Archive for the ‘Search’ Category
Being smart about your data – Adrian Holovaty
Create serendipitous webapps – it increases stickiness and usefulness
Serendipity is providing the ability to drill or browse data e.g. IMDB, Amazon
Search for a movie, click on actors, year etc… Displays the data relating to the facet.
To make this happen we have to be smart about our data.
It all starts with the structure. Getting data structured data is half the battle – the rest is easy
The problem with data blobs – can’t do anything with them. Need clean Key | Value pairs
e.g Flickr lets you browse by subject, who took the photo, location, camera etc…
Give your data “the treatment” i.e take it from the big DB and turn it into efficient and serendipitous Hypertext.
Permalinks are very important. You must be able to create permanent link to concepts or specific information.
Data granularity is great for SEO
More Examples…
Chicagocrime.org
View crimes by date | location | type | hour | precint | crimes on route
Washingtonpost.com – Faces of the fallen
View by age | state or territory | military branch | by year | month | day
Video Game Reviews.com
Mixed Messages.com
THOUGHT: It struck me that our very own Secure Forms is only 1 degree away from allowing our customers to create these structured data browsing pages by themselves on the fly. We need to talk about and promote this more effectively – in fact this is great and could/should be a contender for our next event as it’s a natural extension of the search concepts we promoted at the first event.
Search
It’s not until a small team get together to build an online search tool that you get an appreciation of what Google and others have done. We recently embarked on a project to improve the search in our Enterprise Content Management System – Community Manager. Last year we introduced federated search and we wanted to build on that success and improve search relevance.
Our challenge – improve search relevance whilst adhering to access rights, incorporate a thesaurus and taxonomies, cater for best bets, stemming and spell checking without sacrificing performance. In our industry this is sometimes referred to as a “non trivial” task.
The Team – We put the band back together on this one. It consisted of a formidable team of experts in the fields of search, taxonomy, thesaurus, usability, sql and asp.net.
The Build – From start to beta it took us 6 weeks to build and implement albeit there were some lengthy requirements sessions and sleepless nights.
The Result – A dramatic improvement to search relevance. Initially there were performance impacts but with subsequent tuning we’re now seeing a fast and highly effective search.
Search is a simple concept that anyone can grasp but when you scratch the surface there are layers of complexity. I guess this explains why Google have thrown some of the brightest minds and resources at the challenge. The great thing is that we now have our very own search product and it even has a name – Community Search.
