Of the thousands of entries, Gavin Potter, a retired management consultant with no formal machine-learning training managed to rise to number 17 on the Netflix Prize Leaderboard. Potter writes, "The competition has trained several hundred, if not more, people how to properly implement machine learning algorithms on a real world, large scale dataset...This is, almost undoubtedly, the world's largest set of data on repeated decision making and it's ripe for analysis. The analysis may not win the competition, but it sure should provide some insights into the way that humans make decisions."

The public knowledge acquired from the process of producing these algorithms will not only affect Netflix's ability to suggest customer desires across its movie titles, but it will also form a baseline for other business systems. In addition to streaming entertainment providers, companies like Amazon and Pandora have worked hard to produce the best possible predictive technologies. If these company can tap into our unique tastes, they can suggest products and services we didn't even know we wanted. So a 10% improvement on recommendations can equate to a lucrative sales increase.

A second shorter term Netflix prize is expected in the near future. According to the New York Times' Steve Lohr, the Netflix Prize 2 will be concerned with "taste profiles" based on demographic and behavioral data.

Your favorite current affairs news blog, which couldn't survive on Viagra ads, is now charging subscriptions. Your e-tail site of choice keeps recommending country music, which you outgrew years ago. And your default social network's constant entreaties for donations finally annoy you so much that you do the unthinkable: switch to MySpace (at least it is supported by News Corp's old-media money).

This is too much. So, you pick up a copy of Portfolio magazine and browse the ads for financial products, reassured that at least this medium knows how to target its audience.

Anyway, this extreme scenario plainly won't happen, even if Congress has its way and regulates behavioral targeting in some fashion. The advertising lobby is far too strong to let policing efforts slingshot us back to Web 1.0.

But the battle continues. We'll likely see some form of Congressional hearings, with consumer advocacy groups and watchdog legislators like Congressman Ed Markey (D-MA) trying to sort out whether behavioral targeting companies are really just stealth spy networks. The ability of ad networks, social networking sites, ISPs, and Web publishers to capture user data without disclosing that they're doing so will remain under the microscope. Who knows? We may one day see a national "Do Not Target" registry.

Let's hope not.

The main argument for behavioral targeting is obvious: profitable CPM advertising, which allows advertisers to better reach buyers, supports free sites. Users, in turn, inhabit a less cluttered ad environment with far more relevant advertising.

But it goes deeper than that. Targeting, when it works well, creates a seamless relationship between you, a site's content and the ads on it. Consider magazines like Architectural Digest, Vogue and Travel + Leisure. If you're interested in design, fashion, or travel, you might ending up scanning the ads as much as the articles. Now, consider that the Web can deliver those ads in a far more segmented way (if you're traveling to Turin, you're likely interested in Piedmont wines).

In October 2007, Microsoft paid $240 million to power Facebook's ads, driven by private user data. While jaws dropped over the $15 billion valuation, Facebook has used the capital to quietly build an advertising environment that shows what behavioral targeting can do. You may be too busy browsing your friends' baby pictures to notice the ads (and that is a downside: social networks are not purchasing environments), but as an exercise, go to your personal page and click the "More ads" link. If you're a regular user, chances are it's getting ever closer to nailing who you are.

Rather than showing a succession of Netflix ads, Facebook now more closely resembles a neighborhood store, run by a knowledgeable proprietor who makes smart suggestions about what you might want. Yes, that creepy spy is really just Mr. Hooper. Seriously. The dozens of clerks at your local Barnes & Noble can't possibly remember what you bought a month ago. But Mr. Hooper? He knows.

Let's face it: Facebook's targeting has gotten much better because it uses your personal data without explicitly telling you. Were you given the opportunity to opt out, you probably would. It's almost reflexive for us now: "No, thanks. I don't have time to read the fine print and consider the implications."

So when the behavioralists are trotted before the House Subcommittee on Communications, Technology, and the Internet, their first order of business will be the basics: that, fundamentally, the Web is behavioral targeting. LastFM introduces you to people who like what you listen to; DoubleClick tracks your browsing history to display relevant ads; Amazon tells you a new Dylan album is available because, well, that's what it does best. Digital advertising is killing the magazine industry not because it is a cheap boorish alternative, but because it can simply embed a little Java tag in your cookie and follow you until it (or you) expires.

That is not a bad thing.

