can provide better reviews to consumers and possibly even propel some interesting innovation in the online review space.  I’ll continue to use the restaurant example, but I believe that these improvement areas are present in a number of other online review services (e.g. TripAdvisor/travel review)…

Making reviews more relevant to me (and to you):  Relevance is an essential dimension to the online experience, and it continues to be addressed across numerous fronts (e.g. search, advertising, deals).  This is an area where you see the big players taking the leading role.  From Facebook likes to Google filtering search results based on your web history, a significant amount of attention (and dollars) has been spent on trying to make the web more personally meaningful to the individual.

That said, there have also been much smaller players that have found positive traction in building recommendation engines of higher relevance.  Stamped (recently bought out by Yahoo in a talent acquisition) sought to allow people to ‘stamp’ things that they liked and share them across their network of friends in a more convenient way than Facebook likes.

However, current efforts within online reviews continue to be primarily focused around associating relevance to things preferred by your circle of friends.  While this is a great first step, I believe that it fails to provide adequate results.  First, your friends have divergent interests, and in the case of the restaurant example, your dining preferences will almost certainly better match a complete stranger (found in a massive dataset much larger than you network of friends) than your best buddy.  The key is figuring who that complete stranger perfect match is.  I believe that big data is a far better solution for an online reviews 2.0 world where preferences can be matched to truly statistically significant comparison sets and deliver much more meaningful reviews and recommendations.

Identifying the proper target and offering objectivity: For the far majority of Amazon reviews, identifying the target of the rating is easy.  It’s the product being highlighted.  However, when rating more multi-dimensional experiences (such as dining or travel), it’s much more difficult to discern what is actually being reviewed and rated.  Sam L. above based his restaurant rating on an interaction with the hostess.  Some may view this as helpful, while others might prefer a more objective review of the quality of the food.  The inherent subjectivity of which dimensions matter most can quickly erode the integrity of the overall rating system.

Just this past week, Yelp released an update to their platform hoping to address this concern more directly.  Yelp is adding menu pages to the restaurants that rate specific dishes while also providing user submitted photos of dishes.  It’s a positive step towards better associating objective ratings with more cleanly defined targets.

Providing more nuanced rankings:  When 40+ restaurants in Cambridge are rated as 4-stars or better, it can be argued that none of those restaurants are actually rated at all.  The consumer is simply unable to make a reasonable distinction between them based on the ratings alone.  Much of this clustering occurs from the polarizing catalyst that encourages consumers to submit reviews in the first place.  Feedback is typically generated from an overwhelming experience, either positive or negative.  The more subtle critiques and comments often don’t inspire the effort to submit a review at all.

One possible way to accomplish more nuanced ratings would be to better segment like-restaurants and force rank the restaurants within each segment across a more normal distribution.  This may generate more enlightening ratings by highlighting the best and the worst pricey restaurants in Cambridge, for example.  Current results tend to skew all ratings for pricey restaurants towards the highest stars, but a force rank would allow the ratings to show the best of the best, and the worst of the best in this instance.

 

References:

Real Actors Read Yelp
http://www.youtube.com/watch?v=QEdXhH97Z7E

Yahoo Buys Mobile-App Maker Stamped
http://online.wsj.com/article/SB10001424052970204076204578078842424741154.html?mod=googlenews_wsj

Ex-Googlers Launch iPhone App for Tapping Into Friends’ Reviews
http://mashable.com/2011/11/25/stamped-iphone-recommendation-app/

Yelp adds menus, makes them mouthwatering (or revolting) with food photos
http://venturebeat.com/2012/10/30/Yelp-menus/


read more

Big Data in Consumer Internet – Why I Do Not Believe In the Creep Factor

On Tuesday evening, I attended a presentation on “Big Data” by Scott Friend, a managing director at Bain Ventures. Scott founded and sold a retail analytics company called ProfitLogic. In addition to talking about his entrepreneurial experience, he focused on Big Data in consumer Internet. In this context, Big Data refers to all information consumers leave behind on the Internet when they post Tweets, upload files on Dropbox, or check in on Foursquare. [1] Naturally, the discussion led to a comment from the audience on the “creep factor” involved in Big Data.

In my view, this is the most important question in data analytics today. Consumers are sharing increasingly more information online and do not have the means to identify how this information is used by advertisers and companies trying to capture consumers’ money and attention. As is usually the case in situations with information asymmetries, trust is a necessary factor driving consumers’ online behavior. To continue to build consumers’ trust and encourage them to opt in to share personal information, companies should focus on the three strategies below.

Educate the Consumer

I have been an advocate Big Data in predictive analytics because of the way it has transformed the Internet into a discovery tool for me. I find it valuable to see personal and relevant online advertisements rather than random, untargeted ones.  I continue to discover new artists through Pandora. Knowing what my friends pin to their Pinterest boards enables me to buy better birthday gifts. I eagerly embraced the “Quantified Self” movement and even track and record my sleep pattern online through an app called Sleep101. [2] So far, I have opted in to connect to websites through Facebook every time the option exists and I liberally use the Like button to contribute to the 2.7 billion likes Facebook collects daily. [3] Predictive analytics has enabled me to discover things I did not know I wanted and enabled me to make better purchasing decision. This is the main reason why I continue to opt-in. The value Big Data contributes to my daily life prevails over the creep factor.

Protect Personal Information

I know that along with my Likes, I am making my willingness to pay heard by companies and advertisers. As a result, I may have become a victim of price discrimination. What I perceive to be a discovery tool may have limited my freedom to “browse” the Internet. My social security number could have been accessed and my credit card number could have been stolen.

Until I have concrete evidence that any of the above scenarios actually occurred, I will continue to hope that companies I interact with online on a regular basis, my bank, Dropbox and Amazon, are responsible with the personal information they store on my behalf. I believe that these companies have my interest in mind and invest in the necessary infrastructure to protect my online identity. Only by prioritizing online safety, can Internet companies increase the trust factor as well as the extent and depth of information they can collect from customers.

Be Transparent

Scott Friend gave listeners a business idea at the end of his talk. He suggested that we start working on a tool to help consumers uncover all the ways information they opt in to provide has been used for or against them. Until such a solution can be developed, consumers have to rely on Internet companies to be transparent. By asking consumers to opt in, making “Do Not Track” feature more prominent and enabling consumers to skip ads online companies can build trust. [4]

Consequences of not meeting the above standards are too risky to bear for companies. Encouraging the creep factor and inviting restrictive government regulation to protect the consumers are the first two that come to mind.

Notes:

[1] http://venturebeat.com/2012/02/22/the-7-creep-factors-of-online-behavioral-advertising/

[2] http://techcrunch.com/2011/11/21/the-big-data-bottleneck-in-the-consumer-web/ 

[3] http://techcrunch.com/2012/08/22/how-big-is-facebooks-data-2-5-billion-pieces-of-content-and-500-terabytes-ingested-every-day/

[4] http://www.forbes.com/sites/gyro/2012/09/04/the-end-of-big-data/


read more