People Vertical Search and Personal Network Clouds or is it all just attention
Lijit is a personal content network, a place to manage the content of all your online identities, a similar service is Ziki.
Not only is it handy to manage all your identities, but Lijit and Ziki also allow you to search all your content, search your blogs, bookmarks, etc… in one go.
It gets better, since you are in a network environment, Lijit allows you to add friends (trust a friend), now you can search you, or search you and the friends you trust (not sure if you can search your friends friends), this is searching through a social filter.
Even better again is that it is portable, you can generate a search box widget.
For a visual explanation check out the excellent post, Personal Network Search.
Community Building for success
But what if people you trust don’t have a Lijit profile, this is what is going to sustain this service, they need to attract a massive community, otherwise people won’t don’t have a good enough selection from people to trust.
What they need to do on sign up is have a tell-a-friend feature, because if your friends are not Lijit, then who are you going to trust (add as friends).
For this to work properly all your friends need to join…so what is the selling point:
- Search all your web 2.0 identity in one place
- Search your friends web 2.0 identity in one place
- Search you and your friends in the one search
In order to build a community they need to be popular and to do this they need to hook up with some big players, and this is just what they have done with Feedburner.
They also seem to be following the del.icio.us lesson, that is, there needs to be personal value before community benefits, people need a reason to use it personally, it has to be useful at the immediate level.
Being able to manage and search your own scattered content is certainly a personal benefit, and being able to then connect with others turns into another personal benefit. The fact the others can search through a social filter benefits the community or visitors passing by…I suppose the community benefits in that from one website you have a choice of people vertical search engines.
Outside Lijit
Another way around this is a similar concept, to be able to search your friends without them having to be Lijit.
I think this is key, as an individual could search their network without members of their network even knowing what Lijit is, this will enable Lijit to work for you straight away.
A way to do this is to be able to add your blogroll, which is in the works, Ziki is already there, but not searchable yet.
NOTE: Ziki also lets you tag yourself and tag your friends (network), so it is also an expert locator - find people by explicit tag or find people by searching content.
This way your identity is your blog, bookmarks, photo’s, etc..and your blogroll. This kind of eliminates the need to add friends (people you trust), as you are adding them externally via your blogroll.
But this is limited to just their blogs, your friends on your blogroll would have their own bookmark feeds etc…so maybe you have to collect all these other non-blog feeds and add them to your blogroll before adding it to Lijit.
You could search:
- you and your blogroll (or limit it to just you)
In the ideal world you wouldn’t need to add your blogroll, as all the people in your blogroll would have a Lijit profile, so all you would do is add a friend (person you trust).
You could search:
- you
- you and your friends
- you and your friends friends
Without Lijit
How do we achieve a similar thing without a service like Lijit, this will reveal how useful it is in serving a dedicated purpose.
Let’s clarify this, I want to search my identity (blog, bookmarks, etc…), and my friends as well.
A way to do this is, for each feed in your RSS reader, grab that person’s other feeds, like bookmarks, photo’s, etc…
You could have each person’s batch of feeds in a folder, then if possible in the near future (a shout out to developers 
), select folders to search, you could select your folder (your blog feed, bookmark feed, etc. - yes this is subscribing to your own stuff) to search you, then add other folders to your query to also search people you trust.
In this scenario, you already have your blogroll, but then you are running around grabbing all the feeds that make up someone’s identity, firstly this is a lot of running around, and secondly how to you know what services they belong to, you would maybe search using UpScoop.
For this reason Lijit sounds like a Lijit service…plus it is in an open and social sharing environment.
Limit identity to just a their blog and Reading List
How can we can get the same effect as Lijit but just limiting someone’s identity to their blog and the feeds they read.
Let’s just say your friend = a feed in your RSS Reader.
The idea is to search all your friends RSS Readers, so if you subscribe to 50 feeds, that is 50 people, that means you want to be able to search in the RSS readers of these 50 people.
I guess this idea is restricted to Reading Lists, so for each person their identity is their blog and their blogroll, which unlike Lijit a person’s identity includes your blog feed, bookmarks feed, etc.