Yes, targeting has its downsides:

• Small sample sizes. Remember that gardening book you bought for your grandmother last year? Amazon sure does.
• Psychographic characteristics cannot be determined by browsing alone. Clicking on a Rolex ad doesn't mean you can afford a Mercedes; gawking at eye candy does not a luxury buyer make.
• Perhaps most critically, every behavioral targeting company has to toss out user data on a regular basis or else invest in endless storage. AudienceScience, for example, can track 2 billion events per day but purges data every 90 days.)

Congress should in fact be looking for ways to foster better targeting, not limit it. The more an ad network knows about you, the more likely it will be able to serve an relevant ad and improve your user experience. iTunes makes better recommendations than most for a simple reason: more often than not, it operates from a larger collection of data. You may frequent only one or two sites powered by 24/7 Real Media's Open AdStream, but if you've got an iPod, you've spent time building an iTunes database of, say, a few thousand songs, which directly feeds iTunes' Genius.

To see if Genius is better constructed than other recommendation platforms, create an iTunes database of only 10 songs (which could very well be the number of books you purchased from Amazon in the last year). Then ask Genius to recommend a playlist. You will almost certainly not get that eerie "How does it know?" feeling.

What's the solution? The digital lobby certainly has to do a better job of educating Congress on the benefits of targeting. And self-regulation is critical: NebuAd's clumsy efforts to track browsing histories directly through ISPs hit every wrong note possible for privacy advocates.

The solution, though, may be more technological than regulatory. Suppose ad networks, social networking sites, e-tailers and search engines were to share IP data. The overall sample size would shoot through the roof, and both advertisers and users would benefit from ad placement that is as spot on as iTunes' Genius. Some kind of universal cookie, perhaps, could let e-tailers notify ad networks that, despite your apsirational browsing of ArchiteturalDigest.com, you still shop at Ikea, and so that $40,000 George Nakashima coffee table really is not a good fit.

Suppose further that consumers could control this data. For example, once you returned from your European vacation, you could simply delete your old browsing history on travel, eliminating any more ads for hotels in the Mediterranean.

"Yeah right," you say, "Google would never share its users' search histories with e-tailers, much less other ad networks."

Well, it doesn't have to. Your browser already contains that information. Having this URL...

http://www.google.com/search?q=kindle

... means that you Googled "Kindle." And having this one...

http://www.amazon.com/Free-ebook/dp/B002DYJR4G

... means that you searched Amazon for the Kindle version of Chris Anderson's book Free. Solely by parsing URLs, your browser could even tell whether your Amazon session ended with a checkout.

For certain advertisers, that is valuable information. And if that personal data were kept safe, you probably wouldn't mind if targeting companies used it to display an ad for Free on ad networks.

But let's say you did mind. Let's say you preferred random Viagra and Disney ads. Your browser could resolve this, too. It already stores thousands of cookies for you—in fact, it stores all of the non-logged-in data for sites across the Web. Mostly, this is the Web pages you visit, user names and passwords, and other form data, and it's not easy to discern what cookies are storing what. But your ability to monitor and manage your cookies could be enhanced with a simple extension.

If this extension could communicate with third-party cookie-based systems through a standard protocol, you could set simple controls for how much and what kind of data to broadcast. "Okay, re-targeting bot. If you want to drop a Java tag on me and follow me around the Web, here are the ground rules."

The rise and fall of NebuAd illustrates a rather simple truth of behavioral targeting: the data is yours, not theirs. Yes, ISPs could track your browsing history universally, unlike individual ad networks, social networking sites and search engines. And shared data would enhance the user experience. But because the data is yours, controlling it should be in your hands, not theirs.

Which brings us to the real treasure trove: data you've entered on password-protected sites. Why is Facebook, for example, the only beneficiary of the personal data you enter on that social network? It knows you're an iPhone user, New York resident, Ivy Leaguer, travel lover, and Texas Hold 'Em player for one simple reason: you told it.

If communicating that data through a browser-based protocol, at your discretion, turns Viagra ads into iPhone app ads across the Internet, wouldn't that be something? We all understand the nature of free: someone has to pay. So, why not turn your browser into Mr. Hooper?

Guest author: Chris Kincade is co-founder of DesignBuggy.

