Dan Grigorovici
SemanticWeb.com Contributor

3. Build your strategy and business team for your Semantic Web startup

No matter how strong your technology is, solving the problems I discussed yesterday that are plaguing the discourse in Semantic Web startups in the consumer space today, we cannot do without strategy and business.

To comment on this article, go to the Datamation Blog.

We need to let business people drive the Semantic ship, as they are in the best position to answer the simple questions I mentioned yesterday. In my experience -- as I am sure in any corporate technology project anyone has experienced -- successful technology products have been successful precisely because they were started with the end goal in mind (business preparing the requirements after identifying the problem in need of a solution), rather than the other way around. Having a healthy dose of business reading (Jim Collins' "Built to Last" comes to mind) will help in realizing that a visionary technology is nothing without being embedded with business vision.

It seems to me that the only reason why our community is so keen on evangelizing Semantic Web (other than the fact that we've spent at least the last seven years working on it) to everyone willing to hear, is that (or at least should be) that we believe in its benefits. But again: Its benefits have nothing to do with technology itself; its benefits are about making the use of information to consumers and businesses, better, or easier. Interesting fact though: Whenever I ask a Semantic Web preacher, "would you support a different technology that solves the same problem faster, with the same effectiveness, etc.?," I tend to get answers such as "we need Semantic Web." And this to me is alarmingly sounding like technological totalitarianism, probably because of our need to validate those seven years + we've spent trying to deliver the Semantic Web vision "since 2001".

So my question is: How do we address the consumer/user when the issue of which particular technology delivers the same end result is not really important as long as its needs are met? I suspect that only our business team will be able to show us the path in solving.

4. Don't prove the obvious; don't fix what's working

A few people I talked to in the industry mentioned there is no value in improving what is already working. In other words, if the output of our Semantic Web projects will be a much better organized, really inter-connected, data web -- essentially delivering similar information to what is being delivered today by non-semantic applications -- we have spent our time trying to (im)prove the obvious, or fix what was already working.

The promise of Semantic Web is that not as much that it delivers more relevant information to us, or that it opens up the Web to querying it like a database, especially if these are about what essentially could be delivered by a non-semantic search engine, or our incessant love for browsing loads of pages (which, by the way, has the advantage of making money for businesses today); it should be that it opens up NEW information that was not available/accessible before; that it's "smarter," and I now don't have to spend hours still not being able to find out "the relationship between the price of tea in China and the Kansas hurricane of 1978," or "show me the list of pop singers whose last name is Johnson who are NOT members of band x." If all we are improving on while building the Semantic Web killer apps is the "guts" of the Internet which will essentially deliver similar, but only slightly better information, we need not bother spending time on these. Instead, we should prioritize our development (both technical and business) by focusing on delivery of what we could not have delivered before without it. The reason I am saying this is because a lot of Semantic startups are really delivering the same information as before, only more effectively (while costlier, due to schema "complexity tax," Nick Lothian might argue). Let us take a second in studying what "more effectively" here means.

First, more effectively means "more granularly," and the typical examples from Semantic Web arguments and use cases are related to situations when you're looking for very deep information which search engines cannot retrieve or cannot be relevant because they don't "understand" in machine language what the source information is "about". There is a flaw with the use of this argument for a business case though, and that is: we as users only spend less than 5.2% of the time in active information search, versus 42.6% in content browsing, or other Internet functional behaviors (as per Online Publishers Association's "Internet Activity Index" with data from April 2008, available here), which effectively means that we can't use deep information search as our only Semantic Web use case scenario.

The second means "more accessible," as in "more queryable." The flaw with this argument is that quite honestly, even though it's a great thing to query the web as a database, most of us, common users are not DBA's, don't think as one, and will never do so. So a query interface such as the few that exist today makes no immediate user sense to the large mass of potential adopters.

All in all, I think we need to refine our picks for the top examples when building the Semantic Web business and consumer case; that we need to go back to Business and Product Management/Marketing 101 by studying the consumer, starting with the business focus first, then using Semantic Web as the way to solve the most important problems first. I really don't see a lot of Semantic Web business cases thought out from this perspective, and I think we should start thinking this way.

5. A credibility issue is never solved with a new catchphrase

We missed a boat with the "Semantic Web" phrase; it has gotten a bad vibe through various circles, and there is an air of disbelief we need to break in a different way. Recently, there has been increased talk about replacing this term with the newer "Linked Data" (also the name of the LinkedData Planet Conference, organized by Jupiter Events on June 17-18, where you can find me, and I want you to come and ask the hard questions, and help our community solve them).

But I tend to think that the reason why we have difficulty using the former term was that we did not do a proper job on solving the issues above. Replacing it with a new terminology is a good strategy but will not help us if we won't address the same problems, or if we do it the same old way, upside down. I hope that this post will spur plenty of commentaries, cause plenty of digestive problems, or at least enough to help us re-focus through debate, constructive criticism, and collaboration, in addressing these issues (and others which I will tackle in future posts). Only this way will we be able to "dog food" our own agenda for data/information openness and the growth of a strong Semantic Web business community.

What do you think?

Post your comments here.

Dan Grigorovici is an authority on the semantic web and advertising. He is vice president of data strategies and analytics at Tacoda, a division of AOL that runs behavioral targeting advertising networks. Prior to Tacoda, Dan worked at Digitas as the vice president/associate director of digital analytics. His startup, Disruptive Logic, is building Life Engine, a consumer cross-channel intelligent agent technology that will personalize the Internet.