Emetrics San Francisco: WebSideStory passed away, long live Visual Sciences!


Logo Visual Sciences

WebSideStory will no longer exist. Bob Chatham announced at Emetrics San Francisco a couple of weeks ago, that the nice brand WebSideStory would no longer exist, but don’t be afraid, the company hasn’t closed, it’s just that it changes its name to Visual Sciences, to better reflect the new positioning of the company.

bob_chatham.gifLast year, at Emetrics Washington Bob did already a very interesting presentation explaining the positioning of WSS after VS acquisition: going towards ‘Customer Analytics’. Back then, Bob explained how Web Analytics would enlarge it’s scope from simply Web analytics to Customer Analytics going through Internet Channel Analytics and Multi-Channel Analytics. I remember that this presentation was really inspiring; Aurélie and I were very excited with this new overture in Web Analytics and by Eric’s private demo of Visual Sciences. For those of you who aren’t familiar with Visual Sciences, you may read a couple of posts that Aurélie wrote following Emetrics Washington 2006:

Getting back to Bob’s presentation of last week, here are some of the points that got most of my attention:

First of all the analogy of Cubes vs Wheels: Bob explained that very often the limitations that we face are self imposed and he expressed this by the following 2 slides:

Cubes      Wheels

He then explained how this was related to Web analytics. According to Bob, most Web Analytics tools have been greatly influenced by legacy systems as BI or DW that work mainly as cubes, this implies a long process in order to bring Insights. According to Visual Sciences, this approach is not flexible enough when you have new questions as very often you need to go through the whole process again.

Opposing this ‘legacy vision’ he presented what he called the data wheels revolution, where the continuing processing lead to a real time analysis capabilities allowing real time analysis of data against any parameter (dimensions, metrics, …). This leads to real time analysis: the user can answer any question by querrying in seconds instead of days as it happened in the former model.

As data is randomly distributed around the wheel, this means that you can query part of this data and enlarge the scope of your analysis without truncation. This is interesting as truncation leads to the non-analysis of the long tail, a concept so present nowadays.


Back to Bob’s analogy, another of his principles is that cubes are not scalable. This happens as adding new dimensions can lead to exponential amounts of data that will be difficult to analyse (remember the former DW projects that took ages to implement and lead often to failure?).

According to Bob Chatham, in the Fortune 500 environment, answering a CXO question may lead to 10 days with ‘cubes’ but a few minutes with ‘wheels’ as you still need to discuss the outcome and the action to be taken 😉

Getting back to the announcement what does it imply and what is our opinion?


  • HBX product will continue to exist. Not only that, but it has been enhanced with it’s 4th version as the investment in the HBX platform has been increased. HBX integrates also to Publish, Bid and Search (complementary products of HBX).
  • Visual Site allows to go further and really integrate the whole Internet Channel (please note that data collected by HBX can afterwards be analyzed through VS).
  • Visual Call & Visual Mail allows to integrate the offline world and propose a multi-channel integration of analysis.
  • And for those who want real Customer Analytics Visual Sciences releases its Platform 5
  • Visual Sciences solutions may be On demand, Software of both
  • Visual Sciences releases new education and certification programs (we’ll keep you posted on this one ;-))

My opinion:

HBX is a great tool for most companies (as others in the market such as WebTrends or Omniture). Visual Sciences Platform 5 is for the moment according to what we have seen the best tool on earth, but unfortunately not all companies are ready for this kind of tool… It’s the Rolls-Royce of Web Analytics but you not only need a good analyst (internal or external) to get the best out of it, your company needs to be what we like to call a ‘data driven company‘ that has integrated accountability and is not afraid of testing and failure. Testing will inevitably lead to some results that won’t be optimal, the important thing is to constructively focus on this from a learning and improving perspective.

From OX2 we consider this to be a great move and I wanted to personnaly congratulate Jim MacIntyre and Bob Chatham for the great work they have done pushing the former WebSideStory from just a Web Analytics player to the new Customer Analytics arena. And I have to admit that I’ve been really impressed by the new marketing (I love the new logo), great work!

For those of you who didn’t attend the Emetrics conference, you may download the presentation in PDF (Thanks Bob for providing me with it) or download the new White Paper that has been released.

Have you attended the presentation, leave us you impressions and comments. You don’t agree with what it has been said? Let us know also as we always welcome constructive criticism (how could we evolve without it?).

And we will keep you posted on other Emetrics Insights 😉

Cheers from Brussels sunny Brussels,


2 Responses to “Emetrics San Francisco: WebSideStory passed away, long live Visual Sciences!”

  1. Martin Says:

    The wheels approach seems basically an improvement in speed for ad-hoc analysis if I understand correctly. VS is saying that wheels are more scalable for large data sets vs. cubes, i.e. the old cube based OLAP approach used by companies like Hyperion, and Cognos Powerplay. VS is saying that relational OLAP (e.g. MicroStrategy) is scalable to large data sets but not as fast for ad-hoc analysis as wheels.
    Now all of this is very interesting. But I wonder how relevant this ad-hoc analysis speed improvement is for web analytics? How many companies are really impeded to get value from web analytics because their ad-hoc analysis is not lightning fast? Is that the big problem that impedes web analysts today?

  2. Aurélie Pols Says:

    Hi Martin,

    Interesting question indeed as speed of ad-hoc analysis can be of essence, depending on the users’ level of analysis.
    Personnally, I spend a lot of time making annotations here in there and listing questions because I simply can not get to the data or because the rendering of the requested data is too slow. Try to export some reports in Excel with certain products and you’ve got time to get at least 2 coffees. Furthermore, sometimes you have ideas about a possible insight that you’d like to test out but once you get the data, there’s nothing you can actually say about it. So if you’ve got to wait for it, it gets frustrating after a while, to say the very least 😉

    But this of course depends upon the level of analysis of the company you’re working for, if it even exists in the first place as a next logical step following reporting.

    To be quite frank, rare are the companies that go that far so if they just need reports, fine the wheels don’t really matter. But if they intend to spend human resources efficiently on analysis, you might as well assure that accessing the data is as smooth as possible.

    There are other consequences to the non OLAP vision such as non trucation, which might be of value for certain (larger) sites.

    Overall I’d say nonetheless the biggest problems that web analysts face with today are: technical tagging and data integration issues due to lack of processes, lack of interest and understanding of the data by the business and most of all, lack of action following real insights found in the data.

    Installing a tool is good. Using it is better. Take action based upon found insights is the reason why it actually exists.

    Hope it helps

Comments are closed.

%d bloggers like this: