Mar
22

Definition and history of Web analytics

Article read 3230 times in categories Web Analytics

The web analytics market in France and more widely in Europe is developing. New consulting companies were launched and accompany more and more companies, whether or not pure player. The work of evangelization begins to be gainful and more and more companies in Europe are now aware of the web analytics’ value added, whatever the business model(s) of their website. They only remains to take that step. In north America, the web analytics market is already mature. The firm Forrester also forecasts that market value will reach $ 1 billion just for the United States in 2014,  that said the potential in France and Europe.

Faced with this enthusiasm, I think it’s time to step back and relate the key stages in the evolution of the discipline since the start of the industry by Webtrends in 1993 until today. I found others articles about this subject but they are either outdated or not quite complete. I think this will help us to better understand what is happening and anticipate what will happen this year.

This article includes technical passages explaining the basic features of web analytics’ tools. These passages are simple and understandable by the less technical of you.

It’s composed of six parts  that are accessed directly through the following links:

- Definition of web analytics

- Web analytics, what is it ?

- Back on the main advances of the discipline from 1993 to today

- Focus on 1993 and the beginning of web analytics through log files’ analysis

- Simplified schema of data collection via log file

- Excerpt from a log file

- The year 1997 and the arrival of data collection method via javascript tags

- Simplified schema of data collection via javascript tag

- Development’s timeline of the majors web analytics’ solutions from 1993 to today

- Analysis of this evolution

Definition of web analytics

The web analytics is the measurement, the analysis and the reporting of users behavior in order to understand and optimize the use of your website.

Web analytics, what is it?

Attract, convert and gain loyalty - The three main domains of e-marketing

The web analytics aims to increase the profitability of your website.

For years, traffic acquisition is the field of e-marketing the most popular at the expense of conversion and retention. Too many enterprises forget that their company website does not live through traffic acquisition but with its monetization and therefore with the visitors that the site managed to convert into customers. The conversion rate is 55% in physical stores, that is to say that more than one in two people entering a store buys. This rate falls to 1.8% on the Internet for e-commerce websites. It’s very little and this leaves considerable room for improvement.

Concretely, the web analytics comes down to the collection, the analysis and the implementation of recommendations based on data collected by a web analytics’ tool.

Back on the main advances of the discipline from 1993 to today

The data collection can be done via two methods:

- Logs’ analysis of web servers

- The use of JavaScript tags

Here is a timeline showing the main stages in the evolution of web analytics:

Evolution of web analytics

You will find this timeline and the one after in this pdf file.


1993:  Start of web analytics, three years after the creation of the HTTP  protocol and the beginning of the Internet in public

The web analytics is born through the HTTP protocol which has the particularity of recording each exchange between the visitor and the website in a log file. Thanks to these data the analysis of user behaviour has began.

A protocol is a set of rules defined for a type of communication. If you want to find a comparable in our everyday life, the protocol used to talk to another person could be the language, there would be a protocol french, english or canadian french  (without offend you dear friends across the Atlantic, some of your expressions require an explanation :) ).

The HTTP protocol is the protocol that we use every day to view  web pages. We unconsciously use it in the address bar of our browser, when we want to visit a site:

Location of HTTP and WWW in a standard Internet address

Location of HTTP and WWW in a standard Internet address

You probably know other protocols like HTTPS which is a secure version of HTTP or FTP for example.

The HTTP protocol was invented in 1990 by Tim Berners-Lee with HTML to create the World Wide Web (which is the meaning of the acronym www visible in the majority of urls as well).

Here is a simplified schema representing data collection via log files:

Schema of data collection via log file

You will find this shema and the other one about javascript tags in this pdf file.

Here is an extract from a log file:

Extract from a log file

Extract from a log file

The first work of measurement tools based on logs analysis is to assign each requests (each line of a log file corresponds to a request from the browser) to the right visitor. This allow to replenish each visits.

These requests are called hits. Logs analysis was an adapted solution for the beginning of the web (from 1990 to 1996) when pages were statics pages composed almost exclusively of texts and links. A request or a hit was therefore equivalent to a page. Pages are now richer with the arrival of images, videos, animations, style sheets and  technologies like Ajax or Flex. Since each hit corresponds to a request of a page element, the number of hits measure now more the complexity of the page than its actual popularity.

The number of hit is now useful only for the DSI (Department of Information Systems) to measure the loading of web servers.

The data collection became unusable with the advent of search engines and their robots, proxies servers to surf anonymously, allocation of dynamic IP addresses by ISPs and cached content techniques. All these developments have rendered the use of log files inappropriate to analyze user behavior. The data contained in log files were indeed biased and unique visitor identification almost impossible. There is also the fact that marketing departments had quickly appropriated the web because of its added value. Analysis of log files was therefore a method too technical  for these services to be easily used.

So, audience measurement by logs analysis had become inadequate… The editors of web analytics solution had to change their method of data collection with the evolution of the Internet.

