Can google track visit length without analytics or toolbar data?

[ganeshgrowth]

Can google track visitor's visit length in the absence of

1. Google Toolbar
2. Google Analytics code

[Receptional Andy]

Tracking visit length requires knowing both the start and endpoint of a particular visit. Clearly, the toolbar is the ideal way to do this, since it tracks cross-site (otherwise, you can't easily tell when the visit ends, other than if there is simply no further activity on a site). Analytics is slightly less reliable.

Of course, there are other methods - if a visitor clicks a search result and then goes back to results, Google can clearly know the time between those two events, which is likely close to visit length in many cases.

Is their a particular reason you're interested in finding this out? Personally, I've seen little evidence that this type of metric has a direct or particularly noticeable effect on search results.

[ganeshgrowth]

Of course, there are other methods - if a visitor clicks a search result and then goes back to results, Google can clearly know the time between those two events, which is likely close to visit length in many cases.

So, if there's no toolbar or analytics installed making an analysis about the website will be next to impossible for google. Then in that case it is a good practice from an SEO point of view to either install Google Analytics and/or toolbar.

[darkyl]

I agree with Andy, lenght of visits and bounce rate could both be manipulated to improve your results or to damage other sites, so I don't thins SE's are using this data significantly to evaluate a site.

[ganeshgrowth]

Personally, I've seen little evidence that this type of metric has a direct or particularly noticeable effect on search results.

Visit's user experience is a very important factor that google would like to analyze. Only based on that it could predict if the website is relevant or not for the search query the visitor is typing.

[brinked]

I agree with Andy 100%. I have always believed this. Why wouldn't google track visitors length? Google should be using this more...forget backlinks, if 90%+ of visitors go to a result and do not return to that result page, odds are they found what they were looking for meaning that page was more valuable than the ones above it and therefore should be ranked higher. If I owned a search engine, this is certainly something I would do

[darkyl]

This is what Matt Cutts said in December 2008 about bounce rate:

"... I'll just say that bounce rates would be not only
spammable but noisy. A search industry person recently sent me some questions about how bounce rate is done at Google and I was like "Dude, I have no idea about any things like bounce rate. Why don't you talk to this nice Google Analytics evangelist who knows about things like bounce rate?" I just don't even run into people talking about this in my day-to-day life."

From this post I understand that bounce rate and similar data is too easy to manipulate and that isn't even a factor that is currently being taken into accounts by the folks at Google Search.

[ganeshgrowth]

From this post I understand that bounce rate and similar data is too easy to manipulate and that isn't even a factor that is currently being taken into accounts by the folks at Google Search.

I do not think one can very easily manipulate that. Google can easily keep a check on the finger print of the visitor. I think it is tough to manipulate rather than to do this things appropriately.

[brinked]

I am not saying bounce rate should be the most important factor..but it should be one of many factors if it isnt already

[tedster]

One method Google can use to assess visitor data in the absence of the Toolbar or Analytics is by purchasing data from ISPs. They do sell it for the right price.

With regard to bounce rate - I'm pretty sure it's not a direct metric in the algo, for the reasons already mentioned and more. But that doesn't mean bounce rate isn't a good metric for the websmaster to use. By lowering your bounce rate you are often setting off a cascade of improvements that can be used in the algo. One obvious example - a visitor who bounces right out is not as likely to give that site a voluntary link.

[Receptional Andy]

To add to Ted's comment and also the remark about visit length/bounce rate being noisy, from Google's point of view, what's so bad about a bounce? The aim is to find the information the visitor looked for as quickly and as efficiently as possible.

So, if I search [what's the time in london?] my question doesn't require multiple page views. There's a whole category of searches where if Google do their job well, there won't be a long visit length, and there's a high chance of a bounce. I think it's somewhat too simplistic to view this solely as a negative signal - hence "noise".

[dstiles]

Why does google need external software to analyse time between going to a web site and returning to google results? They have your IP. Surely that's close enough for jazz. It's how most site stats analysers work.

That, of course, isn't all of it anyway, but it's unlikely they could easily track time-in and time-out of a site to any useful accuracy. Even the on-site stats analysers can't do more than guess.

I don't believe bounce rates should be included anyway. It's a dumb idea, as the google guy says: very noisy and spammable.

If I find a site from SERPs it may contain a useful link to another site - happens often here when I'm stuck on a tech problem. I go off to that site, waste a few minutes or hours wandering through other linked sites and assimilating information, find none of the solutions work and go back to google for more, perhaps trawling through the 50 results already gathered on a single page before doing another search or maybe refining the search there and then in the light of knowledge gleaned on the way. Maybe in the meantime I've had an hour off for lunch, gone to a meeting, fallen asleep, whatever. Toolbar or Analytics may show up some of this but it would be very inaccurate.

And then there are speed readers - take in a whole page in seconds, make an order whatever and out again. I'm a medium speed reader - could take me five times longer to read the page but I'm a good typist so could fill in the forms quicker. Some people I know could take ten or twenty times longer to read a page - or as Andy says, all they want is a single word on the landing page.