Webnomena

Short Introduction

Glue is a Firefox addon that uses semantic analysis to connect people around books, movies, music, and other common things across popular sites. Glue can understand and map both structured and unstructured data and then become the bridge (or better the glue) that connects people looking at the same object from multiple web sites. There are few good blog posts that have covered Glue already. This is the reason that I’m keeping this intro short. I prefer to focus on the value of using Glue.

 What to do on glue?

Start Glue-ing by visiting a page on one of the many web sites that Glue supports like Wikipedia, Amazon, Last.fm, O’Reilly books, Yahoo! Finance, Citysearch, Wine.com, IMDB, Last.fm and many more. When you’re looking at the book, music, movie, star, artist, stock, wine, or restaurant you will see the Glue toolbar slide down from the top of the web page. Glue’s toolbar shows you friends and other Glue users that visited the same object. Glue shows you friends who liked the object and you can read their “2 cents” – a short comment about the object (140 chars long). At this point you are presented with lots of ways to benefit (actually, more than I realized the first few times I use it). Here are some of the things to do next:

• Read a summary describing the object
• Check which other Glue members visited the same object (anywhere on the web)
• Read others’ comments (two cents).
• Take action
  • Object specific actions – find it on your preferred web site, read a review, compare its price, find similar objects, and more
  • Sharing option – use Facebook, MySpace, Twitter, FriendFeed, and Delicious to tell the world about it
• Learn more about the people that liked it by looking at their profile
• Follow people and become friends if they follow you back
• Say that you like it – by pressing the heart shaped button
• Add your two cents
• Grow your knowledge and network by moving from things to people to things to people and stop only to connect, comment or to take an action.

What else to do on glue?

• Grow your network on other social networks like Twitter or FriendFeed – most Glue members have their Twitter and FriendFeed accounts linked to Glue. Glue allows you to find out more about their interests before following them on these social networks.
• If you are into the stock market you can use a new service called StockTwits. This is a Twitter mashup that lets you follow discussions about stock trades, find active members to follow on Twitter and build your own portfolio to use as a filter for finding related conversations. When you select a stock to read related twitts you’ll see the Glue toolbar sliding down with all its glory. Now you can see other Glue members that were interested in the same stock and connect on both networks.

Explicit values

• Your network is built automatically as you browse your favorite web site without leaving it
• It is a single, web-wide network that works on popular book, movie, and music sites
• It replaces both search and bookmarking – Glue brings you the information when and where it makes sense
• It’s easy to move from object to people to object. This helps you find great books and music – this is what that I found myself doing on Glue.
• The option to take action no matter what web site you are browsing helps to complete your search tasks faster.
• Glue is very intuitive and simple to use. It takes not time to get on-board.
• Crowd wisdom – you can see what is most popular with friends and other interesting people

Implicit values

• The building of the network is driven by objects you like. You connect to likeminded people around common objects automatically, regardless of the website visited. Since there are lots of objects out there and many curious people looking for them it makes Glue a network building machine.
• Contextual lifestream filter – it shows users relevant information from friends about things they visit. Other lifestreams have a lot of noise and require work. Glue brings you a filtered lifestream of valuable information i.e friends activities wrapped around object and people in the context of an object. 
• Connect around the rare stuff – connecting around objects that are loved by many is a rapid way to build your network but some times it is meaningless, like joining the Facebook beer lovers group . Using Glue you have a good chance for finding new people that are interested in objects that are not so common like this amazing British TV series from the 90th that I like so much – Cracker (I did find a few Glue members that liked it).

AdaptiveBlue

AdaptiveBlue was Founded in February 2006 by Alex Iskold and has 11 employees working from their New York Office. The company has two products: Glue and SmartLinks (patent pending). In its short existence it earned industry recognition and top press and blogs.

Glue use two methods to understand meanings from data on the web. The top down approach using its semantic engine to understand some of the most popular web sites out there that don’t use any of the known metadata format (like RDF). AdaptiveBlue also collaborated on a new format to describe objects attributes on the web called ABMeta. Sites like Oreilly books, UGO and others have already adopted it. This is referred to as the bottom up approach, which is a more robust way to make web pages easier for machines to understand.

 Additional thoughts

After using Glue for sometime now I have a few features that I hope to see in the future. The first one is coming soon and it is the option to discuss with friends about different objects.

• I also would like to get an alert when someone was looking at one of the objects that I have visited in the past (set selectively on certain objects)
• I think that Glue needs a landing page. The toolbar is cool and subtle yet there is a place for presenting some aggregated data like:
  • Most active people on Glue (sorted by object type)
  • Most looked at objects – most liked objects
  • Most connected people on glue – featured users
  • Recently joined and recently visited objects
  • Promotions – Glue knows what people are looking for and like. This informarion gives an opportunity to get some nice deals for its members.

Glue is a simple to use application with great benefit supported by very complex technology in the back-end. It manages to bring a lot of value to the front-end without scarifying usability and ease of use. Is Glue the first consumer application that’s showing us the semantic web finally fulfilling its promise?