Two year old wine social network Snooth announced today that it is now powering wine recommendations for the 25,000 editor tested recipes on Conde Nast's food site Epicurious. Snooth says this is just the first of a number of big sites that its custom algorithm will power recommendations on. That cassoulet? Snooth suggests you serve a Montevina Terra d'Oro Syrah 2002 ($15) with it. Nice.

Wine with food has got to be one of the most familiar kinds of recommendations offline, but the online recommendation technology industry is a fast growing one. The belief is that quality recommendations will serve as searches you never knew you wanted to perform - helping users navigate from one logical option to another, possibly making more purchases as a result and hopefully being better served by the websites they visit.

A food site with good wine recommendations sounds pretty tasty to me. Snooth says its recommendations are based on ingredients, cuisine and cooking method.

Here's how it works. First, the keywords are parsed out of a recipe, then they are run through an extensive food dictionary and a long decision tree is then followed. Is it a soup, is it a salad, what is the primary taste? Beef and nuts tastes mostly like beef; beef and liver tastes mostly like liver. How the ingredients are to be prepared is determined by their proximity to preparation words in the recipe. Recipes with expensive ingredients will see more expensive wine recommendations, inexpensive ingredients (lobster vs. shrimp, for example) will yield less expensive wine suggestions. Goodbye old one-liners about "if you're eating chicken!"

Nibbledish, Cookstr, Chow? All cool recipe sites but no wine recommendations, much less very sophisticated ones. It's easy to see how recommendations can provide a competitive advantage in a niche like this.

If that team withstands the month long period of scrutiny that begins now, it will not only mean fame and (some) fortune for them and a big boost for Netflix - it could signal a key turning point for recommendation technology on the web.

In January of this year, we took an in-depth at the Netflix Prize, asking if 2009 could be the year that the goal would be met. In that post we discussed a New York Times profile of the contest as well, where we learned that the company's existing recommendation engine called Cinematch is credited with driving 60% of Netflix's rentals. That system is especially good at predicting "long tail" movies, older more obscure titles that are less well known but make up 70% of what Netflix customers pick. Improvements in Cinematch's effectiveness plateaued in 2006 and the move to offer a big cash prize for outside innovators has captured the imagination of thousands of engineers and their fans.

How Does it Work?

How do you judge improvements on recommendations? Netflix provides contest participants with huge piles of anonymous data about what movies certain customers rated highly, then the teams built algorithms to predict which movies other customer profiles would rate highly based on past patterns. BellKor's Pragmatic Chaos says it can now guess what people will like with a 10% improvement over Cinematch's success rate.

That gets difficult when movies like Napoleon Dynamite, which some people loved and other people hated, get thrown into the mix. It's nearly impossible to predict whether a person will like films like that.

Most of the predictive recommendation systems entered in the Netflix Prize are reported to be quite similar - so we asked in January whether it was going to take a radical breakthrough to top 10% instead of just continued iteration.

That breakthrough may have come when the four teams put their heads together, or it may have been an iterative victory. Time and science will tell.

Some people believe that recommendation as a technology has the potential to be even bigger than search. In our favorite article on the subject, written eight-teen months ago now, Dr. Rick Hangartner, Chief Scientist at recommendation engine Strands, puts it like this:

In the near term, search engines will increasingly incorporate simple recommender technologies to handle approximate queries (e.g. "You asked for this, and based on similar queries/behavior by others, you might be looking for this."). But in the long term, the recommender industry will be larger, and recommender technologies will be more pervasive than the search industry and search technology as we know it. [Because there will be recommendation going on all over the web.]

Now you can Stumble outside of the shadow of the mega-corporate overlords at eBay - two years after Stumble founders Garrett Camp and Geoff Smith cashed out and handed their baby over to the ecommerce giant, they've come back with a team of investors and bought StumbleUpon back from eBay. It's pretty exciting.

Why is this such a big deal? For three reasons.

StumbleUpon is Back in the Game

Earlier this year we wrote about StumbleUpon hitting 7 million registered users, which was at the time 50% higher than most estimates of media darling Twitter's users. Twitter's numbers have shot through the roof this year, leaving StumbleUpon in the dust.

Some people told us that Twitter and Stumble are like Apples and Oranges. Both, though, are tools of discovery, sharing and social connection. They are different in some ways but similar in others. There's a special kind of magic to StumbleUpon, though, and we're very glad to see its founders back in the driver's seat. A web where Twitter and Facebook are the only social discovery games in town is a poorer web.

Some Suspicion is Mitigated

Dave Winer wrote last week that the reason everyone wants to know how Twitter is going to make money is because eventually they'll come up with something - and we as users might not like it.

There was always a similar kind of feeling with regards to StumbleUpon at eBay. Just what was eBay going to do with Stumble and was it going to be creepy? Did you really want to share your serendipitous discoveries, your likes and your dislikes, with the arguably dysfunctional eBay?

Somehow Camp and Smith, as founders, feel more trustworthy. Who knows though - last time they thought it would be a good idea to sell the company to eBay! Many innovators from the beginning of Web 2.0 quickly grew skeptical after their startups were acquired, hopefully they'll be more careful next time.

Innovation Coming Again

Remember what the last wildly innovative thing StumbleUpon did was? Adding the awesome video section of the site, we'd argue, six months before the eBay acquisition. Now that Camp and Smith are in charge again, we hope to see the same spirit of awesomeness that built Stumble in charge again. The team already says they've got "several new products and features to be released in the upcoming months."

We've got high hopes for this deal, just like we do for the rumored founder buy-back of Skype from eBay.

Founders are where the innovative juice is, often, and if a big conglomerate that buys those startups can't figure out how to make effective use of these incredibly disruptive technologies - then they should get out of the way.

StumbleUpon is one of the handful of services launched on the consumer web in the last few years that can truly be said to have significant cognitive benefits for the people who use it. Give yourself an evening to peruse and explore the web using Stumble, if you haven't. It makes the brain feel good and it's a great way to learn - through play and personalization. It's a whole lot of fun. It's not something that needed to be turned into a way to jump from eBay listing to eBay listing. Stumble is smart, it's useful and we're very happy to see it independent again.

Ginx appears to be a people and news recommendation service built out of a Twitter publishing tool and a URL shortener. We think that sounds great, those lightweight technologies hold huge stores of valuable data. The company has raised about $2 million in funding so it's the real deal, not a fly by night operation. Check out a screenshot below and our full coverage of Omidyar's new gig over on Jobwire, our blog covering new hires in tech and new media.

Update: Omidyar pinged us on Twitter this afternoon to point us to a very short press release confirming that Ginx "is a Twitter client that aims to provide Twitter users with a rich experience for sharing and discussing links. Ginx was created to enable people to become more actively engaged in the news and topics they care about."

For the rest of what we've been able to find out about the service so far, please see our post on Jobwire.

We're impressed by the offering and excited to see what will come of it.

Idiomag has been pulling in user attention data since early this year. The idea is that users can enter their username from sites like Last.fm, Pandora, iLike, Strands or imeem and Idiomag grabs their public listening profile from those sites. Those past interests are used to recommend playlists of music and videos that a user would probably like. Idiomag then brings in semantically indexed articles from syndication partners about those artists, and photos from around the web from live concerts, etc. The coolest part is that the pages of these magazines are created dynamically so that the background color behind the text, for example, works well with whatever the images are on the page.

Content is changed day by day according to which artists the user has liked or disliked. It's all pretty fascinating.

Now all of that content is available to be placed on 3rd party websites through API calls. Music social networks MOG and TheFilter are already using the API to offer up-to-date blog posts about artists on their profile pages.

Data is available in xml, json, rss, apml, xspf and foaf.

What's Coming Next

Idiomag says it's taking its semantic content discovery and social recommendation technologies into new verticals and onto new platforms next. Mobile and HTML interfaces will extend beyond the current Flash magazine format. Film and gaming content pipelines are already built and will soon be integrated along with music content. Gossip content is next. (Who does web content but leaves out gossip? Only people uninterested in monetization or overly concerned about the dignity of the human spirit.)

Idiomag has been around for several years but just two months ago built out a new semantic indexing technology that parses the full text of articles to determine their relevance. The company says it's seeing 80% accuracy in non-music verticals and will reach parity soon.

The advertising on Idiomag is personalized along with the content and the company says it's seeing 8 to 10% click through rates. That's covering 40% of their very slow burn rate (they also have some seed funding) and they hope to soon start cutting some B2B deals.

The Mashups to Come

What will we see built with this new API? We're excited to find out. We expect to see 3rd party sites in all kinds of verticals, but first in music, pulling in bundles of personalized multimedia content that they will then add further value to on their sites. It's exciting to think about. Maybe you, dear reader, will be so inspired. Let us know what you build!

1. Me. Access books, music, movies, restaurants, wine etc. that you liked and commented on via the browser. All your favorites are always synched up and right there when you need them.

2. Friends. When you're looking for social recommendations on the go, you can tap into an intelligent, aggregate list of things your friends liked around the web.

3. Popular. This screen lets you expand your circle and stay connected to what is happening on Glue around the web. You'll find 100s of books, music, movies, restaurants, wines and more that are popular among the Glue users.

As you browse through the items, you can either display them in the standard view as shown above, or you can switch over to a more fun "cover flow" view that allows you to quickly flip through the different films, books, restaurants, etc. similar to the way you browse through your albums on your iPhone/iPod.

The iPhone app is definitely a must-have for Glue users as they will enjoy having access to their friends' recommendations even when they're away from their computers.

In a world more swamped with content options every day, recommendation technology is poised to make a huge difference in our experience online. We identified recommendation tech as one of the 5 most important trends for 2008 but we may have jumped the gun just a little bit. Below is a quick profile of each of the five Strands finalists working to bring more of this paradigm into the present market, followed by our thoughts on which one we're most interested in.

Gravity R&D

Sentimetrix is another four person team, this one from the University of Maryland Institute for Advanced Computer Sciences. This startup analyzes text content around the web and "has automated sentiment extraction/analysis/scoring, the ability to find and quantify opinions in text." If this kind of technology is of interest to you, see also our review yesterday of the new BooRah API.

iletken is a mysterious project built by four Turkish college students at Koc University. It balances personal and social behavior to recommend advertisements "based on relevance."

Reccoon is a stealth project built by Peter Tegelaar and Dominiek ter Heide in the Netherlands. Ter Heide also worked on the Japanese social learning platform iKnow, which launched an API yesterday. Recoon appears to use the iPhone's GPS, user attention data like Last.fm listening history and the GeoNames reverse lookup API to notify you when you're near the location of an event you might like to participate in.

Commendo is a four person team from two universities in Austria. Team Commendo is both the Grand Prize winner and in first place for the Progress Prize in the Netflix Prize leaderboard.

Strands describes Commendo like this: "Commendo uses recommendation technologies to optimize the drug design process in the pharmaceutical industry, including speeding up drug development and the minimization of adverse drug reactions."

Our Take

All of these sound interesting but the one we're most excited to learn more about is Commendo, the pharmaceutical drug development recommendation engine. We're dubious about the political and economic world of big pharma, but we love innovation and that's a field where there's enough money and science on the line that there's a premium put on magic. Strands has products for all kinds of industries (we think their banking service is the coolest) but we'd love to see what Strands plus the Netflix Prize champs Commendo can do in pharma research.

Will Any of These Make a Difference?

Could these startups change the world? With a little bit of funding and possible acquisition by Strands, they could. Strands has customers around the world in everything from music to banking and mobile. They have a lifestreaming service, ala FriendFeed, that doesn't seem to be going anywhere yet, but getting backing from Strands is a great step for any little recommendation startup.

Bring on a smart future augmented by powerful recommendation technology!

Disclosure: Strands is an RWW sponsor. We'd have written about recommendation startups anyway, though, because we think they're really cool.

MyThings raised an $8 million Series A round in 2006 and has been growing steadily ever since, though according to Compete, this growth has stalled somewhat after the company saw its traffic go up considerably during the 2007 holiday season.

Relationships with Online Retailers

MyThings' strongest assets might be its relationships with a number of large online vendors, including Casio UK, Currys, Dell Canada, Tesco, TigerDirect, and Woolworths. Those retailers upload purchase information directly to MyThings and consumers can then access information about the product through MyThings' web site.

The service already has 1 million active users (though it is not clear how 'active' is defined here). MyThings is also very popular in the art market, where a large number of sellers and buyers use MyThings' 'Trace' due diligence service.

The company did not release any information about how it was going to use this new round of financing specifically. Chances are, however, that it will step up its marketing efforts during the upcoming holiday season. Hopefully, the company will also dedicate some of these resources to improving its back-end technology for parsing receipts.

As is typical for Apple, the company is not exactly transparent when it comes to describing how the 'Genius' feature actually works. It looks as if Apple compares your music selection to that of other users and then builds its recommendations based on this. We assume that iTunes looks at data about play and skip counts, beats per minute (which is available for all songs in the iTunes store), ratings, and playlists.

