Google Analytics Funnel Visualization Issue

Google Analytics has a visually appealing Funnel Visualization feature which is very useful for ecommerce applications and web systems especially for an ecommerce accounting system. It is much more visually appealing than Yahoo Web Analytics Scenario analysis feature.  But as we dig deeper into this tool we keep finding more and more problems. The latest problem we found is that when the rules of the Goal definitions are changed they dont operate on historical data but on new data only and that too after couple of hours. There is no way around this problem. Also as Avinash says: " if you set up a funnel today and a month later you remove a step, or add a step, then that change is only going forward. It is important to make a note of it." This basically makes this feature of the tool unusable for most realistic scnearios.

Google Analytics Site Overview issues: Subdomains and HTTPS

On the dimdim.com website you will see that the Orange button on the top takes one to a subdomain behind HTTPS. This means that when I see the Site Overlay for Dimdim on Google Analytics I don't see any information on clicks which took users to this crucial signup page. The surprising part is that this is a well known and documented problem. Google knows about this and has not solved the issue. I guess one of the ways one can win over Google is to make a product where the features are prioritized taking the customer feedback into account.

LifeTimeValue (LTV) calculation in a Freemium Model

In my role in Dimdim I have been working on making sure we can maximise the value of customers to the company while making sure that real value is being provided to all users. We adopted the Freemium model around the time it was put forward by Brad Feld without thinking too much about the impact it would have. I actually remember feeling a bit vindicated that we have chosen a promising path in Dimdim when Brad's blogpost came out. Lately this Freemium Model is being analysed (or fomalised) along with the slicing and dicing that VCs typically expect. Andrew Chen's blog on this is a good place to start off to understand this.  

There are two things that I don't agree with in Andrew's blog post:

Buying traffic (and hopefully customers) through ADs: I understand most startups fail not due to technical issues, or sales not closing but de to the word not reaching the customers. Startups fail die not getting enough customers. Internet ADs and buying clicks and traffic is seen as a way to avoid this. But there are much better ways. I believe the Customer Development Manifesto is an effort in the right direction. Having satisfied customers who can help a company avoid the Big Drop-Off is the best way.

The "Viral Growth Kicker": What does this mean? Does this mean a killer feature which is exactly what the customers want? Or a marketing campaign is wildly successful and sustainable? Is this just wishful thinking of expecting too much from a secret sauce? We should not expect that this parameter should contribute so much to a startup's success the way its factored in, in the LTV calculation in Andrew's blogpost.

What do you think?

I think there are also there are a bunch of complexities around Churn rate, Referral conversion rate and some other parameters. What about the experiments around pricing? They should also be a part of the model. Hope to address all these are more in the coming posts.

Awake ! Arise !

I have woken up from my period of inactivity. I see that the last post on this blog was in April 2006 (Couple of months since we started Dimdim). Its been more than 3 years. A lot of tings have happened in that time frame. I am happily married now. Dimdim has grown phenomenally. Have of things to talk about. You can expect to see regular posts on this blog and in the Dimdim blog my friends. :-)