I'm still gathering my thoughts about the recent Google Instant announcement, and how these changes will impact search tactics for our market. The following list describes some of the issues I think are important, and my reactions:
- Search volumes are going way up. Google says 5-7 fold from the former method of searching. To me, this says that keyword research is going to be a more fruitful process - i.e. the "long tail" search phrases will become more abundant - more target phrases, and more volume of searches for those phrases. Click through ratios, on the other hand, are going to drop. This is the new environment, but not all that much has changed fundamentally. The number of people searching shouldn't change substantially, just how the numbers come back to us.
- The "Three-off the key" rule is now in effect. Google is now saying that users cause a "search impression" any time you stop typing for more than three seconds. The fact you can do that mid-word is going to cause some problems -- both for keyword research and website analytics. Some will see this as an opportunity to optimize for partial words; others will see it as slow keyboarding and muddying the waters. [more discussion below].
- Contrary to some reports, you don't need to be signed into Google. Google Instant shows up on Google.com for everyone, and is moving to the following countries within the week: UK, France, Italy, Germany, Spain and Russia. Google.ca isn't on the immediate release list, but it likely won't be long.
- Google Instant is on by default. Yes, you can opt-out, but don't expect that to happen in any number. If web history tells us anything, when default settings are established by major players, they become the norm (See: Microsoft & Internet Explorer). The majority of users going forward will use Google Instant.
- Predictive Search is different than Google Instant. Predictive searching, a.k.a. autocomplete, has been going on for a while now. Instant search, and the per-keystroke impact on the visible webpage in front of users, is the change that must be measured.
- Google is creating more bid-worthy widgets for PPC. People won't stop at 2-3 word phrases anymore. Giving the 4-5 term search market an increased presence is good for both paid search and organic.
- Getting page-one results is more important (yet again). Ten percent of users have traditionally gone past page-one in the search results. I can't imagine that number doing anything but plummet. Why scroll pages when the search results change with every keystroke?
- Partial word searches may (or may not) have an impact. There's a fair bit of speculation online that optimizing web pages for partial words and phrases may evolve (e.g. "new jersey crim" for New Jersey criminal law phrases). So far, I'm not seeing this. Similar to optimizing for typos and spelling mistakes, I suspect this will have a minor impact. I fully admit it is possible, but after viewing the metrics for a number of high traffic websites today, partial keyword searches over the past 48 hours were almost non-existent. Impact? perhaps, but not a game changer.
- Will users change their search mid-stream based on results shown? Again, this would be noticeable by partial word searches in your website analytics. My suspicion is that users will finish typing the words they are thinking of, and ignore the page results until they've completed their thought; especially compared to the alternative, stopping mid-search and clicking.
- Partial brand names in the search results may become an issue. I'll leave this one to the trademark lawyers, but I'd have my eyes open for new websites that include part of a recognizable brand name. How close can you get to a trademark without causing confusion? I don't know, but Google may have added a new check point with GI.
- Google's suggested phrases are important. Every volume search phrase will have tier-two associated phrases that you must know for marketing purposes. Watch the suggested terms in predictive searching, along with the phrases embedded at the bottom of most Google search results pages. This was a noticeable trend with predictive searching, and should get stronger with Google Instant.
General thoughts on approach: It's still very early in this process to see all the fallout issues. And it's possible that website metrics may paint a different picture in a few weeks time. That said, the biggest change I'm seeing with Google Instant is in the user's experience. The organic search results haven't re-ordered in any critical way. For search marketers, that means reacquainting themselves with the new numbers, and then responding. I don't see a major overhaul in tactics.