Every now and again you hear people talk about the inaccuracies of cookies, but most of the time level of extent of this inaccuracy is not really known.

Everything from campaign reach and frequency to monitoring a campaigns contribution to conversion rely on cookies to work their magic. Sometimes these don’t get taken with a grain of salt and can be misinterpreted - probably resulting in less credit to the channels that kick off the process and some good news, you're conversion rate of "real" user is higher than reported.

With users blocking cookies, clearing at end of session, deleting more regularly, using different browsers, using multiple computers, netbooks, mobiles, public pcs, and spending more time researching touching more online media points, little bits of attrition in the tracking integrity here and there add up.

We wanted to find out what the deal was, so hunting around for answers we found this great little exercise. Angie Brown from showmeanalytics.com explains how to estimate the impact of cookie deletion, and use it as a model that can be extended to some of the other problems that deleted cookies and lost identities have on data analysis. We did it with a site with less frequency than the example, and got 2.1 x inflated unique visitors than real users.

http://showmeanalytics.com/2009/04/calculating-the-effects-of-cookie-deletion/

An interesting point in Angie’s post is how different adoption rates of authentication can help rebuild a profile of an individual and tighten the accuracy of your data collection. When we think about authentication this doesn’t have to be a login. By encoding the users unique id on communications such as emails and MMS links, you will also help tie your customer intelligence back together.

I’d like to see analytics suites focus more on a view of the individual, rather than a view of the device. Does anyone know a service that can do this well?

Do you think analysts should adjust results to something more accurate to accommodate for its flaws?