Because these recommendations are at least partly based on the libraries of other iTunes users, iTunes periodically downloads updated recommendations. You can also force an update from the 'Store' menu.

One fact that surprised us was that Apple often returned playlists for songs that were clearly mislabeled, which has led us to speculate if Apple, during the first run of Genius, actually creates an acoustic fingerprint for every song.

According to Apple, all the uploaded information is anonymized.

Does it Work?

In our tests, the recommendations and playlists were often spot-on, but also a bit inconsistent. Sometimes we would get great recommendations based on songs from rather obscure bands, while we sometimes couldn't get any recommendations based on songs from more popular and contemporary artists. For classical music, the recommendation feature basically didn't work at all.

We also noticed that the recommendations tend to favor more popular mainstream artists, but that could easily be a function of the current user base.

Apple points out that the recommendation engine will get better over time, as more users start uploading their information. It would be nice, however, if Apple also gave users a chance to tweak settings for themselves or at least gave us more information about how these recommendations are calculated.

One minor annoyance when using the recommendations is that if you decide to build a Genius playlist based on a song that is already playing, iTunes starts the song over after creating the new playlist.

What about Last.fm and Pandora?

As Last.fm co-founder Marting Stiksel pointed out in an interview with Wired's Eliot Van Buskirk, the 'Genius' feature basically validates what other music recommendation services have been doing for a long time.

It's also important to point out that a lot of other music recommendation services have strong, built-in social networking functions. Apple, even though it now has information about the listening habits of a large chunk of its users, does nothing to connect these users. One neat function, for example, would be for iTunes to show playlists from other users that have a certain songs in it. For now, though, it doesn't seem as if Apple is interested in adding these social aspects to iTunes anytime soon.

Rediscovering Music

For now, when it works, Apple's recommendations are actually a very nice way of rediscovering a lot of music that had long been sitting in our jukebox but never saw the day of light. We also assume that the shopping recommendations in the sidebar will drive more traffic to Apple's music store, especially once the recommendations get a bit better and users get comfortable with trusting Apple's recommendations.

LinkedIn users are now being shown personalized news targeting their industry verticals on the Business and Technology sections of NYTimes.com and will then be prompted to share those stories will professional associates.

A number of other social networks and bookmarking services have "share this story" links on NYT stories, but it's unclear how much traction those links alone are getting. Last month we wrote about one of those services, social news site Mixx, that's still seeing fewer than 1 million unique visitors per month despite "share this on Mixx" buttons on a long list of the biggest news sites in the world, including NYTimes.com.

How much more compelling is this partnership? We think it's a lot more compelling; check out the screenshots below and imagine the feedback loop this could create between the NYT and LinkedIn. LinkedIn has 25 million registered users and the NYT sees 17 million + unique visitors per month, but the partnership will need none the less to introduce more people to LinkedIn in order to really be a home run. See this NYT page for an "introduction to LinkedIn." That's pretty classy, though it's unclear yet when that link will be displayed and when it won't.

We'll see how the recommendation process works; we hope it doesn't rely exclusively only on explicitly shared links, but we'll see. This certainly gets the mental juices flowing about any number of other integration and recommendation possibilities.

One question we have is about money changing hands. There has been extensive discussion around the web of late about LinkedIn using partnerships as a revenue source and it wouldn't surprise us if the NYT is paying for this integration. LinkedIn may not be a huge social network, but its user demographics are some of the most financially desirable in the world.

We expect to see more partnerships like this emerge, perhaps from a chastised Facebook attempting to relaunch its Beacon program in a more acceptable fashion.

Customers of Spanish bank BBVA will now be offered recommended products and services, individual and anonymized aggregate analytics and personalized goal setting and alert services, all based on their banking activities.

What's Interesting About This Deal?

Using the Strands Social Recommender technology, BBVA will be able to offer intelligent observations and suggestions for personal finance. A demo of the product shows, for example, that users of the system might be given interesting statistics about the financial activities of people in a particular demographic group, then asked whether they belong to that group. It's like having a private, personal, math-powered financial adviser available for your use on demand.

With interfaces for the iPhone, BlackBerry and Nokia phones - analytics and recommendations will also be available outside of the desktop web browser. This is the kind of heavyweight application to see coming from online recommendation services.

