Richard MacManus’ post 5 Problems of Recommender Systems.

In a Guide to Recommender Systems, a blog post written by ReadWriteWeb founder and editor, Richard MacManus, 4 types of recommender systems are identified:

Personalized recommendation – recommend things based on the individual’s past behavior

Social recommendation – recommend things based on the past behavior of similar users

Item recommendation – recommend things based on the item itself

A combination of the three approaches above

Each of these approaches has distinct advantages and disadvantages and the post outlines how both Google and Amazon use a combination of all three to drive their recommendation engines.

As I look at the current state of online fashion it’s pretty clear no one has figured out how to leverage these recommender systems in fashion affiliate ecommerce.  There are simply too many styles, they have too short of product lives and the diversity of consumer wants and needs are too segmented.  The complexity in making fashion recommendations makes creating an algorthimic solution extremely difficult.  Past behavior is not a good tool because the trends are always changing.  Item recommendations don’t work because there are simply too many product attributes in fashion and each attribute (think fit, price, color, style, fabric, brand, etc) has a different level of importance at different times for the same consumer.  Social recommendations, though challenging, hold the most promise.  So far, though, social shopping has fallen flat.

The problem, I think, is trust and scale.  In fashion, unlike say electronics or travel, you aren’t going to trust just anyone’s advice.  You want to know who is making the fashion recommendations and they need to have some credibility.  To overcome this trust issue, many fashion recommendation sites focus on “friend recommendations”.  The problem here is this model doesn’t scale.

By example, consider TripAdvisor, if you are looking for resorts in Bermuda, you will see the top-ranked 4-star resorts at the top of your search. The top-ranked resorts have credibility as “being the best” in part because of the many, many lower ranked resorts further down the list. The consumer knows that all resorts have been carefully considered and these are the best. Friend recommendations don’t pass this scale test. You just don’t have confidence that all the styles out there have been discovered and considered. Therefore, the recommendation has much less weight and credibility.

At StyleHop we believe the way to get around the trust-scale tradeoff in fashion social recommendations is to let users explicitly identify their “fashion peer group” when they are performing product searches.  So in addition to the typical product attribute refinements you can make when shopping for clothes online, we want to allow you to filter your results based on the 5-star rankings (think netflix) of folks you trust to help you shop.  For a nineteen year old going to college in the fall for her freshman year, she may want to see the styles highly ranked by women at the University of Michigan.  For another woman moving to New York City for her first job after graduation, she may want to see the top ranked wear-to-work clothes of 20-something Manhattan women.

By allowing consumers to identify their fashion peers and adjust who their peers are for each search, the user is in control and the results are broad and deep. Most importantly, you are giving the shopper validation -before she makes the purchase- that what she is buying will be acceptable among her peer group.  This goes well beyond surfacing top styles the consumer might like and acts as a powerful social reinforcement of the purchase decision.

Will eBay incorporate social shopping into its fold?

Link to another Vator.tv interview with Erik Stuart.  Erik clearly understands the dynamics at play in social shopping and has pretty good advice for ThisNext.

Charlene Li, co-author of “Groundswell” and a thought leader on social and emerging technologies made some predictions for 2009 that included one for Social Shopping:

Shopping Goes Social. After a devastating holiday season, retailers will eagerly seek a way to improve results other than driving demand with deeper discounts. One option they will investigate will be how to insert people and social connections into the buying process, illuminating and influencing for the first time the Black Hole Of Consideration. As they lick their wounds in the first half of 2009, retailers will watch from the sidelines as media companies implement open social technologies like Facebook Connect and the Open Social Platform. But as the holiday season launches early after Labor Day, shoppers will find options to see what friends are recommending, buying and rating integrated into the shopping experience.

Charlene is spot on.  There are going to be lots of different takes on how to make this a reality.  At StyleHop we are building the first fashion affiliate e-commerce engine that allows you to order your product search results based on the rankings of your self-identified fashion peers. Take three women living in Des Moines. One may want to see the top designer jeans as ranked by other women in her neighborhood, another by her friends, and another by women in the East Village of NYC. Each may have similar demographic characteristics and even similar initial “clicks” but, by identifying their unique fashion peers, they each get highly specific lists of styles that work for them.

This is so much better than black box behavioral analytics which consistently give back poor recommendations in fashion. Peer review has credibility and gives a woman shopping online the ability to shop quickly and confidently knowing that, when she buys an item, the people she wants to look good in front of have already pre-approved her purchase.