Headup – the semantic web Firefox addon

I recently started using Headup. I’ve been looking for this kind of addon for some time now. When bits of information are missing from peoples’ profile pages, product specs, media, and other online content it is crucial to combine multiple data sources to piece together a complete picture. Headup does this!

Using its smart semantic mapping of entities and relationships Headup gathers and links information from multiple online sources. To complete the picture it then personalizes the results using your presence on multiple web services like Gmail, Twitter, Facebook, Digg, etc.
Headup is not only innovative in its semantic approach to linking data, it also integrates nicely with your Firefox browser and offers you a few ways to access the data it discovers. One example is Google searches: After installing Headup you can expect to see your search term annotated “Headup:[search term]” with a thin orange underline at the top of Google’s results. When your mouse hovers over the term a click-able circular plus sign loader will allow you to open Headup’s overlay  interface.

 

The starting point – googling eagle eye.

The complete picture – headup-ing eagle eye

I recommend you visit Headup‘s website to learn how to use it but as a whole it’s pretty intuitive and I prefer dedicating this post to the reasons you should get it:

My eight reasons for using the Headup Semantic Web Firefox add-on :

1. Because hyperlinks simply aren’t enough – Relying merely on arbitrarily selected outbound links that send you to find info related to the page you are browsing is limiting. There are more relationships among the different entities on the page that could be leveraged to retrieve associated information. Headup already mapped out these semantic links and makes them available for you in a neat and accessible interface. The experience doesn’t end with search results.
2. Because you can save valuable search time - Both the user interface, and the way information is presented, require less clicks to complete an in-depth search through multiple search sources.
3. Because the information comes to you – Search can be an exhausting task. In many cases it involves either a recursive drilling down into multiple levels, or traversing the search vertical up and down for additional information. Google itself is aware of this potentially laborious process and is making an effort to bring associated information to the first SERP: Recently when I googled the term “movie” I got three results that were movies playing in theaters in my area. Headup provides multiple data types as a default: Using Headup on the “Pink Floyd” will get you a summary relating to the term, the bands albums, see photos depicting it, listen to the bands songs while reading their lyrics, find news blogs and web activities related to it, and much more.
4. Because it brings down the chances you’ll miss key information – “Headuping” people is a terrific way to learn more about them. I “Headup-ed” my friend Bill Cammack on Facebook and immediately discovered that he’s a video editor with an Emmy award to his name. In this case the extra information regarding the Emmy award was brought in from Bill’s LinkedIn profile.
5. Because you can learn and find information you didn’t expect -  If the example from my previous item wasn’t proof enough here’s anoter example: I ran Headup on “Kill Bill” (what can I say? – I’m a Tarantino fan) and discovered this blog post published today (1-2-2009): “More Kill Bill on the way” – Tell me this isn’t cool!!
6. Because it’s personalized – When configuring Headup after download, or later via the “Settings” option, you can choose to connect Headup to the online services you are subscribed to. Headup connects to a wide variety of web services like: Gmail, Delicious, Twitter, Facebook, FriendDeed, Digg, Last.fm etc. The information Headup retrieves from these services allows it to personalize the info it discovers for you: If you Headup a firm you’ll get friends of yours that work there. If you Headup a band you’ll see who in your network likes them. This is another example of how Headup is not just a search tool but a browsing experience.
7. Because you don’t lose your starting point – Headup is designed as an overlay window that keeps your starting web page, and anything else you have open on your desktop, visible beneath the interfaces’ SilverLight frame. Inside Headup you can drill down endlessly, but when you’re done you are back where you started.
8. Because your information is safe – from Headup’s Privacy Policy – “In plain English”:

“We here at Headup treasure our privacy and that’s exactly why we made every effort to create a browser add-on that would live up to user privacy standards we would be comfortable with. We’d be embarrassed to let you download an add-on we wouldn’t download ourselves.”

 

**You don’t need to sign-up for using Headup and your information is stored on your machine only**

 **Bonus: one additional reason – because on some pages it ROCKS! Try it on last.fm and you’ll see why it ROCKS…literally! By the way, the Headup user interface lets you watch videos and listen to music like a regular media player.

My questions for the Headup team

I plan on occasionally checking Headup’s blog for updates. At this point Headup supports Firefox on Windows and on Macs but I know that they plan to support more browsers in the future. I think that at this point the key thing to focus on is that the Headup concept works.

I do have few questions for the Headup team:

1. Do you plan on adding vertical derived classifications? I can see some use cases for health (and maybe even for software development). Just as headup was able to map out “Actors”, “Films by the same director”, “Web Activities”, “Related News”, “Trailers”, etc. for a “Film” type entity. I can see it applied in a similar fashion for a “Health” type entity – retrienving things like: “Case”, “Treatment”, “Clinics”, “Pharmaceuticals”, “News Groups” etc…
2. Do you see enterprise usage for Headup? I still need to give it more thought but having Headup in my email could be cool. Another possible implementation is supporting corporate CMS tools.

