is cookie tracking dead?

If users delete cookies is this tracking method reliable nowadays?

[jarboy]

Following reading this (http://www.clickz.com/experts/crm/analyze_data/article.php/3492951) and other articles that tell us ever bigger percentages of users are deleting (either intentionally or otherwise) cookies, are cookies viable as a tracking method?

I've read about flash based 'cookies' (http://www.markme.com/jd/archives/007257.cfm) that can be used to retrieve information from the user.

Does anyone have any opinions as to how sites will continue to get (relatively) accurate measures of returning visitors?

[Adam_T]

Well, i think it is definetly on the 'slow decline'. i think that these figures are likely to appear slightly exaggerated(?) and the actual figure is not as high as this, maybe around 10-15% i would estimate. Comapny's with stats tend to like to stir up shocking news i find, when the reality is it really isn't as bad as they make out.

I do however, believe that unique visitor figures are quite unreliable, and we do not depend on them in the slightest, as they can be quite deceiving. I mainly use cookies to track navigational data, how people are moving through my sites and possible reasons for this - then altering creative accordingly.I wouldn't rely upon them for unique user data. I would say that cookies are and will remain however a reliable medium for recording visitor navigation, but possiby less reliable for visitor numbers.

[jarboy]

Hi Adam_T

I agree regarding a slow decline - I was also wondering what it might mean for A/B/Multivariate testing.

I was wondering what other methods the analytics industry will come up with (if any!) to measure returning visitors to a site. I was at a conference last year and much play was made by some vendors of their superior (usually 1st part vs 3rd party) cookie tracking methods.

It will be interesting to see what new measurements they come up with :)

[bobdole]

For most folks cookie tracking is probably still the best option for tracking return visitor data. It really depends on your site and the type of users you get.

If your site recieves alot of traffic from government domains or you run a "paranoia" type site then cookies would be a poor choice. However if your site deals mostly with the avg computer user then cookies will still give you decent data.

If you use a forced authentication system you could set something up to identify a returning user vs a new sign up fairly easy, then use some sort of customized report/script to give you stats on returning visitors.

-b