Deals & More

In recent pieces for VentureBeat, I’ve been proposing ideas for new companies. This next one was proposed by PayPal writer and fashionista April Van Scherpe.

Style is like pornography: you know it when you see it, but few can do it well. It’s a skill of creativity and pattern recognition, knowing the working combinations of color, line, and form. Many of us would like to have it but don’t have the time or inclination and resign ourselves to fashion disasters. In the words of Gilda Radner, we base our fashion decisions on what doesn’t itch.

This is a problem, not just because clothes reportedly make the man, but because it makes shopping hard. If we hated shopping for groceries like many of us do for clothing, we’d starve like supermodels. But in problems of pattern recognition, technology can help.

What many of us could use is a Pandora for fashion. Pandora learns your tastes in music and suggests new songs and bands; a similar approach for fashion could learn who you are –- your size, personality, preferences, and purpose -– to suggest wardrobes that make you feel good within and without.

Let’s call this idea Stylize. Users would sign up, enter age, sex, and measurements, then take a visual fashion test to tease out preferences. Do you prefer this white polo or that black tee? Are you an hourglass or a pear? Mary Anne or Ginger?

Stylize could then present personalized items and wardrobes from retail partner sites. Like Netflix, users could rate, share, and buy items, refining Stylize’s knowledge of each user and general trends. Stylize could also sharpen its suggestions with collaborative filtering –- “users like you liked these items” — and basic fashion rules: use vertical lines to look taller, favor dark colors if you’re pale, only wear plaid in Switzerland.

New imaging technologies might enable advanced features. Like.com’s visual search could suggest items that resemble your favorites. Facial recognition may let users upload pictures to calculate measurements and skin tone.

Stylize could monetize well in the nearly trillion dollar clothing market. Middle class Americans spend $1,680 on clothing per year; generating just one checkout in this high-margin market could yield tens to hundreds of dollars in commissions. Targeted advertising would also be lucrative and even welcome, as fashion consumers browse ads as much as content. Stylize search may also be embeddable on clothing sites to generate suggestions specific to their inventory.

I didn’t find much competition in this space. Shopping search engines such as Like.com focus on aggregation or pattern recognition. Fashion social networks like Fashionspace focus on connecting users. Content sites like Glam focus on photos and articles. Social shopping sites like StyleFeeder have some interesting suggestion features but broadly include other categories like electronics and food. Fashion is a large enough niche to warrant a focused discovery engine.

A risk of Stylize is that though online clothing is a multi-billion dollar market growing fast even in a recession, many shoppers still hesitate to buy clothes on the web without trying them on. That may be especially true if a suggestion site is inching people beyond their comfort zone. Eric Seidel, co-founder of defunct fashion site ShoutFit, says sizing is a real problem for women. “Women use the web to buy things that don’t have a size or have well-understood sizes like purses, shoes, and accessories. Only men can actually buy clothing on the web.” Clothing sites like Zappos offer money-back guarantees, but the hassle of returns is still a problem. At the very least, Stylize could spark demand and drive more shoppers into retail stores to try and buy.

Overall, discovery engines in long tail markets can be powerful sources of serendipity. By applying proven technology to a subtle art, Stylize might make the web a more attractive place.

What do you think?

Update: The Like.com team just launched Covet.com, a virtual personal shopper and stylist. It has a few of the above features like a visual fashion test and personalized recommendations. Hopefully future versions will add more fashion logic and tips to help the majority of us that need them.

Mark Goldenson is half Swiss so he can make fun of plaid. He is starting an innovative venture in health care. To submit an idea for the What’s Next series, email Mark at mjgold3@gmail.com. Selected ideas will receive attribution.

Next Story: 6 essentials for turning sales page visitors into clients
Previous Story: Apple sees 11 million Safari 4 downloads in three days