Over the last few months we have seen Google innovating at an unusually high rate. We are going to be in for some interesting times. Nothing ever stands still when it comes to marketing on the internet, but it is all too easy to miss the signs of fundamental upcoming changes.
I believe the signs are there in three areas:
1. local search changes 2. pay per click (PPC) changes and tests 3. fundamental changes to Google's infrastructure (the so-called 'Caffeine' update)
## Executive Summary
I believe that the Caffeine update is a sign that big things are afoot. There is relatively little change to be seen in the results so far and while I'm sure there are incremental internal benefits in server efficiency and code maintenance, I think there is more to it than that. I think they are seeking improvements that will enable the faster roll-out of future wide-reaching changes. The changes we have seen in local search and the AdWords display layer are indicative to me of the kind of tests that I would feed into the fundamental re-write of underlying systems apparently taking place under the Caffeine project.
There is a case to be made that this is all due to Bing's entry into the market but it could simply be the pressures from Wall St for continued earnings growth. Either way, it's the trends, not their causes, that are most important from an online marketing strategy perspective.
But now onto the detail of what we are seeing:
## Local search changes
Local search has been a little bit broken for some time. Tom has written previously on SEOmoz about some of the issues facing legitimate businesses attempting to market themselves in Google Local. He talked about problems ranging from the prevalence of spam through language issues for international businesses to the difficulties of managing large numbers of physical locations (such as a hotel chain might face). At the time, we recommended that business owners don't rely on bulk upload unless you absolutely have to. As Tom put it:
> Given the ability of webmasters to edit local listings and given the relatively untrusted nature of the bulk upload, I urge you to try and register individually the locations/properties you want in Google Local.
This should resolve one of our biggest criticisms of the way that local listings were handled which was that unless you verified each listing manually (via the "PIN on a postcard" method), the data you provided to Google was treated as untrusted. This meant that other users could edit your information. This was a big source of spam and also a very difficult process to manage across large sets of physical locations.
Here is the direct link to verify your bulk upload.
Any trend towards giving business owners (and their agencies) more control over how their business is presented is a good thing for legitimate businesses that have been stumped, up to now, by the lack of flexibility in the system.
## Pay Per Click (PPC) changes and tests
Google in particular seem to be testing things constantly around AdWords. At the display level, they recently moved the adverts closer to the search results. A small change, perhaps, but one that appears to be having an impact on user behaviour.
At a more fundamental level, they have been experimenting with possibly moving into the lead generation business (although so far they mainly seem to have hit legal issues), displaying "sitelinks" in paid search and including universal results in paid search.
I was wondering if a next step might be to allow simple forms to be included in AdWords ads - e.g. newsletter sign-up directly on the search results page?
These kinds of test fascinate me both because I am fundamentally interested in tests and testing, but also because the basic text ad has been the most effective form of online advertising for as long as we have been managing clients' campaigns. I am intrigued by whether it's possible to do (significantly) better.
I also wonder what the wider effects will be on how campaigns are sold and managed. Do changes like this make the field flatter or make it harder for new or small businesses to break in?
## The Caffeine update
I think that there are a number of incentives for Google to refactor their code (e.g. greater efficiency and maintainability) but that the biggest business case for what must have been a fairly major internal project will be an ability to roll out changes like those described above more quickly in the future.
The announcement says that this is mainly a rewrite of the underlying technical architecture rather than an attempt to change the results deliberately. There do, however, appear to be some results changes and I have some speculation of my own:
What follows is essentially guesswork at the moment, but I thought it would be fun to throw out some of our theories about deliberate and accidental changes as a result of the Caffeine update:
- We have occasionally noticed that the "suggest" results that appear in a drop-down below the main search box as you start typing have been fresher on Caffeine than on regular Google. We have seen, for example, references to current news stories in the Caffeine suggest box but not in the main Google one. This would support a widely-held belief that Caffeine has increased the reliance on fresh and time-sensitive data. It would also seem to indicate bigger peaks in search volume for breaking stories and bigger potential reputational issues for the individuals and companies concerned. The fact that we are struggling to reproduce this effect at the moment, however, could easily mean it was a data-centre artifact. - I feel that there is slightly less reliance on exact-match anchor text links in Caffeine than in the current search results. One example where this can be seen is in the rankings for the word caffeine itself (compared to the same search on the new system!). The sandbox page allowing people to test the caffeine update doesn't mention the word 'caffeine' anywhere on it and so the fact it ranks at all is entirely due to inbound links. The sandbox page ranks 3rd in the current live Google.com and 5th in the new version supporting my anecdotal belief that this could be true.
## In other news...
We have recently begun running conference calls on which we share insights and things we have learnt about. If you'd like to hear when we're running future ones, you can sign up here.
## Please tell your friends
If you find our conference calls useful, we'd love it if you would tell people about them - you can point them at the sign up page - they will not only be told about future calls but will also get to download videos and notes from previous calls like this:
You can also [download the slides](/files/seomoz-tools.pdf) (4Mb PDF).