21 replies to this topic

[nethy]

http://googleblog.bl...-for-words.html

Optimising for spelling mistakes was always ugly.

[Orpheus Descending]

I wonder how this will change the way keyword queries are counted for tools like Word Tracker and Keyword Discovery? Also, I wonder how this will affect long tail searches? Will searchers automatically use the suggestion or will they continue to type their specific phrase? And most importantly, are the suggestions based on the keyword phrases of PPC campaigns? Therefore driving traffic to the more generic keyword phrases used by top tier advertisers and slowly eliminating the quirky or long tail phrases that some of us like to sprinkle in our copy? I smell uniformity and it makes my nose itch.

[nethy]

Also, I wonder how this will affect long tail searches? Will searchers automatically use the suggestion or will they continue to type their specific phrase? And most importantly, are the suggestions based on the keyword phrases of PPC campaigns? Therefore driving traffic to the more generic keyword phrases used by top tier advertisers and slowly eliminating the quirky or long tail phrases that some of us like to sprinkle in our copy? I smell uniformity and it makes my nose itch.

I imagine the effect on search volumes will be subtle. Not sure that the direction would be away from long tail (not sure that it's not either). Though it may move people away from searches with few results & away from searches that are what you say when you can't think of how to say it.

[incrediblehelp]

I think the effect will be bigger than most people think. Now end searchers will have a search bias towards whatever Google auto-complete shows them.

[Orpheus Descending]

hi incredihelp - that's exactly what I'm thinking too.

[Randy]

Spot on Jaan. I'm thinking the same thing. So I guess I'll have to make sure their auto complete thingee is actually giving them the appropropriate keyword phrases that I'm already targeting. I've already started doing that a bit for a couple of sites actually.

I wonder how many phrases that simply weren't worth going after before are going to be worth the effort now.

When I was checking this for one of my sites I noticed that I have one sort of 2nd level phrase for one of my sites that I'm going to have to start tracking just to see what happens with it goes live. Google's Keyword tool has said for ages that this particular phrase gets a shade over 12,000 searches per month. And adwords for the phrase is fairly competitive, seeing as the tool has it at a 6 of 10 on the little bid competition bar.

Not that I believe the 12,000 number for a moment mind you, seeing as how my site ranks #2 for the phrase and has gotten exactly 20 visits on the phrase in the past month, which is pretty typical. I'll grant you that I didn't try really, really hard to optimize for this particular phrase since my previous (non-Google) keyword research said it didn't get much traffic, but I did include it because it was easy and ended up with a good ranking. But you can't tell me that I wouldn't get more than 20 visits out of 12,000 searches by dumb luck, being ranked either #1 or #2 depending upon the day. This is one of the reasons I don't trust the numbers the Google Keyword spits out very much. The numbers just don't make much sense sometimes.

That said, what I'll be watching to see if is I actually start to get more traffic on this phrase now that it's one of the prime choices being offered in the Google Suggest box. I kinda hope so, but we'll see. If it happens it could sure make keyword research a lot easier!

<edit for a bit of clarificaton>
I don't mean to leave the impression that I see this disconnect in the numbers given by the Keyword Tool and Analytics data with just one phrase. It's not that at all. I can find multiple examples of the same sort of thing for every site I run. Lots of 'em, in fact.

It's just in this case one of those supposedly bigger traffic phrases that is totally non-performing according to my analytics happens to be one that they've included in the Suggest tool. It'll be interesting to see if the clickthru's on this one phrase suddenly rise. If it does skyrocket up into the hundreds each month it'll be an indication that people are probably using the Suggest tool, since I've never seen more than 30 visits on the phrase in a month before. With several years of data to play with.

Edited by Randy, 26 August 2008 - 10:36 AM.

[Orpheus Descending]

But Randy, won't it also make the keyword gene pool more shallow? And by default more competitive? So if a searcher types in 'holiday' and they are looking for a 'holiday home,' but they get derivations like 'holiday cruise,' 'holiday resort,' and 'holiday aboard' (which, by the way, are top tier CPC phrases - just for the sake of argument - i have not done the rearch) and so this fellow thinks 'Oh, 'holiday resort' - that must be what I should be looking for (because Google suggested it) and off he goes. My holiday homes site just went belly up because we didn't target the top 'holiday' phrases because they were so competitive and we get plenty of traffic from the lower level phrases, but now that Google net just caught all the fish before any of them made it down stream so there ain't nothin' for me to catch. I want to call it 'Google Puts Words In Your Mouth'!

Also, if you click on the options menu is that query impressionable for companies like KD? Or does it circumvent their data retrieval systems? I just get the feeling that Google is creating a monopoly by strangling out everything that doesn't have it's genises with Google. But you know me, I've been singing this song for a long, long time.