Maybe it is wise to think that you don’t need to search within a person’s RSS Reader, as searching in their blog and bookmarks is the cream of what they read in their RSS reader and from elsewhere as well.
HOW WOULD YOU DO THIS?
Bloglines allows you to make your RSS Reader public, this way using their advanced search you can search a users RSS Reader.
Other things you can do is search your subscriptions, or exclude your subscriptions.
Now what if Bloglines allowed you to add friends, then you could search friends subscriptions (excluding or including your subscriptions)…I think Rojo once attempted this with their contacts feature.
Spokeo is a social network RSS Reader, where you can actually subscribe to another Spokeo users RSS Reader.
What about non-Bloglines users, the key here would be for each RSS Reader on the planet to have a public version or at least a spliced feed, this way you could grab the spliced feed of your friends RSS Reader and put it in a folder in your own RSS Reader called friends of friends.
Your RSS reader would have 3 folders:
- your blog feed
- feeds you read (blogroll/Reading List)
- the feeds (blogroll/Reading List) your friends read
ie. each feed in your RSS Reader is a person who has their own RSS Reader
Then you would have to be able to search at the folder level.
OPML instead of spliced feeds
An alternative is to share your OPML, see SYO.
The idea here is for your RSS reader to subscribe to an OPML (not bulk loading feeds, but subscribing to the OPML itself), there are only a few RSS readers that can do this at the moment, the pioneer being Blogbridge, and FeedBlitz for email.
If you have 50 feeds (people), and you want to find the OPML of these people (their RSS Reader Reading List), you go to a people search registery like SYO.
Once you have found a person, grab their OPML and subscribe to it in your RSS Reader.
What if SYO was more than a registery, perhaps a social network, you sign up and add your blog, then add your OPML Reading List. A user could browse or search for people, once they have found a person they could search their blog or search their Reading List or search these together. This is searching you and the feeds you trust.
The social network part would be bookmarking other users, this way you can search that user and their world from your account. But what about if it was more than bookmarking a user, what if this was intergrated into your profile. This way someone could come along and search:
- your blog
- your blog and your Reading List (feeds you trust)
- your blog and your Reading List and your friends Reading List
You could also have an outline view in an OPML tree:
root OPML
….OPML include - You
……..RSS (your blog)
……..OPML include (Your Reading List)
…………RSS feed 13
…………RSS feed 2
…………RSS feed 3
…………RSS feed etc…
….OPML include - Friend A
……..RSS (Friend A blog)
……..OPML include(Friend A Reading List)
…………RSS feed 2
…………RSS feed 22
…………RSS feed 345
…………RSS feed etc…
….OPML include - Friend B
……..RSS (Friend B blog)
……..OPML include(Friend B Reading List)
…………RSS feed 19
…………RSS feed 22
…………RSS feed 567
…………RSS feed etc…
…and so on.
This example only shows adding friends at one level, but what about your friends friends.
This outline could go really wide to the right and repeat, eg. my friends are Friend A and Friend B, but Friend A could have friend B in their OPML, so this means Friend B appears twice in my OPML, but that’s ok, we are just browsing.
Also note that Friend A and Friend B both have RSS feed 22 in their Reading List.
What would be the ultimate is to be able to select multiple nodes to search, example, I could search within:
root OPML - search within the full-text of the whole tree
OPML include A - search within the full-text of just Friend A
OPML include You, OPML include Friend B Reading List - search within the full-text of You and Friend B’s Reading List
RSS feed 19, OPML include Your Reading List, RSS Friend A blog - search within the full-text of RSS feed 19, Your Reading List, Friend A’s blog
Personal Network Cloud
So far we have talked about searching you or your network (eg. Lijit searches your identity and the identity of friends you trust, a blogroll would just search your blog and your RSS Reading list).
Other features to consider is comparing your OPML blogroll to other people’s like you can do in SYO and feeds.reddit. This may show you your personal network cloud without having to manually add friends, your friends could already be shown to you based on how your interests match to their interests.