Privacy Concerns

How will bank customers feel about having their personal and financial details thrown into the collective pot for analysis of recommendations to other customers? We think it may take some getting used to, but that kind of information is undoubtedly being aggregated inside of banks already. The prospect of allowing users to benefit directly from their collective data is an appealing one.

Will the recommendations offered all point crudely toward buying more services from the bank? Given the huge war chest that Strands commands and the caliber of hires they've made over the last year, we hope that the company's banking recommendations and observations will prove truly useful and engaging for customers and not just for the bank's bottom line.

Only time will tell, but we've said for some time that in a world drowning in data - powerful recommendation technologies that help point towards personally meaningful information have huge potential. Financial services are the next frontier for these experimental technologies and we hope that Strands will disclose statistics in time demonstrating the impact their service had on the financial lives of users around Europe.

Disclosure: Strands is a RWW sponsor.

As Corvida noted earlier, the next step for social media should be filtering, which makes the timing of The Filter's launch perfect. As web users and social media addicts become inundated with choices, there's now more need than ever for noise-reduction tools as opposed to just more aggregators.

In fact, Peter Gabriel's own reasons for creating this service echo the complaints of the information overloaded netizen. He says, "the first freedom the internet brought was the possibility of access to any content, at any time, or anywhere. Now that many of us are drowning in choice, we need good tools to help us make smart decisions."

The Filter Homepage

How The Filter Works

Some systems make recommendations based on your actions and history where others leverage the power of the crowd to find the best content, but The Filter combines both methods and uses data learned in one area to augment the other.

When you first sign up at The Filter, you begin by stepping through a brief profiling wizard in order for the site to establish an initial set of recommendations. As you begin to use the site, you can continue to personalize your recommendations in a way that's very much reminiscent of Amazon's "Recommended for You" section. On Amazon, items are rated with starts but The Filter uses a + / - sliding scale instead. However, like Amazon, you can mark items you own so they won't be recommended to you again while also helping the system get to know you better.

Getting to Know You

A second part to discovering your personalized tastes comes from "The Filter" which you download. This tool is a plugin application that works with either Windows Media Player, Winamp, and, of course, iTunes. Unlike other music player plugins like iLike, this piece of The Filter's system is not designed for social sharing, but actively collects data from your computer and sends it back to The Filter's servers. Where iLike uses the data it collects to help you discover friends of similar taste, The Filter solely uses that data for the purpose of improving recommendations.

However, that's not to say there isn't a friends element to The Filter - you can add friends and socialize with them via onsite mail and a very Facebook-esque  "Wall." However, besides inviting new friends via email, there doesn't really seem to be any good way to find new friends whose interests mesh with yours. This is one area where social music services like iLike and Last.fm have The Filter beat. For example, when you're viewing an artist's page, you're recommended more items like that artist, but not the profiles of other users who also like the artist, so it's hard to know where to begin with the friending process.

Another feature designed to improve recommendations is a profile import tool. To give The Filter a jumpstart, you can import your profiles from Last.fm and Flixster into the service to improve your recommendations quicker than if you had to start from scratch.

Is It Worthwhile?

What's most interesting about The Filter is the way that it combines your manually rated items, your buying history, your playing history, and your friends' likes (assuming you can find some) to provide an overall recommendation service. Its ability to stretch beyond just music to include videos, TV, and movies is also unique.

Yet it still feels somewhat lacking when compared to Last.fm or Flixster because, despite the social element it purports to have, it's difficult to locate other users to befriend on the service. Last.fm makes that easy - displaying other listeners when you go to play a song and offering numerous Groups where users can bond around a particular genre or artist. Flixster also has a tab on its homepage to help you "Meet People."

Additionally, the site was slow (although that could be launch day jitters) and occasionally buggy. For example, somehow clicking into the Genres section logged me out. The site's player was good - offering songs I enjoyed - but then again, Last.fm's radio is good too and they also offer software for scrobbling your tracks from your music player back to their site.

The Filter is still an interesting experiment, though. Instead of focusing on just one genre of entertainment, it has taken several different recommendation techniques and mashed them up to provide one overall aggregation and recommendation service for many different types of media.