1997 :  arrival of data collection method via javascript tags

The method chosen was the use of javascript markers. This is the most widespread method of data collection today. The logs analysis method is still used but rarely. There are also some web analytics tools which use a hybrid method that combines the use of log files and javascript tags.

Simplified schema of data collection via javascript tags

A javascript tag must be carefully inserted in all pages of the website. The tag is executed at page load, collects datas of the current visit and sends them to a local or remote server depending of the measurement tool used. The tool receives informations from the visit, associates them with the good visit and finally stores them in a database.

Here is a simplified schema representing the data collection via the javascript tags:

Schema of data collection via javascript tag

You will find this shema and the other one about logs analysis in this pdf file.

The javascript tags have resolved the problems inherent in log files and now allow pratically to measure everything, the e-commerce transactions to the number of clicks on any page element.

Only the number of unique visitors remains and will, in my view, always approximate.

Development’s timeline of major web analytics solutions from 1993 to today

Here is a timeline tracing the evolution of web analytics’ solutions market to nowadays:

Evolution of web analytics' solutions

You will find this timeline and the one located at the beginning of the article in this pdf file.

Note that the market has moved up moderately up to March 2005. At this date, Google has entered the market and gave the pace by buying Urchin Software in March 2005 and launched Google Analytics in November 2005, only 8 months later. The Google brand is a huge advantage for the adoption of their solution Google Analytics which is free and whose features are evolving very rapidly. It’s clearly a priority for the Mountain View firm. Google nibbles slowly away market share held by the historics measurement tools despite the fact that these tools offer more advanced features that are not available in the Google tool such as data ownership, warranty service etc… For now, mature companies in terms of web analytics use paid tools to respond fully to their needs and companies that start in the field use firstly free tools. Google’s solution was a huge boost for the industry and paid measure tools must redouble their efforts to maintain the lead they have on Google Analytics. There is also Yahoo! Web Analytics which is a free solution as well. Yahoo! Web Analytics offers a very good alternative to Google Analytics because it is more advanced on certain aspects. This solution can be used by mature companies in terms of web analytics by offering the advantage of being free … It remains still unknown in Europe, I will try to make you discover through this blog in future articles.

As you can see, the market is changing very fast. Knowing that in the Internet a major key success factor is execution speed, we can expect further announcements very soon and often!

It’s at your turn!

Any questions? What do you think of this article? What are your predictions for this year? How will evolve the coexistence between free and paid tools? Are you interested by Yahoo! web analytics? …

I look forward to your comments! Remember that my main goal is to have your opinion in order to exchange all together, thank you :D

Sources:
- Person in charge of Twitter accounts of web analytics tools
- “Visual History of Web analytics” by Eric T. Peterson
- Wikipedia

Print and share this article :

  • Print
  • Twitter
  • Facebook
  • del.icio.us
  • LinkedIn
  • viadeo FR
  • Wikio FR
  • Digg
Brice Bottégal
About the author, Brice Bottégal :

I'm product manager at Tag Commander where I'm responsible of 3 products: TMS (Tag Management System), AMS (Attribution Modeling System) and Privacy.

You can leave a response, or trackback from your own site.

8 Responses to “Definition and history of Web analytics”

 
  1.  SEO St Louis says:

    Hello, i feel that i saw you visited my website so i got here to return the prefer?.I
    am attempting to find issues to enhance my site!I suppose its adequate to
    make use of some of your concepts!!

    [Reply]

  2.  Jeffrey says:

    Great post Brice. Going forward I think we will look to real-time tracking, actionable insight (presentation of metrics that a business user can act upon, not just an individual metric), and even predictive analysis.

    Part of this will be combined into dashboards, typically used by executives, and partially onto a mobile device (i.e. iPhone or iPad, Android, Etc).

    There are already companies starting to pull in GA data right on the iPhone / iPad. Check out this blog post here: http://www.trendslide.com/blog/google-analytics-for-the-iphone/

    [Reply]

  3.  Patrick LeMay says:

    Well explained the history of web analytics. Thank you very much Brice.

    [Reply]

  4.  Ken Quandt says:

    Thank you for this history, Brice. Well done. May I blogroll you on my blog?

    [Reply]

    Brice Bottégal the March 24th 2010 at 08:28 answered :

    Of course ! Thank you :) I will create a blogroll too but not now. Good continuation in writing your blog !

    [Reply]

    Ken Quandt the March 25th 2010 at 20:29 answered :

    Thanks, Brice.

    Take care.

    Kind regards,

    Ken

    [Reply]

  5.  Justin Kistner says:

    Great history of web analytics post, Brice. If only the Wikipedia version was this good… ;)

    [Reply]

    Brice Bottégal the March 23rd 2010 at 16:01 answered :

    Thank you Justin :) English is not my mother tongue so…
    I tried to do my best !

    [Reply]

 

Leave a Reply