Tuesday, October 20, 2009

Google Analytics new features: my take

It is now entrenched in tradition: Google Analytics makes announcements at the eMetrics Marketing Optimization Summit. So here's the latest, straight from the eMetrics floor where analytics evangelist Avinash Kaushik is addressing a jam packed room of enthusiastic web analysts!

Introducing Analytics Intelligence

The feature that excites me the most is something I discussed privately with Avinash a while back: once we spot a dip or spike in a metric, I shouldn't have to manually dig for the other metrics that are correlated with this significant change. The tool should be smart enough to point me to the other few metrics that have the highest correlation with this change.

Here's an example: conversion rate is significantly down. Ok... What does it really mean? Is there a spike in traffic? Or is there an issue in the process leading to conversion? In the former case, it means I'm bringing more unqualified traffic to the site. I can then review my Traffic Sources to easily pinpoint the issue. In the latter case, it means my conversion process as deteriorated or is failing at persuading people to convert.

Other features

There is also a host of new features (on top of the previously announced beta features):
  • Alerts! Once you have a good baseline, you want to be alerted of outliers: values significantly going beyond the historical trend. You can easily set up daily, weekly and monthly alerts.
  • Multiple custom metrics: adding "custom attributes" to visitors/visits is an essential feature of enterprise-level web analytics solutions (but I wish there was also custom metrics at the page level Can have custom metrics at page, visit, visitor level - without limitations!)
  • Multiple-rule filtering: now you can more easily filter reports by building multi-criteria on-the-fly filtering rules.
  • Goals, more goals! (20) of them instead of four.
  • Segment unique visitors: something that should have been there from the beginning
  • Mobile analytics
More info at the Google Analytics official blog.

My take

Google keeps adding new features that challenges the traditional web analytics vendors. I have witnessed exactly the same phenomena in the high-end 3D animation/special effects software industry. When you are an industry leader enjoying high margin on product and services, you accordingly spend a lot of money on R&D. However, when you reach a certain feature set level, adding new innovations becomes exponentially expensive. At the same time, lower (and free!) tools can easily look at the top-line products and replicate and improve upon those features. Basically, the industry is leveling itself off, becoming a commodity: data collection & storage costs are being eliminated, intelligence and automations are being built in. This is obvious  when a competitor you dismissed as being "irrelevant" can relatively easily replicate your unique value and offer a much cheaper (and free!) product. Even more if the true value of a product is being subsidized by other sources of revenues (ads in the case of Google).

The only alternative becomes to expand horizontally. Develop a suite of products. And this is exactly what Omniture has been doing over the past few years. Once it's done, or when you can't financially sustain the model, the way out is to move higher in the food chain... Adobe's acquisition of Omniture for example.

What does it mean for the web analytics industry?

As a member of the Web Analytics Association board of directors, any major shakedown in the industry raises some interesting questions. It is clear to me web analytics industry as we know it today won't exist in two or three years from now. Commodization of web analytics is not inherently a bad thing and is a demonstration of a maturing industry. To me, the path clearly points toward analytics, analysis: business analysis and process optimization, or what some call "business intelligence". On the way to the future, Google mammoth is likely to put several players to extinction...