Saturday 9 June, 2007

Starting Web Analyics - Part 1

How to start collecting Web Analytic metrics?
By - Akshay Ranganath


Introduction
The following tutorial is a short primer on how to start collecting the analytics on your website. The goal of this article is explain what is Web Analytics, identify why it is necessary and explain how to start an implementation. The key take-away is help the audience to get a basic understanding and use of Analytics.

No individual product is examined or recommended at this stage. The following diagram explains the basic strategy.



What is Web Analytics and KPIs?
According to Wikipedia (1), “Web Analytics is the study of website visitors”. Using analytics, you can see what the users to the web site are doing, where are they coming from and are they performing the actions that you were expecting them to do.

The question that comes into the mind at this point of time is, “Why do I want to collecting these details?” Well, the answer varies from website to website but, the basic reason is to measure the performance of your web-site against some objectives that you had in mind when you created the site. Some of the possible objectives could be:

  • Are people coming to my site whenever they search “Book” as I am selling books online?
  • Is the new ad that I put on Google search working fine? Is it attracting anyone at all?
  • I just re-designed my site to show the new globally-local image. Is it a hit or are people still vary to use my services?
  • And so on..




These measures are called as Key Performance Indicators (KPIs). According to Wikipedia (2), “key Performance Indicators (KPI) are financial and non-financial metrics used to quantify objectives to reflect strategic performance of an organization. KPIs are used in Business Intelligence to assess the present state of the business and to prescribe a course of action.”

So, Web Analytics is used to track various defined KPIs for a web site. The primary aim of any organization is to define the KPIs correctly. If you don't know what you want to measure, you'll probably end up measuring what you never wanted. (Something akin to Cheshire cat and Alice in Wonderland!)


How do I start Web Analytics?
If you are a blogger or a an enterprise embarking on the world of web analytics, the easiest way to start is to follow the three step approach:
  • Start collecting basic metrics on your site
  • Brain storm and identify the KPIs for your web site
  • Tweak the implementation and focus on the reports that delivers value to you

Let me explain each of the steps in detail.



Start collecting metrics
The first step in analytics is to first “go and get the hands dirty” concept i.e., start an implementation on the site without too much of brain work. Let the brain work go on in parallel.

The reason for this type of an approach is that most often, we don't know how our site is performing. For example, when I had implemented analytics on my blog, I just wanted to see where people are coming from geographically. Instead I found some interesting behaviour where Google was directing users when they had searched some terms like “VIT” and “Cognizant campus”. I had never imagined that someone would land on my blog via Google since that had never been my intention.

In a similar way, the basic implementation will throw up some real surprises to you too. For example, people may be landing on your site due to a keyword search on Google but, that keyword would be a happenstance and not related to what you do. Or better yet, something that you'd never imagined that people would be interested in!

This basic implementation will also give you a starting point to examine and brain storm in identifying the metrics that should be measured on the site.

If you are simply trying to implement analytics for a blog, then try a free product like Google Analytics. If you are an enterprise, the product to choose is a bit more tough (and beyond the scope of this article). Whatever the product you choose, as a first step, go with a very basic implementation – something akin to out-of-box usage.

Brainstorm and identify KPIs for your Web Site
Once you've jumped into the ocean (so to speak), you need to dedicate energy into identify what is it that you are trying to measure. This is the core of analytics and will need a lot of elaboration. At a high level, this is what you should do:
  • Check what is your business
    • are you a retailer
    • an online book seller
    • a news publishing company, switching to online media
  • Identify why you are on the web
    • do you intend to generate leads on the web
    • do you want to sell more online
    • do you want to create interest for your shop
  • Correlate with the findings in the basic reports obtained earlier
    • are people coming to your site because they wanted to buy something
    • are people coming in for something that you don't do
    • are they guided correctly from your keyword ads on Google/Yahoo and so on

This is an area that is very specific to the area of business. The main aim here is to identify why you are online and if the web site's performance aligned with the strategy of the core business. Once this is determined, you can decide on the following factors:
In my line of business, what should I start to measure
Looking at the performance, is my site fulfilling its role or are people being misguided

Once this step is complete, jump into the third and continuing phase – Optimizing.



Optimizing
This is a stage where you tweak your analytics implementation, key word marketing, email marketing or SEO results to better suit the KPIs that are relevant to the business. For example, if the above analysis shows that people are hitting your site for a wrong keyword, use a SEO technique to ensure that your site appears for the more relevant keywords.

The ability to optimize the implementation of Analytics is dependent on the product being used. For example, Google Analytics is out-of-box while Omniture SiteCatalyst offers many ares of tweaks. Understand your product and implement the tweaks.

Conclusion
The stage of brainstorming and optimizing is a continuous one. Never get too complacent about your site's performance. Keep measuring and keep improving!

The article is intended as just a thousand feet high view on the process of analytics. A lot of area has been covered in here and over a period of time, I'll be building on the concepts introduced here.

References:
  1. Web Analytics From Wikipedia.
  2. Key Performance Indicators from Wikipedia.