So you can find friends by comparing blogrolls, revealing similarities, but let’s go beyond the subscription web.
Other ways to find friends is by comparing your bookmarks with others.
Other data you can gather is analysing your trackbacks/inbound links, and analyse your outbound links, also comments on your blog, and blogs you comment on, see my post Blog clouds.
Somehow this would all be churned together and result in a personal network cloud, with you in the center.
It could start off with you in the centre and you could say:
“show me people who read the same blogs as me”, then you could add another layer and say;
“show me the people who bookmark the same links as me”, then you could add another layer and say;
“show me the people who I comment on”, then you could add another layer and say;
“show me the people who comment on my blog”, then you could add another layer and say;
“show me the people who link to me”, then you could add another layer and say;
“show me the people I link to”.
Further, it would strengthen a person’s proximity by comparing the variables above:
eg. people who comment on my blog - do I read the same blogs as these people, do I comment on these people’s blogs or other blogs they comment on, do I bookmarks similiar links to them, do they link to my blog posts, do I link to their blog posts.
You could add/remove any of these layers, you could also say “mix all or select multiple layers and show me a personal network cloud.”
You could even click on any person in your cloud, and their cloud would layer on top (this could get messy).
Also when you click on someone you can access and search any of their content (blog, bookmarks, etc…), plus as before in the outline tree you could choose to search any mix of variables from the cloud.
Besides your personal network cloud, you could also generate a personal recommendation cloud, see my post RSS reading recommendations.
Many RSS Readers recommend feeds, even 3rd parties like Outbrain, UltraGleeper (this also recommends posts)…also see SYO and feeds.reddit. Sphere also has related content (just to your blog not to your identity), and megite and Tailrank will recommend posts based on your blogroll…Findory and Spotback do this based on your reading behaviour. Feeds 2.0 fits in here somewhere.
Plus comparing your bookmarks is great for content recommendation, gre.gario.us used to be good for this, also see Groop.us
Attention
Maybe this whole post is about attention, and analysing it to derive person centric networks/clouds/recommmendations.
Perhaps ziki and Lijit are attention management services in essence…see APML for more attention variables.
Maybe this is the easiest way to achieve my goal. Record your attention, nearly everything happens in your browser, it could record everything you do.
Simply plugin it into a service for recommendations and a personal cloud.
The ultimate is an attention registery, plug it in and it compares you to other people, and build a personal network cloud.
Oops, is this spoonfeeding Big Brother…they are already tracking when I use public transport, use my credit card (shopping, eating), etc…
Related
Feedshow sharing OPML’s
OPML sharing with Feedvault
RSS Overload: NusEye and others
Meta-search via an OPML URL?
RSS readers and OPML implications
View comment reactions
![email.png](http://services.nexodyne.com/email/icon/GxbiQvgbwBgCnw%3D%3D/oIp%2BJMQ%3D/R01haWw%3D/0/image.png"){kind=icon}
Thanks for taking a look at Lijit. Your analysis is right on. We want to help publishers search the information they already have and the people they connect to. Last night we pushed a new release with enhanced network discovery from a users blog.
The feedburner guys are great and in the near future we will expand those searches out to the individual publishers networks as well. Look for more good stuff.
Take a look and give us your feedback.
Comment by Todd Vernon — February 14, 2007 @ 4:53 pm
Hi John,
Thanks for the kind words about Lijit. We just did a big release last night so be sure to take a gander at what’s new.
I totally agree that the key is to draw value from people who aren’t (yet!) Lijit members. With our release last night we get much closer to that ideal: We will automatically sync your search with your blogroll, and include content from those blogs in your search, as well as people on *their* blogroll. And in the next release, we’ll automatically pick up any delicious/flickr/LinkedIn/etc accounts for those other blogs too. So for example, if you have Fred Wilson in your blogroll, his blog, tags, and photos will all be automatically included in your search engine as well.
