Monday, March 17, 2008

Confessions of a Web Analytics Manager

1. Its a mess out here.
2. Numbers don't add up and therefore the numbers are not accurate
3. Where do I belong? I'm having an identity crisis, I don't know if I 'm in marketing or technology
4. What is a visit & a pageview.? How is this different from a unique visitor?
5. How important is a click?
6. Why do I have to make the same reports again and again
7. What is my contribution to the business?
8. Am I a cost center or a profit center?
9. Should I switch over to a new Web Analytic vendor, i think the existing one is not good?
10.Should I pay or get free Web Analytic software?

There are many unanswered question for an budding webanalytics manager. The usual problems in his life will center around The server logs not matching the web analytic
results. He would frequently have to encounter Tech complains that the web analytic
codes are increasing the pagedownload time (There may be javascript error due to these measurement codes). And best of all, the support center may follow US timings and resolution is at least 24 hours away! What ever may be the reason, the position is not very easy, answer lies in the cliche- persistence and determination.

Simply put, a Web Analytics is a set of measurement software tools or hosted online service that records and measures the actions of visitors to a website. The existence of this tool lies in the fact that if you can measure, you can therefore improve. It emanates from the fact that since internet is highly measurable, so there should be a tool. And there is also the flaw- since there are no standard definitions and hence there are multiple interpretations. Case in point, the online publisher might give you a set of numbers affecting your marketing campaign, however, the web analytic tool might give you another picture and these numbers won't be necessarily same. The server side logs will indicate figures for a download, which might be 25% in sync with the Pageview report from the web analytic solution. This is not just the end of woes, there are hardly any marketing matrix available that can provide benchmarks eg how many pageviews are required to increase the visitations of the website by 10%. The Web Analytics manager will swing betweeen the Yin and Yang and discover the true dualities of this existential world.

The truth and lets be honest- most of us don't understand Web Analytics and it is still to be ingrained in the planning process in India at least. I have hardly come across people who judicious use this platform to define goal and not just for number reporting and diagnostics.

The origin of Web Analytics, dates a few years back, when the server side logs were used to identify the call made (and hence the relative popularity) to the server, which is called the log file analysis. This coined the word 'Hits' or the number of times call was made to the server. However of you had 4 images, stored on the server, you will get 4 hits on the website (and hence not accurate)which will be equivalent to a page download in the current context. An improvement on that was the server side web analyser, a good number of these tools is available on the net. One had to run these logs as a batch file in the analyser and you would get the relevant details. There is huge amount of data in these logs and has to be analysed periodically to check the web analytics data accuracy. The only limitation of these logs is that, you can conclusively track an event (defined as a goal) and study the possible cause for this event. There is also proxy/caching inaccuracies.

The method which is gaining popularity is the client side data collection systems, which involves coding each page on the site, which is usually a single snippet (tag) of the code referencing a separate JavaScript file. This is cookie based and data is collected from visitors web browsers. There are tins of advantages in using this method...

1. More accurate session tracking
2. More accurate information of the browser (what type, screen resolution, javascript version, windows platform etc)
3. Can track events (say payment and its entry page along with acquisition source) and hence a great tool for eCommerce
4. Real Time data, no procession of the information required.

The limitations is due to the cookie and its privacy (first party/third party). Most of the popular Web Analytic tool, like Omniture, Corematrics, Webtrends & Websidestory (now a part of Omniture) use this method. Omniture is far superior tool, but is cost effective only for large websites. Webtrends is great for content sites and Hitbox (a product of Websidestory and now a part of Omniture stable) is great for eCommerce business.

By and large the features offered are the same. The criterion that you should use for choosing a web analytic tool should be
1. Cost per event (you buy a block of events)
2. Service agreements, it is very critical, as all the support system is sitting in US.
3. Education and consultancy. You should make this a part of the contract, as it will a good platform for the management and stakeholders who sample the report
4. Implementation, 80% of the success of WA tool rests on thorough implementation and
testing.

The cost should range from 4-6 cents CPM (please negotiate hard before finalising the true Indian way with the firangies). Before anything, you should be ready with the standard Webanalytic definitions. WAA, has a a document which is available at http://www.webanalyticsassociation.org. It is a great site with a number of articles, but few chosen articles are only for paid member subscription.

There are a number of Web Analytics Solutions available in the market place today.

1. Omniture: (Clearly the leader in the market place). They have recently acquired Websidestory. In India, Futurebazaar is using the flagship product- Site Catalyst.

2. Corematrics: Have JC penny and circuitcity.com as one of their clients.

3. Webtrends: Though their site mentions Microsoft as one of of their clients, I still have to come across widespread use of Webtrends. It is a great product and is quite popular.

4. Visual Sciences (erstwhile Websidestory and now a part of Omniture): Airtel, MakeMyTrip and ICICIBank is using the flagship product, Hitbox (which is now Site Catalyst).

5. DC Storm: This is DGM's main technology platform. Please connect with them if you are considering DC Storm.

6. Clicktracs: Tribal Fusion Network and Pioneer use this for Web Analytics.

If you want to check out the web analytics platform for the website you are visiting, you can use this great tool. All you have to do is to submit the url into the browser and it will find the Web Analytic application.

There are other free tool available in the marketplace, the most popular is Google Analytics. It is a great tool, however, the data is not available realtime. There are other few products also...

1. AWstats: awstats.sourceforge.net
2. W3Perl: www.w3perl.com
3. Webalizer: www.mrunix.net/webalizer/

These analyse logs and are processed batchwise to give results.

Web Analytics have to go a long way in India. Though Omniture has opened shops, it is still to get aggressive with business development. The support center sits out of Hongkong and US. Like I said earlier, one critical stumbling block is the implementation and support, which will determine the success of this tool.

This is vast and interesting subject and very new in Indian context. Though we at Makemytrip have been using web analytics in India in some way through PPC and Impression and click tracking, however, the real utility is when Web Analytics will be seen when it is able to establish itself as a strategic tool. I hope I can share more on this subject in the near future ;).

Till then you can always look up Avinash Kaushik's blog. He's an institution on webanalytics and is the writer of Web Analytics an Hour a day. Currently, he's Analytics Evangelist for Google.

Believe me Web analytics is that simple. Just requires an hour a day!!

Cheers!



AddThis Social Bookmark Button




Digg!


Add to Technorati Favorites

Add to My Yahoo!

No comments: