This talk looks at how RDFa will impact on both the accuracy of search, and how the results are displayed. It also looks at how RDFa can help to create a range of vertical search engines.
It was given in a panel discussion, entitled <a>A Shift in SEO Strategy</a>, at SemTech 2009.
7. @pmika said today that
richer UI improves
click-through by 15%
@othar_hansson said the
increase was enough
for the big players to
adopt without question
Over the coming years we'll see substantial changes in the way search works, thanks to a growing use and awareness of semantic technologies. In this presentation I'm just going to focus on the changes that I see coming about as a result of RDF.
RDFa will impact search by enriching the results returned, by improving the accuracy of search, and by allowing vertical searche-engines to be created.
Let's look at how search results can be enriched.
Yahoo! was the first to enter the fray, by using Microformats and RDFa to enhance individual search results. Here we can see that the normal textual summary that you'd expect for a search result has been replaced with information that is pertinent to the particular item that was found; since we've found a film, we'd like to know how long it runs for, when it was released, what reviews are available, and of course, the nearest cinema that will be showing it.
Note that this is not just about improving the UI for the sake of it -- from the point of view of the search engine, it means that you'll spend more time on their site, because not only do they help you find the things you want, but they help you to do, the things you want.
Given the enormous potential this has for turning the search engine into a kind of portal, Google was never going to be far behind Yahoo!.
We should stress that click-through is becoming more important than ranking for many people. According to SEO experts, a click-through improvement of just a few per cent would get companies excited (particularly large ones) -- with richer search results, it would appear they are getting many times that.
We've looked at improving the search results, but what about the accuracy of our searches?
By adding mark-up to our HTML pages we can clarify ambiguities. For example, we may have a news story that simply says "Today that Primer Minister said...", and we can add RDFa that says the particular PM being referred to is Gordon Brown.
It's not clear yet whether this will help rankings, but it is almost certain to help click-through.
Finally, an area that I've not seen anyone talk about, creating vertical search engines.
To illustrate, we'll imagine that we're trying to create a search engine for chemists, to help them find articles and research on particular chemicals.
If you look at a search like this, you'll see that the top results for 'asparagine', articles that would not be of interest to a chemist (general articles from Wikipedia, for example).
Verticals like chemistry try to set up their own search engines, tracking articles and research papers on chemistry. However, it's actually a pretty difficult task, and usually requires sites to register with the central search engine. If you play with this site long enough, you'll quickly find sites and pages within sites that are missing.
The important point here is that there is a real need for specialists to be able to use the internet as a tool, in the same way that the rest of us can. And for that they need to be able to search across documents that are within their field, rather than across the whole web of documents.
This is where RDFa can help, because by marking up a page with things like chemical identifiers, we can be sure that the crawler understands what it is receiving, allowing the data to then be partitioned into different vertical sites.
Hopefully we've seen how RDFa has a key part to play in changing search yet again.