@FIRST_LIGHT

There is a great post by Josh Kopelman, Managing Director of First Round Capital over on his Redeye VC blog.

He talks about how he analyzes data to evaluate pitches by web startups. It’s a great tutorial in how to cut through the hype fog and get and understand what’s really going on with a web property.

There is a great section on the concept of “virality” as follows:

Virality — So many people misunderstand virality. Virality is not “word of mouth”. And having a product go viral is not easy — nor is it something you can just “sprinkle on a product” after creating it. If making a product viral was as easy as adding a “share with your friends” button, there would be no reason for the $100 Billion advertising industry. (I can see companies asking themselves — “let’s see, should we spend millions on advertising…or should we just add virality…Hmmm”). I believe a viral product is one where a consumer’s basic usage of a site/product brings new users (and therefore additional utility) to the site/product. Facebook, LinkedIn and Paypal are all great examples of viral products. If you’re pitching your business, you should know your viral coefficient. That is, how many new users get added virally from each additional user. And if you can get your viral coefficient greater than 1.0, then you’ve built something really special.

Josh also touches on the topic of how to measure engagement, which is a really important concept. This is an area where I would love to see some innovation in Drupal. Let me explain.

Google Analytics can tell you how many people have come to any page, but translating that into engagement information is hard. For any given visitor or piece of content, be it a blog post, or a photo, or a video, or a wiki entry, what I want to know is the level of engagement - depth and intensity of multiple interactions viewed as a whole within an experience.

As a crude example of how this might work, a visitor who just navigates to a page where the content is displayed gets a score of 1. A visitor who does that, but also rates the content gets a score of 2. A visitor who reads and rates, but also comments gets a score of 3. And a visitor who does all that and then clicks on “related links” or does a search for related items (in the taxonomy), gets an even higher score.

I want to be able to track the total engagement score for a visitor, their individual visits, and pieces of content across visitors/visits. Then I want to figure out where my most engaged visitors come from, and think hard about how I can get more of them. I also want to know what my most engaging content is, and figure out how I can generate more like it.

It’s unlikely that Google is ever going to be able to provide what I described above, because their unit of observation is a page, there is no intelligence across pages, let alone the items on a page . What I’m describing above is a set of behaviors that spans pages. I think it’s really about a node and the operations on and relating to that node — viewed as a unit. That requires application/framework awareness at the Drupal level, not the page level.

So, what do you think? Is this something worth doing? Is it possible? Has anyone done anything like it out there in the Drupalsphere? I would love to get your thoughts.