Posted by Sam Battin, Senior Search Strategist
A couple months back, we took a look at Google’s new search results on Caffeine. The “Caffeine” engine represented Google’s experiment with a new way of looking at their own data, and was presented in a “sandbox” format separate from its standard results sets. For example, if you visited
You could enter Google queries, and in some cases you’d come across different results than you would find on Google.com. Performics checked out these listings and we found that Caffeine was faster and presented a broader set of query matches. We also found that the results were slightly different on Caffeine for retail keyword searches and some long-tail searches (e.g. searches with three or more words).
As of recently, the “sandbox” has closed, and the only message from Google that remains at the URL is that Caffeine will soon be ready for prime time. What does this mean for your site? It could mean a great deal, so read on.
Most sources on the Web are saying that the new update will be rolled out in January 2010. This should be a relief to most retailers, as they don’t want to have to deal with a bunch of search engine result changes during holiday.
Some sources are making predictions about the changes taking place. So far, the source with the most authority is Mr. Matt Cutts; he actually works for Google and is more or less their public voice about SEO. However, even if the word is from Mr. Cutts, it may not be accurate. He doesn’t run the company, he only works there. He’s the head of their Web Spam department, which means it’s his responsibility to figure out how people are trying to trick search engines and cut them off at the knees before they get to be too successful. He’s a good man to listen to, but he’ll be the first to admit he’s not the final word on what Google is or isn’t going to do.
Google is excellent at maintaining informational security; they will not knowingly release information that would allow people to reverse-engineer all or part of their algorithms. The accuracy and speed of Google results keep them at #1 in the search market, and if everyone knew how they did it, then Google would have a lot more competition.
That being said, the following are theories being developed about the upcoming release. I include them here, along with comments, only to give you an idea of what may be in store.
Rumor #1: Caffeine is Already Live
This may actually be sort of true. Google’s standard roll-out procedure for updates is to put them onto one data center and see how well they work. Last-minute bugs are fixed at the single data center, and then the updates are rolled out to all data centers. Google has multiple data centers located across America (and there’s no official report on how many there are), so it’s entirely possible that a couple of search queries you’ve made have returned Caffeine results. Most results, however, are from Google’s existing algorithms, and this will remain so until after the holidays.
Rumor #2: Domain Age Won’t Be Such a Big Deal Anymore
Separating spam from useful results is an important task for Google. One way to identify spam is by the domain age of the site. If a site’s been around for years, then by default it’s more trustworthy. If a site was registered one month ago with a one year contract, it’s probably not going to get high ranks. While domain age is a useful way of detecting potential spam and putting well-known sites in the top ranks, the result can be certain sites showing up again and again for a variety of similar searches. It’s possible that Google may attempt to correct for this through revising the effect of domain age. Site owners may soon find out whether it’s only their domain age that’s been keeping them in the top ranks.
Rumor #3: More Frequently Updated Sites Will Get Better Listings
In other words, if the data on your site is constantly being replaced with the latest information, you may get better ranks on Caffeine. Some sites have been up for years with little or no changes, and it’s not unreasonable to presume sites that are frequently updated (whether on the main pages or in forum pages) will have the latest and most useful information. Google has always had a preference for recently updated data, so it’s certainly possible they may have found a way to increase the effect of frequent updates on result pages. If Rumor #2 is true, then they’d need something to replace domain age anyway.
What Should You Do?
More important than the rumors floating around is what your response should be. In the waning days of 2009, you should absolutely get to know the words that are currently driving the most traffic to your site. If your site depends on search engine traffic, this should be a priority. The algorithm update is probably going to go wide in January, and so if your site is among those affected, you should know as soon as possible. With your Web analytics, determine what your top traffic driving keywords are, and how many of those visits are coming from Google.
Compare your high-traffic keywords of December 2009 with your high-traffic keywords of January 2010. If there’s a steep drop-off, then action is required. Look at the sites ranking in January 2010 for the keywords you used to rank for, and see what the difference is between their sites and yours. According to the rumors listed above, something to look at may be the activity level of your new competitors. Do they have forum pages? How often are they posting updates?
In your review of the competition, your primary presumption should be that your competitors’ sites are connecting better with your customers than yours. This is the driving force behind the Google updates; Google believes their algorithm revision will make their results more useful. So far, Google result pages demonstrate a good knowledge of what search engine users need, and they’re not likely to change their formula.
To compete more effectively on search engines in the new year, you will need to know even more about what makes your company different than others, and learn more about what your customers need from you on the Web.