After having a chat with Daniel Harris, founder of Kendra Initiative, on March 13th, 2006 CE we decided that I should try to write down my thoughts about the various aspects of Kendra Inititative that we discussed. This mostly focused on catalogue publishing and search (although I haven't focused very tightly on this here).
For those that have been thinking about this for a long time I expect that most of what I say isn't all that new. I just hope that I'm saying it in a slightly different way to how Kendra Initiative describes itself. Maybe that will help a wider audience to understand it too.
The demonstration system which is part of the Kendra web site shows that it is possible to create a generic database engine for storing content catalogues. Although a nice system it isn't what I think should be the main focus of the effort (although the Initiative must have available demonstration catalogues so I'm not saying that building it is a waste of time). I have my doubts that content owners will all by happy to upload their catalogues to Kendra's systems, but in any case it seems to me that running catalogue servers is not what the Kendra Initiative is all about.
The system currently allows for the upload of content catalogues from any database schema. It allows arbitrary data structures to be created through the web interface and most interestingly also allows equivalences to be set up.
So, for example, if one catalogue's definition of track includes performer it is possible to state within the system that this matches a second catalogue's song and artist. An end user should then feel free to search asking for either songs or tracks by Thievery Corporation and expect to have both databases searched. At the moment this works as expected because the demonstration system hosts a copy of each catalogue.
In order to clarify what I think the direction of the technology that Kendra is building I'm going to try on some different "hats" to show where I think the highest value will come from.
I have a pile of DVDs, stacks of CDs and a reading habbit that consumes whole libraries of books. I'm always on the lookout for new films, bands and books. This is all the sort of content that Kendra wants to enable me to get at.
The aspects of Kendra Initiative that are going to help me to get and pay for these media is not what concerns me here. What I want to look at is how I find it in the first place. The key part of this is from Kendra's avowed goals:
[…] consumers can use any device or application to browse, search and purchase from the globally distributed collection of content catalogues.
I do indeed want to be able to use any device. As well as my computer I expect to be able to use my mobile phone and I'd love to be able to search using my portable MP3 player. I expect that the searching won't cost me anything.
What I'm not going to do though is to input my own mapping of what I think the search parameters are going to be. For example, I don't want to be presented by both track and song as search terms. I expect my search vendor to make these decisions for me and if I don't like the ones they've defined then I'll go search somewhere else.
I also don't expect that I'll go visit Kendra's web servers to search for music or films any more than I'd go to W3C's servers to search the web.
The obvious places to look for content are the existing search engines, the on-line music stores and the content owners themselves.
I suspect that the existing search engines will pick up on the catalogues wherever they find them and in whatever format they are presented in. I also think that the content owners have a clear commercial interest in making their content searchable (even if they don't all see this at the moment). More interesting are the middle-men; they have an interest in allowing search in order to match their customers with things they want to buy.
Two clear examples are Amazon and iTunes music store. They have a clear mandate to offer as wide a selection of content that they can and this means having an efficient way of finding new content to sell. Both of these players have existing ways of getting content owner's catalogues into their systems, but if the content owners start to publish their catalogues with a richer data set then this is only going to benefit the store owners (and in turn the content owners themselves).
If they can gain access to the long tail of small independant publishers by getting structured data to search and offer to their customers then, again, this is of clear commercial benefit to them.
This data publication is also going to open up a better market for new content distributors. At the moment I can repackage Amazon's catalogue under my own terms, but I can't go direct to every content owner. Opening the catalogues from the content owners will open this marketplace up and allow them to take direct advantage of every new niche.
The big players have their own agenda and their own strategy (whatever we may think of it), but there are huge numbers of smaller content owners who would love to be able to plug in to existing merchant systems to gain wider exposure.
I can already sell the content that I own through my own web site and I can add it to Amazon's catalogue if I dance through their hoops in the right way. But, the marginal cost for me to add a new distribution channel is high with this way of working. If I can publish the catalogue telling merchants what they'll find and how to understand the information then my marginal cost for each increased distribution channel is almost zero.
For a content owner who isn't able to brow-beat retailers or just announce a new title to have shops flocking to my door to sell it for me this is a major boon to my business.
I see that there are several opportunities to sell software and services based on this:
I don't see any threat from Kendra Initiative building demonstration catalogue and search technologies and even them giving them away. The market here is large and diverse enough that there will be many opportunities for all sorts of software and service sales.
To this end FSL Technologies Ltd will build a demonstration catalogue. I think to have a system that demonstrates Kendra Initiative thinking as a seperate implementation gives a lot more credence that the schemes work as they should.
I think Kendra Initiative is going the right way in focusing on content search (via catalogue publishing) as the first thing to focus on. If the users can't find content they may want to buy then every other part of the chain is moot.
I believe that the number one goal of Kendra Initiative in terms of this search aspect is to define the protocols and the standards to let everybody read the content catalogues and most importantly to understand them.
What I don't think should (or will) happen though is Kendra Initiative being a host for the catalogues, and I don't even think Kendra Initiative can be a host for all the catalogues' schema mappings. I think that all Kendra Initiative can and should do is to host the standards that the catalogues are published under.