Is Lijit like Ziki? We both allow you to identify other places where you create content, but that’s where the similarity ends. Most importantly, Ziki seems to be aiming to be a destination site whereas Lijit is aiming to be a service. We don’t expect you to send your friends to your profile on Lijit, rather, we hope that your friends will find value in using Lijit’s features on your own blog.
Also, Lijit is committed to not being “yet-another-social-network” site. We will use the relationships that you’ve already published rather than ask you to explictly make connections again. Your blogroll is one set of relationships, your delicious network is another, your LinkedIn network yet another.
And lastly, I agree that in the end this is all about attention: For searchers, wanting to give your attention to what’s relevant. For publishers, wanting to understand what of your content people are paying attention to. I’m proud of the work that I did for AttentionTrust.org last year writing the Attention Recorder, and Lijit is yet another step to putting attention-givers back in the driver’s seat!
-stan
Founder and CTO, Lijit
Comment by Stan James — February 14, 2007 @ 6:17 pm
Stan and Todd,
Thanks for commenting.
The network feature is exactly what I had in mind, it doesn’t rely on others joining Lijit for it to work for “you”…perfect.
Plus, you say for every person on my blogroll, in the next release, you will somehow be able to harvest their other profiles eg. bookmarks, photo’s…all this without them even having to know what Lijit is.
I was wondering if you could include OPML import, instead of a feed at a time.
The option to import an OPML file…or to be able to import an OPML URL, this way when ever I add a feed to my blogroll it will automatically be added to my Lijit Network. In this sense Lijit is subscribing to my blogroll.
Being able to import my OPML, almost extends to sharing OPML’s like SYO…when I click on Stan’s profile, http://www.lijit.com/users/stan/profile
I see his content section, but I can’t seem to click on his blog or stumble upon to add it to my network.
I see his network section, but I can’t seem to click on an item here to add it to my own network, I’d also like to click on his OPML to include his whole blogroll.
Another thing is that you can add another Lijit user to your profile, but this adds in the network section, shouldn’t this add in a friends section.
When you do add a friend that you trust, does this mean your search engine will now include their content, network and friends…you may want to add a friend you can trust but exclude their blogroll (Network), you might just want their content section.
This brings me to browsing, I’d like to browse a directory of the content section:
- show me all blogs on Lijit, show me all del.icio.us on Lijit, show me all Stumble Upon on Lijit
I’d like to browse network limited to blogrolls:
- show me all blogrolls and who they belong to, and then click on an OPML to subscribe to this blogroll
I’d like to browse people:
- and add any section or any item from a section to your profile
One other thing, in the content section I’d like to see a heading for “comments I post on other blogs”.
OK, one last thing, can I get a bookmarklet like the Rollyo Rollbar, this way I can search My Lijit from a bookmarklet, and like the Rollbar have a drop down menu so I can search my content, or just part of my content eg. del.icio.us, or search just my blogroll, or just a friend, etc…
Just like the widgit’s you offer, it can also be portable as a browser offering.
It seems you are not going the Ziki way to actually stream content, even though you are archiving it to be able to search it, I guess this is what you mean the difference between a destination and a service site.
Congratulations on Lijit, I believe this to extremely appropraite for the moment we are at…scattered profiles, the need for social filtered searching, etc…
Cheers,
John
Comment by Johnt — February 15, 2007 @ 1:02 am
Hi John,
I’m about to get on a flight so this is short, but wanted to let you know that we do support OPML! Just go to “My Content” and click “Add a URL / RSS / OPML feed”, and enter the URL of your OPML.
Any feeds from the OPML file will be entered as being in your network.
Comment by Stan James — February 15, 2007 @ 6:12 pm
Will Social Networks and Vertical Search combine to challenge Google?
Publishers and advertising agencies have a very difficult challenge ahead as traditional “horizontal” media like newspapers, TV channels and magazines see their traditional demographics and advertising revenue streams fragmented by the increasing preference of consumers for online access and the huge presence of Google eroding their audiences and potential future revenues.
