i had the privilege of seeing you speak recently and i'm a new reader, forgive me if this info exists in a previous post. but i've noticed that much of your material (including this post) references multiple analytics packages such as clicktracks, omniture, etc. obviously GA is free and you are the 'evangelist', so you recommend GA to all.

what are the basic thoughts on using multiple packages at once – and pros/cons of having multiple strings of tracking code for conversions and full blown analytics packages? (any performance setbacks?)

is there a recommended max? for example is having 2 or 3 recommended whereas more than 3 overkill?

i ask because i have access to clicktracks but am not currently using it because we have about 5 other tracking tags going at the moment, GA included. feels like overkill… tough to get my techs to keep them all up to date… easy to under utilize all the data.

thanks for any input.

For starters it is not a "real" a/b test, more like A is we had nothing, B is that we now have the forum. Typically it would also not be prudent to show the link to the forum to half the people and not to the other half. :)

I would think of this less as a A/B test when trying to identify value.

It would be more like a classic value proposition exercise. What's the cost of doing this (probably two dollars because so much good forum software is free)? Software plus people overhead. What is the benefit of doing this? Measurements would be:

* Improved customer satisfaction.

* Lowered support costs.

* A unquantifiable improvement in people's perception of the company (or maybe quantified via a survey and the "likelihood to recommend company" question).

* Various interaction metrics. (How many people contribute vs the number of customers (or the number of Unique Visitors)).

* Growth in posts / pages / whatever.

* Sales on the eCom site from referrals in the forum (notice I put this last).

Things like that.

You give it a couple months. Go through this exercise. Then kill or invest more based on the cost benefit analysis.

Hope this helps a small bit.

-Avinash.

As an example, let's say I was thinking about adding a forum (or wiki or customer reviews or other user generated content) where prospects could post questions and view problems that had been resolved. The forum would not have much value in the beginning while there are few answered questions. So I imagine an A/B test would take a long time to show any effect from the forum. In fact, I think the A/B test would delay the value creation from the forum because fewer users would see it and potentially post.

Do you have thoughts on how to perform a test on this type of content?