« BI Is A-Changin | Main | Tag, You »
December 19, 2006
Information Magnetism
Search and Query are both incredibly poor instruments to find the right information, when we think of information as a market. A market is defined by a match between supply and demand, and that is the case with information as well. There is demand for information, people wanting access to what they are looking for, and there is supply of information, on the internet, in databases and many other sources.
However, search and query only recognize the demand for information, the supply side is totally reactive. Let me explain what I mean by exploring how buying shoes would look like if we were to use search and query techniques.
Scenario 1
You go into the shoe store and ask for black shoes of a certain brand. The salesperson says Sorry, I dont have them. Disappointed, you leave the store.
Scenario 2
You go into the shoe store and ask for black shoes of a certain brand. The salesperson says Wait a minute and comes back with size 6 through 15, as you forgot to specify which size. You think the person is crazy and you leave the store
Scenario 3
You go into the shoe store and ask for black shoes of a certain brand. The salesperson says Let me see and starts looking for it. In the meantime, you eye catches another pair of shoes that are nice too. You decide to buy the other shoes, and you are much happier with the new choice.
The first two scenarios show how poor search and query behave in real-life situations. The third scenario is more realistic, but is based on coincidence. However, that is not how a successful sales interaction works.
Scenario 4
You go into the shoe store and ask for black shoes of a certain brand. The salesperson says It looks like you wear a size 10, sure, let me get them for you (he may not even have them), and continues In the meantime, please have a look at these shoes as well, as they fit really well with that beautiful suit you are wearing. You look at the other shoes and although they are more expensive, you buy the other shoes, happy that you found something new and even more worthwhile.
Both search and query do not work that way, it is all demand driven, where the supply side of information simply responds. Like the sales conversation in Scenario 4 was bidirectional (supply and demand both influenced each other in the outcome), search and query should be bidirectional as well.
- If someone searches for a certain aggregation of data for a number of states on the East Coast of the USA in a database, the tool could/should respond with here is the data for Region East, which is almost the collection of states you asked for, as using the standard hierarchies may be more important for comparisons of data than a specific question.
- If you ask for planning data on a certain level of granularity, perhaps the system responds with the nearest aggregate level of available data.
- If you query for a certain filtered set of data, the system may respond with a slightly larger or narrower set of results, to produce a statistically more relevant bell curve.
- If you define a query that is quite near the query within a standard report, it should show the standard report.
- If you ask for certain information, the system should suggest, not unlike amazon.com, what other people with the same search or query asked for as well.
I would like to call this principle information magnetism, where supply and demand of information are attracted to each other based on common rules and understanding.
Thoughts?
Posted by Frank Buytenkijk at December 19, 2006 3:11 AM
