By Lindsay Fraser
My all time favourite Far Side cartoon features two polar bears looming over an igloo. One says to the other: “I just love these things – cool and crunchy on the outside, warm and chewy on the inside”.
This cartoon comes to mind whenever I am talking to colleagues about the use of Google on Canadian government intranets. I can’t help but think that Google is better on the “outside” than it is on the “inside”. Now, don’t get me wrong, I am not a Google-basher. I love Google. I love Google today the way that I loved and adored Yahoo back in 1994, when I couldn’t imagine anything better.
Google is a tremendous resource, second only to GEDS in my personal favourites list. It is such a marvelous Internet search engine that it leads decision makers to think “anything that good must be a great solution for our intranet too – right?”
NO, No, an emphatic no. And I hate to say it – I really do.
I am the first to agree that any search is better than no search – absolutely.
But the black-box that is Google does not typically help workers inside the organization find the materials on the intranet that are best suited to their working requirements.
Granted, we don’t understand all that well how the Google algorithms work (a closely guarded secret as you might well appreciate); but we do understand that Google focuses on concepts such as word counts, proximity of words to other words, location of words in web pages (titles, top of page) etc. This focus allows Google to do a great job of bringing to its users potentially relevant returns from a massive and totally random world wide web.
But on the inside, where the information collection is much smaller, far less random and actually has some structure and logic to it, Google’s effectiveness is greatly reduced. Employees want access to particular types of information (leave forms, for example), and don’t want a results set that contains every instance of the word “leave”.
Employees also want easy access into the top level of various information resources (the key entry and summary documents) so they can understand the context of the information returned to them and then choose the depth to which they will perform their reading and analysis.
Owing to significant challenges at the time, Google designed its algorithms to foil search engine spammers and other unreliable publishers who were misrepresenting the nature of their information content in order to have it come up in as many relevant results lists as possible. Google couldn’t trust that the information that publishers provided in association with their Web resources (metadata) was truly reliable and therefore built algorithms that placed little credence in such data.
Internally though, organizations understand their information holdings and understand who the users of said holdings are. Organizations also have information architects, publishers and editors, who can assign reliable metadata such as objective indications of page importance, page “type” in a hierarchy and resource type (e.g. form, policy etc.). It is this trusted metadata that allows search engines to return truly relevant search results to organizational users – and sadly, metadata is not something that Google has been developed to exploit.
For internal intranet search purposes, organizations should look to search tools and appliances that will also allow for the rich exploitation of trusted publisher input. If Google is the only option, then organizations should look to acquire the most flexible of the Google offerings and to take full advantage of the “weighting” options that are available. Google is undeniably great – but perhaps better on the outside than on the inside.
Lindsay Fraser is the Principal and Practice Lead for Systemscope’s Information Management practice.She can be reached at fraser@systemscope.com.