Perhaps they should remember the words of Sun Tsu, who once said “When the enemy is too strong to attack directly, then attack something he holds dear. Know that in all things he cannot be superior. Somewhere there is a gap in the armour, a weakness that can be attacked instead.” Google’s major strength – the clean search box and the ease of use, commoditised ad revenues, perhaps masks its principal weakness. As media content and advertising revenues fragment to serve thousands and thousands of “vertical” online communities based on lifestyle or profession, Google may suddenly seem standardised, commoditised and lacking a sense of unique community. Is Google becoming Wal-Mart, while vertical communities may prefer Harrods?
Whilst “horizontal” media companies are similar to supermarkets, specialist professional “vertical” publishers are very specific in serving niche communities with totally relevant content and requirements. However, the publisher’s principal operating difficulty in becoming adaptive to this asymmetric Web 2.0 opportunity is that most tend to run each of their print, exhibition and online titles/businesses as separate profit and loss items on their balance sheet. As a by-product the vast majority tend not to have a centralised IT infrastructure or the human IT skill sets to manage a large scale data centre or web spidering facility – the prerequisites needed to datamine and aggregate open source, user generated and blog content to create vertical slices of the Web that are relevant for their audiences. Publishers will also need to integrate this content into the online extensions of their print brands and thereby allowing advertisers the opportunity to target high value communities. In addition, the datamining, crawling and hosting to identify relevant open source content will also need to be a continual process due to the continual growth of user generated and open source content.
Convera have two very large data centres, an extensive web spidering capability and a web index. Convera are now partnering with a significant number of specialist B2B publishers to create a range of vertical websites for specific professional communities. The first example of this is Searchmedica.com with UBM.
In building the deep vertical search portals, the key is to reach into the specific professional community in a number of ways. First, you can combined the trade publisher’s knowledge and contacts in the profession with community appeals that engage the specific audience in a way that general search cannot, and also by taking special care to use the taxonomies common to the targeted profession in organizing search results so that the user feels more at home and among peers. Building a good vertical engine can be costly and time consuming, and getting a critical mass of users to de-Google their search habits into more specialized engines is potentially a tough sell. However, in tests with focus groups from different professional communities to test these vertical search properties against Google, the results are hugely encouraging.
In building the beta test sites, the specialist publishers are providing Convera with “white lists” of data sources online and websites that would be most relevant to its readers so that the searches are restricted to reliable and trusted information. Publishers are also securing agreements with owners of key proprietary content not normally crawled by Google by leveraging some of its contacts and resources so that Convera can crawl and deliver some of their proprietary content. Another key consideration is getting the user community engaged in the process as co-developers. No matter how bad the results at Google or Yahoo may be for a given professional segment, the interface is familiar and the destination is always at hand. Getting users to think of a specialized brand as the go-to place for business information is the challenge.
A number of publishers are actively assessing the potential of adding social networking to the mix in order to get professionals interacting with each other and adding weekly podcasts by industry experts on issues affecting the community – these additional services will create more community loyalty and also additional advertising and sponsorship opportunities.
The publishers can also use their print titles to drive the audience to the new online areas and this will also assist the transition of their high value print ad revenues to online. Publishers also have exhibitions, seminars, events and email newsletters to assist this transition – and recent research suggests that professional communities will actively attend seminars and events to meet peers and other members of their community. The theory goes that once you get some professionals involved then the viral mechanism or behavioural “Hive Mind” also kicks in and professional workers start referring to the vertical portal as a community source. It is also allows advertisers and public relations organisations access to a clearly defined, affluent, influential and stable audience.
Google does not allow you to have a beer with a potential business partner - it doesn’t have that sense of community. But Google is fighting back – the recent launch of Google Custom Search and acquisition of teenage social network sites indicates they are aware of their weakness – but specialist publishers see this as a Trojan Horse. Social networks for teenagers are highly transient and target a demographic that is volatile, unpredictable and has a low level of disposable income – whereas a social network alongside a vertical search service for 22,000 bio-chemists, 55,000 UK GP’s, 55,000 insurance risk assessors or 120,000 US psychiatrists is stable, affluent and attractive for advertisers.
Comment by Andy Black — March 19, 2007 @ 12:58 pm