I have tried to remain neutral in the fallout over the WASA launch since I currently serve as Head of Research on the OPA Exco and much has been said about it already. I don’t want to thrash out the merits of having a second website representing body in South Africa nor nitpick over what chain of events followed, but I believe it is in the public interest to publish the first official response sent to OPA members this morning from Nielsen//NetRatings, who supply the stats backend and audit the rankings for the OPA, as it gives some good insight into how their system works and some of the perils in using different data sets for ranking websites.

Dear Member

In light of the past two weeks events concerning the OPA site stats and WASA, we thought it would be appropriate to let you know Nielsen Online’s stance on the matter.

Technically, trying to compare figures coming from two different data sets is by itself a source of mismatch, even though the method is the same, in this case, browser or tag based technology. Different analytics packages work on different algorithms and filters to correct bias from the data collection method.

In order to cope with the scenario where a browser has cookies turned off, the Nielsen Online system first identifies which records come from browsers with cookies disabled and then assigns the null cookie id to these records. The raw figure of unique visitors is then scaled by the proportion of records from browsers with cookies turned off to the proportion of records from browsers with cookies enabled. This is not the way that all provides necessarily will work.

Another point is that Nielsen online filters impressions by using our latest automated search BOT list. If the two providers don’t use the same list, the impression figures will be different.

We have always found in the past that mixing two data sets from different sources is problematic. As you well know, even though SiteCensus and Market Intelligence use the same code and algorithms the data is not 100% congruent. All other issues notwithstanding, from a purely technical standpoint, we do not advise you or condone the use of Nielsen stats along with any others.

Whilst we are not deliberately looking to block usage of our data we wish to maintain standards on a worldwide basis that are correct and methodologies that we are confident in. For this reason we do not agree to the process of comparing our metrics with any other bodies data.

I hope this gives some out thoughts on the potential issues that will arise using different data sets. If you would like to discuss any of the points above, please feel free to give me a call.

Kind Regards

[name omitted]

Nielsen Online

Not that I use Outlook any more, but I know enough people that do and have heard many complaints about syncing with other calendars systems such as Google Calendar.

Well, it seems that there is now a fix to the problem courtesy of a new desktop application developed by the Google Calendar team which acts as a “2-way synching application between Google Calendar and the calendar in Microsoft Outlook”. Great news if you’re an Outlook/Google Calendar user as you’ll seamlessly benefit from the best of online/offline calendering from wherever you are.

Now if I could only find a fix for that damn Outlook/iCal timezone bug…