2007: The Implicit Web

My partner Brad and I were at lunch with Josh Kopelman a month or so ago and we got to talking about a number of exciting new web services we are seeing and Josh blurted out "web 2.0 is the explicit web and web 3.0 is the implicit web". That's been rattling around in my mind ever since.

It's true that many of the best things about web 2.0 (tagging, posting, digging, embedding) require explicit activities. Those activities provide great value to both the person doing the explicit action and in a social network, great value to many others.

What Josh was talking about when he used the words "implicit web" is myware, something I've been big on for a while now. For those who care about such things, I did not coin the term "myware", it was Seth Goldstein who first coined it.

Enough about jargon, the explicit implicit web is all about the value that will accrue to an Internet user when their every action is tracked, recorded, and used to provide value back to that user. There is also a second order play when that clickstream activity is shared with the user's permission with everyone else.

My favorite example, which I used in that original myware post, is last.fm. I give last.fm the permission to capture all my iTunes listens. I publish that data on my blog (left sidebar) and the data is also published on my last.fm profile page. I can go back and look at the what I listened to most last week, month, year, etc. But more importantly, I can use the data about what I am listening to currently to surface new recommendations via musical neighbors. And because I share all that data with the entire network, my listens inform others in their search for new music.

This concept can be used in almost any activity you do on the web. Take shopping for example. Amazon does a decent job of capturing the activity in our account and recommending new stuff for us. But we use an aggregated account at Amazon, so the Gotham Gal's books, Jessica's music, my gadgets all get lumped into one profile. What if I had a profile of all my web activity and I could express it to any web store the minute I arrived? Amazon could do a much better job of recommending stuff for me. So could eBay. So could Netflix. But take it one step further, connect me to other people who are looking at and buying stuff just like me. Let me see what they are buying and where and why. That's social shopping in my book.

What about paying bills? Now I am sure that this example is going to give some of you the creeps. But what if you could pay all your bills via a free web service that aggregated what you were buying and paying for and recommended ways to get those services for less? What if that service aggregated the buying power of all of its users and negotiated for better prices from vendors where it had significant market power?

You can take this concept and apply it to most any activity you can do on the web. The idea has been around for a while but privacy concerns have held everyone back. But people are starting to get used to profiling themselves and using it to add value to their Internet experience. They are starting to trust certain web services and let them profile them. That change in user behavior is a big deal. And as a result, the implicit web is going to start taking off in 2007.

Comments

[cem sertoglu]

Isn't this exactly what Seth's working on with root.net? I think, once the mainstream wakes up to the value of clickstreams (or bytestreams, since there's much activity, such as listeing to music, that does not really involve surfing), there will be a rush to manipulate (in a good sense) and distill insight from clickstreams. Google is probably the most likely destination for my ultimate clickstream, since so much of it originates from search.

[CoryS]

In marketing, 'look a like' micro-marketing techniques can be valuable, and rely upon a massive amount of data to be relevant. Sure, Amazon isn't good because you're whole family is on one account, but they (and Netflix, etc) have bigger issues to fight with to provide guidance on future interests and personalized choices. If you entered Netflix's preference prediction contest, you'd see how little consumer information they gave to work with (interestingly, since the contest, they have been building it up).

There's one thing that itches at me on "myware", RSS: variety v. consistency. As a consumer, if I am very consistent, then I am predictable. If I like The Shins, I'll likely enjoy The Decemberists - and the fact that you listen to them allows a system to offer smart recommendations on them to me. But what if I never hear AC/DC because I'm tagged as Emo (and have friends that are as well)? Variety in being exposed to new things has value and is hard to build into a logic system without feeling like noise. Good counter example to myware - I continue to read the NY Times, primarily because I wouldn't naturally have a feed reader about a story about Uganda, but I'll read it and enjoy learning from it.

[Ric]

We are learning to trust a FEW web services, but how much better would this be if we didn't have to trust them all, but had a decent identity management structure that we controlled - roll on Infocard/Higgins/etc.

The "implicit" web would explode.

