Tuesday, October 09, 2007

The Power of the Long Tail

This Post is mainly to respond to some questions Point2 users have about long tail search in the real estate vertical, and about a particular stat we've quoted: "60% of local homes searches have a neighborhood component to them." The comment came from a Google official in a slide show presentation on Search and the Real Estate Vertical.

The context is that of all the search conducted for real estate, homes for sale, or property (i.e. the real estate vertical), 60% of the search terms contained a neighborhood component. This is also called the longtail of search, which constitues more narrowly defined search terms. While "San Francisco real estate" represents a broad search, an example of a longtail search with a neighborhood component would be: "Nob Hill San Francisco real estate". I don't have the specific Powerpoint or research that were presented,nor do I want to misrepresent what they discussed. But I can demonstrate what this means, leveraging research we conducted at Point2.

If you are not Google and you want to measure the amount of search volume, you need to have a site that generates a lot of long tail traffic. Plus, you need to rank at the top for the broad terms, because they drive the most traffic. This way you are likely to have a good representation of the relative size of long tail traffic vs. broad search traffic.

We used Point2 Homes and chose the city of Ottawa Canada because Point2 Homes ranks well for top search terms used by consumers, like "Ottawa real estate" and "Ottawa homes for sale." Of course Point2 Homes represents a good platform for this test because it is optimized to generate a large set of long search traffic.

Over a 2 day period, Point2 Homes generated 547 visits to Ottawa and it's neighborhood level pages. Of those 547 searches, 259 or 47% contained an Ottawa neighborhood name, proving that consumers use neighborhood terms when searching for real estate in this area.

The ugly screen shot above shows you a piece of the working spreadsheet. Where the first column lists the keyword term, the second column is simply identifying ones with a neighborhood name. The 3rd column shows the search volume for the keyword term.

We have tested several other Canadian cities and the results are very similar with longtail traffic, generating between 45% and 50% of the visits.

You can see from that screen shot that the first neighborhood searches that we identified include: "homes for rockliffe ottawa," "house for sale ottawa west," "houses for sale glebe ottawa," "condos in the riveria + ottawa" and, "country club village ottawa."

Real estate is very local and while not all areas are the same, the patterns we see across the US are very similar. For this research, we did not use US cities because we do not yet have a good set of cities on Point2 Homes that rank at the top for broad terms.

It's clear that there is a lot of traffic out there in the long tail and thus it's clear that we need to take advantage of that traffic and describe content accurately to match what search engine users type into their search bars.



suzstephens said...

Jeff, I appreciate your taking the time to post this explanation. I think that the best solution for many agents will be the one that takes them the least amount of time while still covering their needs. I work with a lot of different agents: they all have trouble getting everything done that they need to do, so I wouldn't want to have to tell any of them that they now need to go in and click several hundred counties. They need solutions that will save them time, not create extra work for them.

The highest traffic site on which I work and follow stats is RealEstateUtah.com, which has high 1st page ranking for "real estate utah". I don't have Windows running so that I can check her Ultrastats, but her Statcounter account has traced the last 311 searches on her site. Out of those, only ONE was for neighborhood related information. Meanwhile, there were over 150 searches on variations of the "real estate utah" keyphrase. However, there was one search for October activities in Utah and another for autumn activities in Utah. By that measure, it might make just as much sense to tell Diana that now she needs to go in and click the 4 seasons and the 12 months of the year. After all, there were twice as many visitor looking for seasonal information as for neighborhood info. And that would save her 13 clicks.

My client site that covers the Dilworth neighborhood in Charlotte doesn't have great rankings, but with Trellian telling me that there have been 244 searches on "Dilworth" and 4580 searches on variations of "Charlotte real estate", I still question the figures you got from Google. But right or wrong, I think the biggest issue may be that this change has created a lot of unnecessary work for P2 members as well as a lot of confusion with your neighborhood expert program.

Jeff Tomlin said...

You make a lot of good points. For your real estate Utah site, remember you won't see any traffic for search terms the site isn't optimized for, naturally.

For your example in Charlotte, it is clear that "Charlotte real estate" will generate more search terms than the neighborhood of Dilworth, but it's way more competitive to rank for that's for sure. The main point is that if you add up all the neighborhoods, you could generate more traffic than for the one or 2 main search terms.

Here are a couple more examples: Point2 Homes does not rank for the term "Las Vegas Real Estate" and is on the second page for "Las Vegas homes for sale" but for a selected period in September, it generated 2100 visits from 1500 different keyword combinations. We estimate half of which contained a neighborhood.

A second example: Point2 Homes does not rank high for "Charlotte real estate" or "Charlotte homes for sale" for the same period in September, it generated on 100 visits from those 2 terms because of the low SERP rankings, but it generated 2280 visits in total from 1887 different keyword combinations containing the term Charlotte. Of those searches, around 45% contained a neighborhood name. Examples of those searches would include:

- chantilly real estate charlotte
- copper ridge condos, charlotte, nc
- homes for sale park vista charlotte nc

Lastly, I know this has created work for our members but hopefully they only need to go in this one time and select their new selling areas. I also know we have created a ton of confusion with the neighborhood expert program for Point2 Homes and we are working hard to try and fix that.

Omar Cruz said...

I like this blog is fantastic, is really good written. Congratulation. Do you want to see something more? Read it...: Great investment opportunity in Costa Rica: houses for sale, homes for sale, investment property. Visit us for more info at: http://www.jaco-bay.com/