[incrediblehelp]

Right. Look at the keyword "new york...."

Then give you some popular other new york phrases like new york yankees, new york times, etc. Providing this to the end searcher drastically changes your searching habits/behaviors. You were looking for one thing and now Google can or has essentially offered you other suggestions to change your mind.

[Randy]

Well, it's all pretty speculative at this point. But if you're looking for opinions...

Will it change search habits? Maybe. But I don't think it'll change search habits with the initial search all that much. And I think it's going to end up depending quite a bit upon the individual search being conducted and possibly the searcher doing the searching.

I hate to compare, but Yahoo! search has had something similar for I think at least a year now and I've not seen any huge shift in my keyword phrase traffic numbers from Yahoo over that time. And yes I looked for it. That said, Yahoo's version seems to stick a lot closer to what you actually type in. Meaning the first word you type in probably isn't going to get you what you're looking for. They produce a short list and it gets more focused the more you type in.

I have a feeling it's probably going to end up being much like is is now. People will type in what they want. If they know exactly what they want they'll type in their long tail search phrase and ignore Google's suggestions. If they take Google's suggestion and don't find exactly what they're looking for in the first 10 or 20 sites that brings up, or if most of the top 10 are way off the mark, they'll try another search that is more focused. The more they try, the less they'll rely on the suggestions.

Google etal have always said their biggest hurdle is to try to infer what a user is really looking for when they type in a search phrase. When you try to shortcut this, where the whole phrase isn't typed in, this hurdle can become infinitely bigger. So they'll need to make a choice in how wide or narrow their suggestions are.

So I truly don't expect things to change much right away.

Now, to ask the question that nobody has yet...

How does all of this fit in with Personalization? Will they eventually start offering different Suggestions based upon the personalization group they've assigned a certain user to? Dunno, but if I was trying to make it a more valuable resource that's exactly what I'd be doing.

[incrediblehelp]

Now, to ask the question that nobody has yet...

How does all of this fit in with Personalization? Will they eventually start offering different Suggestions based upon the personalization group they've assigned a certain user to? Dunno, but if I was trying to make it a more valuable resource that's exactly what I'd be doing.

Ahhh good point. Sure why wouldnt they. Google has been saying for years that they want to tailor search for each user. In the end that is how you will provide the most "relevant" results right?

[Randy]

Well, if you want another guess...

Mine would be this is why they put Google Suggest out there in the first place Jaan. Determining user intent has always been one if the main issues confronting any type of search functionality. Heck, I run into it myself with my own Site Search schtuff, and I'm dealing with a limited number of documents.

And after all, those who used Suggest before it went Prime Time had to sign up for the service via Google Labs. So in theory all of those could have been working on some version of personalization.

[incrediblehelp]

True Randy.

One thing I don’t understand about Google Suggest is why does Google list the # of total results next to the auto complete word? Does this help/hurt/influence the end searchers reason to use Suggest?

[Randy]

Goodness only knows Jaan. The numbers they show certainly won't help Webmaster who want to optimize for those terms since it's not a true indicator of competition anyway. Though I'm sure many might view it that way, regardless of whether it's right or wrong.

Frankly, it's a silly thing to show since it means basically nothing.

On the other hand, if you want a more cynical opinion...

From Google's perspective it's better to show the total number of results than it is to show the total number of Adwords competitors or how much they're spending per click. Not much better mind you, but better.

[Disclaimer: No I've not bothered to look to see if the Suggest terms are those that have Adwords bidders or not. But it won't surprise me when someone does and finds that the majority are competitive Adwords phrases.]

[incrediblehelp]

[Disclaimer: No I've not bothered to look to see if the Suggest terms are those that have Adwords bidders or not. But it won't surprise me when someone does and finds that the majority are competitive Adwords phrases.]

No reason to check. The keywords that are listed there are obviously the most popular ones so I am sure a majority are being bid on.

[nethy]

No reason to check. The keywords that are listed there are obviously the most popular ones so I am sure a majority are being bid on.

I don't think that's true at all. It's the main niche phrases that get the most bids. IE main keyword phrases within niche topics. IE 'home loans' gets more then 'domestic lending products' but not more then 'home.'

Many of the most common search phrases get very little advertising because they're about 'how to make tea' or Willy Nelson's knickers or whatever.

Edit: made things a bit clearer

One thing I don’t understand about Google Suggest is why does Google list the # of total results next to the auto complete word? Does this help/hurt/influence the end searchers reason to use Suggest?

Strange huh, But of all things they also show this in the SERPs. Always have.

Makes sense for individual sites' internal search, where you have a danger of coming up empty.

Edited by nethy, 27 August 2008 - 12:53 AM.