[Sumit Chachra]

Stumbleupon anyone? I love the fact that I see stuff my friends like, faster than del.icio.us or digg can ever be! All I have to do (most of the times) is just give a thumbs up or down. Sweet!

[adam]

you should check out wesabe.com - the personal finance site... its pretty cool...they attempt to solve the below problem from your credit card statements

"What about paying bills? Now I am sure that this example is going to give some of you the creeps. But what if you could pay all your bills via a free web service that aggregated what you were buying and paying for and recommended ways to get those services for less?"

[Scott Loftesness]

Like adam, I'm struck by the implicit nature of what Jason and Marc are doing at Wesabe.com. The notion of most all of the payments that you make (not just at Amazon) being among the best recommendations you could share seems particularly interesting.

[Dustin]

Internet wide? That is quite an app. I think it has already begun but will take more then next year to begin to become mainstream.

[Andy Gadiel]

Indeed, it seems that Personalization, Recommendation and Discovery are becoming all entangled as simply different sides of the same coin...loving watching it blossom, what's beyond that?

[Greg Linden]

Building the implicit web -- recommendations for content, advertising, and information based on your past behavior -- is what Findory (http://findory.com) has been doing since 2004.

I guess Findory was three years early to the game, but better early than late.

[Peter Lindberg]

I think "explicit" in the fourth paragraph should be "implicit", right?

[NST]

I love the concept. A minor note, is there any concern of the influence of observation? I know that in marketing this is always a huge concern that skews the collected results. In other words, will people change their behavior now that they know it's being tracked?

[lawrence coburn]

Yeah, the implicit stuff is the coolest of all. It allows us to create additional value for ourselves by doing what we were doing anyway, e.g. listening to music, shopping, browsing, etc. And I don't think it's limited to just web based activities - for example, GPS should enable a nice data stream of places you go while offline.

[Colin Henderson]

Great post - implicit web lines up nicely with the tools Doc Searls is speaking of, that he happens to call VRM (Vendor Relationship Management). The concept is to place the power in the hands of the customer, and let them in essence, RFP to vendors, who have to come back with quotes. Your concept takes this further my having suggestions offered.

[fred wilson fan]

fwiw...fortune magazine ran an interesting piece on this phenomenon in november.

http://money.cnn.com/magazines/fortune/fortune_archive/2006/11/27/8394347/

[Martin Cleaver]

I did a similar blog post just recently about organizational transparency.

http://martin.cleaver.org/blog/2006/11/14/the-transparent-enterprise-implicit-communication/

Regards, M.

[Jake Lockley]

Flip the model. Put it on my desktop and set it up so I am analyzing the data and generating algorithms that are exchanged with the server.

It disintermiediates the power and control so the consumer does not have to be exploited.

The problem with this model is business is only interested in exploiting consumers instead of giving them their own black box that people such as advertisers would have to pay to have access to.

Imagine if every cable box came with a record of what we watched and no one was allowed to use it but the owner, and the owner was empowered to sell their information only to who they want to have access to it.

[Raj]

I certainly agree that 2007 is the year of the implicit web.

Don't know if you saw the rumor a few days ago, but Last.fm is apparently in play for $450mm by Viacom. Personally, I think Apple should acquire Last.fm and bake the social features right into the iPod itself in addition to the iTunes plug-in.

http://www.dealbreaker.com/2007/02/the_rumor_mill_lastfm_being_pi.php

[Jud Valeski]

Implicit is indeed the future. Rather than understanding and building implicit value propositions on a service by service level we, me.dium.com, are building it for the network at large.

http://me.dium.com

[jer979]

Agreed! I just sent Reed Hastings an email w/some 2.0 Netflix ideas:

posted here:
http://jer979.blogspot.com/2007/03/netflix-in-20-world.html

[km4]

For 'Implict Web' to happen you need sophisticated machine learning and information agent technology for extraction, user profiling and modeling, record linkage, sentiment classification etc. on large numbers of web sites, without requiring the monitoring logic to be re-specified for each site.

[Michael Lacy]

The first step in building the implicit web is aggregating explicit web content and normalizing it so that it can be remixed and redistributed.

http://neurokinetikz.com