Attract, Convert and Retain members: Advisory calls and Dating Site Startup Bootcamp & Consulting Services Sign Up Today

Browser History As The New Personal Ad

lumi logo

People on dating sites are liars. Every last one of them. Old photos, lying about weight, misrepresenting themselves to an astounding degree and generally making themselves out to be quite the catch.

Today I had a long talk with Likebright (I’m an advisor) and one of the many topics we discussed was mining and matching on people’s “social exhaust”: Facebook posts, music tastes, books read, photos uploaded, products Liked, etc.

The conversation with Nick Soman reminded me that years ago (2006) there was a company called Root Markets. The founders believed, like Doc Searls, that your attention has literal value. Around that time The Attention Trust had a plugin that tracked your browser history, uploading it to a server with the goal of anonymously running your history through an algorithm that would result in better targeted advertising.

Unfortunately, as recently as six years ago, people were not ready to let all their personal information hang out in the wind. Suffice to say, it all imploded. How times have changed. Fast forward to today, and millions of people have been divorced, fired and expelled due to refusing to adhere to the adage “Think before you post.”

One company that has greatly benefitted from users freely uploading their data without user backlash is Last.fm. Their Audio Scrobbler technology uploads the last songs you played through your media player (eg. iTunes), and uses that data to better tailor the songs that are played for you through Last.fm.

TechCrunch recently said that Felix Miller and Martin Stiksel, two of the co-founders of the music discovery and streaming site Last.fm, went into a kind of startup hibernation two years after selling their early-moving, prescient startup to CBS in 2007 for $280 million.

Turns out that the Last.fm Founders have just launched Lumi, a site that uses your browsing history to help you discover things on the web.

Back to TechCrunch: Taking your browsing history, and then pushing it through its algorithms, Lumi suggests other content and other sites that will be relevant to you, presented in a series of Pinterest/Fancy-style, image-led boxes that populate infinitely down the screen.

Companies have been using and abusing browser cookies for 15 years. Every day you pick up 100 new cookies, not to mention the insidious and deeply-ingrained Flash cookies. But nobody has every figured out how to make a sustainable business based on accessing that last bit of personal data on your hard drive that web servers can get their hands on, your browser history file. Lumi wants to change that, and if their past success is any indication, within a year or two we’ll all let social, dating and shopping sites access our browser history.

Let’s temporarily engage in a collective hallucination and pretend that Lumi is not going to sell our personal browsing habits to the highest bidder, or the FBI for that matter.

What is more personal than your browser history? Mine would probably make a seasoned psychologist ball up in the corner and cry, and I’m not that weird.

Lumi Co-founder Martin Stiksel: A lot of businesses are trying to hook into your “social graph” to make content suggestions, but — in the name of personalization and in defense of privacy — Lumi is taking the exact opposite approach. Lumi never knows what you’ve browsed, and neither do your friends (shudder).

But your browser history knows all. You know that joke about the person you would have run to your house and clean it up if you die so your parents discover your embarrassing proclivities? Better ask them to delete your browser cache as well.

It’s all there: Amazon books, p0rn, illegal downloads, sex toy orders, political sites, various group affiliations, cat videos, blog comments, Internet trolling, etc. Given the tragedy last week, it could even know your political stance and what you feel about gun control based on your browser history.

You see where this is going, right? You can say whatever you want on your dating profile, but its never good enough. You are terrible at telling me what you’re all about, and you’re even worse at trying to explain the type of person you want to be with.

I’ve talked for years about all of the amazing data you spew onto the Interwebs each and every day. And now Lumi is poised to mine that data to provide recommendations. Totally reminds me of Stumbleupon, which used to be awesome but seems to have taken some wrong turns over the last few years as it removed much-loved features and went all Pinterest on users.

While writing this, I joined Lumi, and here is a mid-page snapshot of the recommendations they sent to me 20 seconds after they sucked up my browser history. I’m fairly amazed at what didn’t show up, given I was doing some client research for, well lets just call it a casual dating site, ok?

Lumi browser history recommendation engine

Imagine if the results were people’s profiles or Facebook pages, now that’s what I call people discovery. Lumi could grab your browser history and quickly identify gender, age, sexual preference, relationship status, psychological attributes such as clingy or independent, Type-A or follower, depressed or homicidal maniac, spiritual or Westboro Baptist Church member, active outdoor enthusiast or couch potato, artisan or entrepreneur, the ways they could categorize you is practically limitless.

The difficult part (besides convincing people there is value in installing a browser extension and uploading your deepest darkest secrets), is identifying how the sites your visit correlate with your personality type, tastes and preferences. I can imagine that it would take a large room of people to sit down and categorize sites based oncertains “buckets” of people. ESPN.com and Boston.com/sports, definitely sports. Foodspotting and Epicurious, definitely a foodie. But what about the nuances of certain sites and what if I’m miscategorized? I could go to Fling.com, then a church website, then Al Jazeera and Reddit (which has thousands of interest categories). Who do they think I am then? Pandora is the case study here. They had many musicologists categorizing individual songs in order to provide the spot-on playlists that make the service to popular. Interesting thing about Pandora, I heard the CEO speak at MIT years ago and he talked about how categorizing classical music was straighforward, but they had to to be extra-diligent when it came to figuring out all of the sub-genres of rap.

This “Who am I, really?” conundrum reminds me of this clip, “My Tivo Thinks I’m gay”.

To date, dating sites have not done a great job leveraging the social graph and your Likes. I expect some sites to nail this in coming years, but it’s going to take a long time for them to figure out how to get it right. Meanwhile, Match is creating incredibly-detailed composites of users and matching according to type, not on a user-per-user basis, which is far too computationally  intensive, not to mention mind-bogglingly complicated. So is eHarmony and POF and OkCupid for that matter.

Regardless if you are going to hate on them for the results, top dating sites are doing incredible work with their matching algorithms, but it feels like they are hitting a wall now, which means matching effectiveness could stagnate while the number of users increases, which cannot be good news for singles.

And now we have Lumi, which could potentially deliver a people-matching system that could go head-to-head with any dating site. Or at least deliver a personal ad that’s orders of magnitude more comprehensive and authentic that what dating sites offer today. I’m sure they will go for the easy money first, but why wouldn’t they at least take a look at the dating industry and see if there is a fit there?

Dating Industry Advisory Services

Are you a dating site startup looking to make a splash or an established player in the marketplace looking for answers? Online Dating Insider Advisory Services focus on product development, revenue generation, site traffic, site reviews and *much* more. Contact us to learn more.

Comments

  1. Match is creating incredibly-detailed composites of users and matching according to type, not on a user-per-user basis.

    You seriously expect me to believe that major dating sites aren’t using user-per-user matching? composites is so 2006, and you only do that if you lack data.

  2. I’m not so sure that single guys would want to know that a single woman has bookmarked hundreds of wedding dresses and engagement rings. I also don’t think men want women to know about all the “adult videos” they watch. Some things are just better left to discover with couples privately. However, Lumi can become a good “adult content” discovery engine since that is probably what most men look at frequently or have bookmarked.

  3. Not quite agree with the author, you can log in to our website, through the signature, we dating compatibility.

  4. i agree people either struggle to describe themselves accurately or cannot be bothered to in their profiles. anyway to automate that could be a nice value add but i doubt many people will want their browser history fiddled with

  5. I agree that people lie on profiles, sometimes unintentionally. I still think the best way to meet someone is face to face. I trust my intuition over a computer match any day.