If they can work out the kinks on the site and enhance the socializing aspect, The Filter could have a shot since it does have some unique features that make it interesting. For example, it offers entertainment news you can subscribe to and it recommends not just web videos, but also movies that go with an artist or band, and vice versa. And since you can continue to use Last.fm and Flixster (and hopefully other services in the future) to enhance The Filter's recommendations, then it's not really necessary to switch entirely. Instead, you can simply use The Filter more like an aggregator and filter for all the activity you do out on the social, entertainment-focused web.

An RSS reader like illumio could have a real shot at marketshare among the digerati if not for a few issues. The app, more of a competitor to the Newsgator desktop reader product line than to online readers, provides automatic filtering of your news feeds. Unlike technologies like AideRSS, which filters by popularity, illumio personalizes your feed reading experience by determining what's important to you and then displaying those top articles in a newspaper layout.

Illumio is not new, but its latest incarnation - illumio version 3.0 - was only released a couple of weeks ago. With this version, the app, although downloadable software, now launches within a web browser window. The UI has also been completely redesigned so articles are easier to read and navigation is simplified.

illumio

However, what's really interesting about illumio is not the fact that you can read your RSS in a newspaper layout - it's the built-in filtering technologies the software employs to do so. In order to determine your interests, illumio scans the files on your computer to discover your interests and expertise. Not to worry, though - this information is never shared with others, nor does it ever leave your PC, according to the company. The privacy-conscious set might find that a little bit disconcerting, but you have to admit - that's certainly a unique approach to uncovering someone's interests in order to personalize their news.

Using illumio

When configuring the software for the first time, you must initially specify some default interests, but after completing the configuration you can remove any unwanted feeds and upload your own OPML file. (It's too bad you can't just start with an OPML upload, though.)

Once you're up-and-running, your feeds are displayed in a newspaper layout that features a tag cloud of topics on the right and your feed groups on the left. You can rate articles with star ratings to further train illumio as to what you like.

There's also a tab at the top of the newspaper called "Questions," because, if it wasn't enough that illumio was a filtered RSS reader of sorts, it's also trying to be a Q&A service, too. Here you can ask questions and respond to those posed by others in the community. While that might be useful in a business environment where team members review feeds together and then discuss as a group, it's hard to see exactly how this would really benefit a typical user.

Recommendation Alone Doesn't Ensure Success

Unfortunately for illumio, their great strides in improving recommendation technologies are going to be overlooked by the community they wish to engage for two major reasons: 1) it's downloadable software, 2) it's Windows only.

While they are working on a Mac version (sign up here to be notified of its release), the fact that illumio is a software download is going to be a huge turn-off for many users. Those at illumio maintain that the reason for it being a download is due to privacy concerns - since it scours your hard drive to assist with its recommendation and filtering services, you wouldn't necessarily want that private data stored online. That being said, most users are looking for filtering and recommendation to occur within their web apps, so illumio doesn't have a chance at converting anyone beyond the already small niche of desktop reader fans.

Instead, Try Newsgator Online

By partnering with a company called SenseArray, NewsGator Online is now offering RSS feed recommendations to its users. These recommendations come from the data NewsGator had been collecting en masse from their users as well as from an individual's actions - like a thumbs up or thumbs down - that were performed in their desktop reader (either FeedDemon for Windows and NetNewsWire for the Mac).

This latest news comes on the heels of last month's announcement about Newsgator's incorporation of our favorite filtering service, AideRSS. While both of these technologies are currently only available in Newsgator's online reader, there are plans to make them available in the desktop readers as well.

Additionally, according to a blog post on Venture Chronicles, the company has also been working with mSpoke to provide a categorization capability to their products that will soon offer Wikipedia-style indexing of content.

Newsgator Online, image courtesy of Jeff Nolan

Who Will Win?

With illumio's commitment to being a downloadable product only, Newsgator has the advantage, but even it still faces opposition from the still popular online reader provided by Bloglines as well as the increasingly-popular Google Reader.

Although users are demanding products that provide filtering, it's yet to be seen whether or not they will ditch their currently preferred online readers just to have access to these tools. If anything, Google Reader is one of the worst when it comes to filtering - in fact, its social feature that lets friends share stories means that you are likely to read the same story over and over again. Yet, it is growing in popularity despite its lack of filtering. (That is, unless you just read your friends' shares, which could be see as a type of human fitltering for your RSS.)

However, it is nice to see some forward movement in the recommendation and filtering space, even if it's not available in all readers yet. The fact that it's out there will mean other web apps and desktop readers will need to start offering similar technologies in order to stay competitive.