42 Quirks

Part II – I Seek You, and your meta-data, too…

What does making sense out of data mean?

In Search Engine terms, it would refer to contextualizing the huge chunk of uncontextual data that is the World Wide Web into information and eventually knowledge. To me, as a human, it simply means tagging certain keywords to any given chunk of data (e.g. a lecture, a passage, a book, a chapter, a conversation) in order to be able to recall it at any time – especially, when one of these keywords is mentioned.

For instance, the conversation in the previous post was about a traveller, (an out-of-towner) looking for directions to a tobbacconist. As I keep reminding myself, Web 2.0 is not a product, it is a process. The process has a lot of conversational threads that keep getting picked and dropped as newer and more interesting threads or new participants appear in their place.

So what would a contemporary Search engine have to consider in Web 2.0?

‘Weight’ing for Information.

From being a static display of items-for-sale behind elegant window panes, the Internet slowly transformed into a bazaar of sorts, with hawkers all around the place plying their wares. The markets grew to accomodate the new and the old. With the advent of Web 2.0, contextualization of information became the norm and not an option.

It all began with a nifty bookmarking site called del.icio.us that allowed you to access your favorite sites across the web. Technorati extended the concept to Blogs and induced bloggers to ‘tag’ their posts with their choice of keywords/tags.

With the Web evolving like a democracy, the obvious question of authority in the Web-democracy arose. Which voice among the loud babble was to be trusted? As the web evolved, so did the concept of it’s franchise. Only, in this virtual reality, links were deemed votes and tags were your campaign ads. Let’s take a quick look at the foru weights that influence your vote.

1. Tags – Powerful Keywords

Each tag is a keyword that associates a particular context, a topic, with a given chunk of data.

2. Time – The ‘other’ Long Tail

All topics & data have a peak presence time. The freshness of a particular keyword is of prime importance in its influence.
Consider this simple example: When Iraq was attacked, almost all of the Search Engines across the world were buzzing with Search queries consisting of corresponding keywords, viz., “Iraq” “attack”. The “hotness” of the Search cooled down as the days progressed, as the world got other topics to discuss about.

3. Trust & Authority

Even in flat hierarchies like the Internet there are obvious postitions of Trust and Authority. People who blog well, and blog often gain a large following, and effectively, the crucial element of Trust.

4. Authenticity

A news on a Microsoft blog would obviously be rated higher in all terms than a news quoting a “trusted Source at Microsoft”. The only exceptions to this rule are:

• The news is a really good bit of juicy gossip – like a rant or a ‘leaked’ secret
• The blogger has high levels of Trust & Authority

There’s a common thread that binds all of these.. Do you see it yet?

(To be concluded)

Note:
I profusely apologise for disappearing from the Blogging scene, all of a sudden. I was forced into a short hiatus by unforeseen circumstances. We updated our website platform to a new version, recently. although the beta is pretty stable, we are still working on a better UI. As a result, I had to spend some sleepless nights and a few Blog-less weeks. ;)

Once again, my sincere apologies for the same…

Technorati Tags: search, search-engine, tags, web2.0

powered by performancing firefox

Part 1: Search and the Web 1.0: Gorblimey!

Those of you who reached here through Google, Yahoo or MSN are probably laughing as you read this. But do go on, there’s more. :-)

(Un)common Recurring Searches

Often our searches are simple keywords crafted with central themes in mind:

• A name (e.g. Shrikant Joshi or Performancing)
• A topic (e.g. Corporate Communications)
• A context (e.g. “Spanish Omelette” +recipe)

Some of us might even burden the spartan box (or in the old days, the Butler) with an entire question. The faithful zombie then crawls its way through the innards of the webs, looking for that occasional diamond stashed away in the back alleys. Usually, in the common cases such as the ones defined above, results are returned in the correct context of our request. Often, the SERPs also throw results that are related yet not within context.

Robert Scoble’s post on Optimization had this line that caught my attention:

It all starts with the blog. Now, why can’t I put my blog on the map? When you go to Live.com and search on “Scoble” why can’t I customize my results there with more information for you?

Well, I don’t agree wholly.

Search for my name on Google. There are at least three different people called Shrikant Joshi who turn up in the top 3. We keep exchanging the first three ranks. And all of us are pretty active bloggers it would seem. The see-sawing of rankings in the Organic Search results is not a matter of concern for me. Nor do I want to customise these search results so that I would get more result-space.

I am not a key-word

What are search engines? Simply speaking, search engines are content-aggregators assigned the additional job of classification. As humans we need to have everything classified into a taxonomy so as to facilitate recollection. Our knowledge depends upon storage which in turn depends upon collection and classification of data. Classification helps recollection and hence improves perceptive retention of knowledge.

Or, in simple words:

The more you know, the wiser you are. Hence, classify and remember.

Similar to how we retain knowledge, Search Engines classify the data they crawl according to keywords. A huge index is built up and referenced and cross-referenced until all the possible avenues of keywords linking to pages and vice-versa are covered. But you probably know all that and more already.

Keywords, mmmm… Aah!

The next step would be making sense out of the data, which eventually leads to contextualization. Don’t get it? Well, simply put:

“A search engine’s job is to make sense out of all that data.”

Let’s take a simple case. Someone in your town happens to own a convenience store named Uncle Tom’s Cabin. Let us imagine that an outsider in your city is searching for it. Here’s how the conversation would go:

Outsider: “Where can I find a convenience store?”
You: “That would have to be Uncle Tom’s cabin. Go straight down for about two blocks and then take a left. It’s right across the street.”
Outsider: “Would I be likely to get some cigarettes there?”
You: “Oh! If you simply wants cigarettes, there a tobacconist just round the corner!”

A normal conversation, eh? Well, let’s take a look at it again. Only this time, we’ll look at it the way a search engine would. Let’s insert some key words into it for understanding the flow of the conversation:

1. “Where can I find a convenience store?”
[New Search Query, keyword: "convenience store"]
2. “That would have to be Uncle Tom’s cabin. Go straight down for about two blocks and then take a left. It’s right across the street.”
[Response keywords:"Uncle Tom's cabin", "directions"]
3. “Would I be likely to get some cigarettes there?”
[Refine Search Query, keyword: "cigarettes"]
4. “Oh! If you simply wants cigarettes, there a tobacconist just round the corner!”
[Response keywords: "Tobacconist","Round the corner"]

With me so far? Here’s the stumper:

If each of these sentences corresponded to an entire blog-post in the Blogosphere, how would you track this conversation? How would you rank each post with respect to the keywords. Would those keywords be enough to cover all aspects of the conversation? Would you call those keywords as appropriate descriptors of the conversation? Where would these posts appear in SERPs for the combined keywords {“Your Name” +directions}

To be continued…

Disclaimer:
I am no Search Engine Expert. These opinions are simply my $0.02 worth. Or may be less. :)

Technorati Tags: search engine, optimization, keywords, web2.0, content, context

powered by performancing firefox