Epilogue – Is Headup’s “Top Down” approach the face of the future Semantic Web?

The Semantic web promises to make information understandable by machines. If you follow Alex Iskold‘s excellent series on Semantic Web on ReadWriteWeb you are aware of the multiple approaches to make this happen. The top-down method implemented by Headup helps brings the future to us a little sooner. I think Headup is giving us a taste of what future browsers will look like in an age when they, and other tools, will be able to understand more than just hyperlinks. When using Headup it feels like I’m doing more than “browsing” or “searching” I feel like I’m experiencing a new web!

One last thing: using Headup for some objects didn’t yield complete results. Don’t judge them too harshly for it, instead please focus on the concept. My experience with Headup so far is that in most cases the relevancy of the information provided was more than reasonable. I think that for a small company just out of Alpha what has been accomplished in the short time the company has exited is impressive and promises that improvements will be fast coming.

I’m using Headup and gave you the eight reason I have for doing so. If you are using it too I’d be happy to hear why…

  I recently noticed on LinkedIn that one of my old friend is now the CEO and founder of a new start-up company named Datanetis. I sent him an email because I was intrigue and it was also a good opportunity to catch up.  Elery Pfeffer and I worked together few years ago in a small start-up called Nester Software (later the name changed to Plataine) he was a student back then and I was thinking about moving to America. He is one of the brightest people I ever met.  Elery graduated from Tel-Aviv university in computer science and in a country with so many bright people and very few universities, it is as hard to become a student in Tel-Aviv university as it is in one of the Ivy League education institutions over here. Yet the thing that makes Elery a great friend is his strong integrity and generosity. Later he become the President of Nester, I made it to the state and we kept the friendship and mutual respect going.

Elery was kind enough to spend an hour with me over the phone and through GoToMeeting session he presented his new creation. Now, I was even more excited. Datanetis is the real thing! It is not another dot com web2.0 bubble company. The company has a real product with multiple patent pending applications describing algorithms for finding influencers, highly complex (real barriers to entry), and scalable. Datanetis is selling it to the enterprise as a hosted solution and providing new data about whom should the bushiness focus his best marketing effort on – leads to influencer on other leads. And they already have large customers around the world.

Elery view of the new marketing is revolutionary.

“The new marketing is not just about customer’s monetary value but also about the customer’s social value to the organization.”

For someone that has been working building software for the marketing automation industry over 8 years now and is familiar with multiple solutions for finding the right prospect out of many, it was an eye opener. I’m evidencing the progression from mass email campaigns through marketing to target individuals with a matching/relevant offers (data mining, behavioral pattern, collaborate filtering, recommendation engines) to finding customers that can market for you – agents.

Finding subtle connections between individual and causality.

Followers buy only in networks where influencers buy first

Quoting Harvard Business Review:

“The only path to profitability growth may lie in a company’s ability to get its loyal customers to become, in effect, its marketing department”

So who are these loyal customers?

Datanetis’s software is capable of both automatically generating social networks from low level CRM data within the customer database and to mathematically identify influencers and followers. This information could be used to increase the return on every marketing dollar spent on new product adoption(x5-x70), new customers acquisition(x12), churn prevention (x10-x30), conversion, product cross sell(x6), higher product virality, and significant cost reduction.

From Datanetis’s experience, social network marketing using influencers is comprised of two cycles. The first is the closed friends cycle, from the results it seems as if they almost decide simultaneously to follow the leader. This first wave peaks around 4-5 weeks from the beginning of the campaign. The second wave peaks around 9-10 weeks traversing through the rest of the social sphere. This may indicate that the customers are going through different decision making processes at different locations on the social network graph (influencers affinity). This knowledge offers Datanetis customers ways to fine tune their campaigns over time.

In my opinion this disruptive technology will force changing some of the operations and thinking in the marketing department going forward.

Viral marketing is not just hit and run but a multiple acts campaign.

Another interesting finding that Elery shared with me is the lack of influencers trivial characteristics. Any attempt, so far, trying to map influencers based on demographic, product adoption or any other factor distinguishing them from the population failed.

This is what Datanetis found about these individuals only after finding them using their software:

• Social Influencers are not celebrities and resellers
• Influencers typically
  • represent 7-15% of the total population
  • has influence in 3-5 different subject areas

Datanetis successfully executed hundreds of large scale social marketing campaigns and is globally active in Retail, Telco, Gaming, Internet, and Hospitality industries.

Here is what that Elery suggests to the marketers out there:

“Don’t focus only on turning leads into sales, focus on turning influencers into ambassadors for your company.”

The influencers phenomenon is covered vastly on the web ever since the invention of digital social networks and the social graph. From what that I read so far it seems to me that we are still in the experimental/research phase. I was excited about finding a real application with proven method for finding influencers. Maybe Datanetis will help pushing this new science forward.