Good post about the impact of the economic downturn on startup entrepreneurs:

VCs to Entrepreneurs: Outlook for Software Startups Is As Good—or Bad—As Ever

Bigelow’s report from a VC panel discussion supports what I’ve been saying for a while now….it has always been really hard for startups to raise capital.  Just because you are struggling mightily fundraising doesn’t mean it’s because of the collapse in the financial markets.  It could just be your business/idea isn’t that great.

Smart investors will continue to invest in startups…..the biggest risks for tech startups have actually gone down significantly.  Here’s my risk assessment for StyleHop in light of the economic crisis:

Risk #1:  Competitive Risk – Lower.  It is much less likely that someone will come in behind us and steal our market away.

Risk #2:  People Risk – Lower.  It is much more likely we will be able to attract and retain the right talent.

Risk #3:  Financial Risk – Higher.  It will be harder to raise cash (but that risk drops after investment).

Risk #4:  Market Risk – Neutral.  Its is likely that fashion and apparel companies will continue to seek out our demand forecasting tool because it is a high ROI investment with a very fast return.

Manageable, right?  Would love comments from other entrepreneurs.

An excerpt from Gavin O’Malley’s post Lines Between Media Channels Increasingly Blurred

From 2006 to 2008, the share of U.S. consumers using shopping Web sites doubled from 17% to 35%, according to a study released Monday by public relations firm Ketchum and the University of Southern California Annenberg Strategic Public Relations Center.

More revealing, perhaps, is that 44% of those online shopping consumers reported reading consumer reviews and comments found on the sites, the study found.

When I started StyleHop last year, one of my friends in acadamia shared Duncan Watts seminal work that essentially debunked the concept that marketers can predict hits by looking to the “influentials” in their product categories.  Watts revisits this topic in Sunday’s Washington Post article: So You Can’t Pick the Hits.  Neither Can Anyone Else:

Why is predicting so difficult?  Well, for lots of reasons, but two fundamental ones stand out. First, individuals are much harder to predict than they seem, not because people are infinitely complex, but because how we are apt to behave depends on subtle details of the situation.

The most interesting part of Watt’s work was his collaboration with Matthew Salganik and Peter Dodds to explore how certain songs become hits.  Here is what he found:

When participants knew what others liked, the popular songs became more popular and the unpopular songs less popular than when people made their choices independently. More surprisingly, however, we found that which particular songs become the most popular also became more unpredictable — in some cases social influence caused luck and randomness to overtake intrinsic appeal as the main factors driving success.

In the fashion industry, there has always been a strong belief that some individuals have near divine powers to predict what will be hot.  What we find over time, however, is that this almost never plays out unless the predictor has become a truetastemaker or brand unto themselves.  In other words, some individuals like Mickey Drexler, Ralph Lauran and Anna Wintour become so broadly followed that they in fact do have enormous outsized influence.

.

If Watt’s theories are correct, though, the magic fairy dust that helped these three rise to tastemaker status had less to do with their predictive fashion abilities and probably a lot more to do with hard work combined with a lot of luck.  All those buyers and merchants trying to become the next Mickey Drexler may be wasting their time .  The ability of individual merchants to consistently guess in advance of the season what styles will sell in which quantities will never rise much above a mediocre distribution.  Okay, I’m saying it:  Merchants simply can’t see or intuit the subtle nuance that affects consumer decision making in advance nor can they see how the trends will evolve and be accepted as the social influence evolves.

.

This is what makes StyleHop’s model so compelling.  Leveraging thousands of real live targeted consumer’s feedback we can improve upon the individual’s merchant’s forecasting ability leveraging the wisdom of the crowd.  Think of it this way, if the ability to predict fashion is a loser’s game (historical markdowns and high variation in sales performance in fashion support this) then wouldn’t it be better to just ask a large number of the target consumer which items they would like?  By the way, consumer products companies like Procter & Gamble have been doing this market research for decades prior to new product launches.

.

Longer term, with traction in the consumer marketplace, StyleHop’s model will be reinforced through the social influence Watts discovered.  Think of StyleHop like the leaderboard in Watt’s music study – directly influencing consumer behavior by showing consumer’s which items other folks have already said are great styles.  Imagine hangtags, in-store signage (or better yet geo-location iphone updates) highlighting the top-ranked StyleHop styles while you shop.  This peer validation, in my view, is more relevant and targeted than any fashion editorial in the magazines today.  Women would love to know that other real women like a particular style before they buy it.

.

Would love comments if you have them.