Monday, April 12, 2010

Web Analytics, IT and Business Intelligence: friend or foe?

This article was published in the IT World Canada magazine, Volume 3, Issue 3, March 2010.
I'm speaking about web analytics and business intelligence at the Salon Business Intelligence of Montréal, April 14th. 

As more organizations are using web analytics to measure online marketing performance and optimize their websites, those experienced in business intelligence are increasingly looking at  web analytics with a mix of awe and suspicion. The Web Analytics Association official definition of web analytics is "the measurement, collection, analysis and reporting of Internet data for the purposes of understanding and optimizing Web usage" while Wikipedia defines business intelligence as "the skills, processes, technologies, applications and practices used to support decision making". Two definitions that can be competing, interchangeable or complementary - the confusion and misunderstandings are easy.

What's the role of IT in web analytics?

IT responsibilities traditionally include listening to business requirements and objectives and coming up with sound recommendations and solutions to optimize business processes. IT is often the guardian of data integrity and security. Single handedly giving the reign of web analytics to marketing is uncomfortable, especially when data is freely flowing out of the business walls and safeguard to be stored and accessed using SAS (Software As a Service) solutions. Fears of abuse and misuse of corporate data, confidentiality and lack of integration with other corporate data sources are often cited by IT as reasons to hold back and look carefully at web analytics projects. This often leads organizations to address web analytics as a business intelligence project. On the marketing side, we hear over and over about IT's lack of understanding, entrenched mentality and strict rules which impair their ability to act.

What should IT do?

Even if marketing campaigns could be optimized to the extreme, eventually, prospects and customers will have to go through business processes for registration, purchase or support requests. Coincidentally, while marketing can optimize the brand, bring people to the site and improve the customer relationship, those processes necessarily rely on back end systems which reinforce data quality, integrity and security. As with other processes, IT's role in this context shouldn't be different: support the business by providing the most appropriate solutions - including support for web analytics and business intelligence.

The goals of web analytics are slightly different from traditional business intelligence objectives: web analytics aims to measure user interactions from online marketing awareness, social media, mobile, video interactions, and of course, the web itself. Those are non-personally identifiable activities, largely non-transactional (in classic IT terminology), they have a relatively short lifespan and are much closer to the realm of statistics - with implications of sampling and margin of error. On the other end, business intelligence relies heavily on historical transactional data where integrity is reinforced, customers are identified and the timespan require longer, more thorough analysis which will lead to improved business process and better strategies.

Who should be the champion of web analytics?

In "Analytics at work", author Tom Davenport, a leading authority in analytics, suggests IT is better skilled at building data infrastructures and installing and maintaining application that generate transaction data, and less oriented to help the organization use data in reporting and analytical processes. However, no one would think of challenging IT's crucial role in a Business Intelligence Competency Center. The Online Analytics Maturity Model, which address the six critical process areas to succeed in online analytics, also recommends empowering a multi-disciplinary team where IT, business/marketing and analysts work hand in hand to grow the online analytics skills and maturity. As it stands today, and because of its focus on the overall user experience, marketing is usually the best home for online analytics - but IT should definitely have a seat at the table.

What's easy, what should be done now?

One of the most common issues in web analytics is the distinction between the "informational" part of the site and the "transactional" side. The necessary instrumentation, called "tagging", is often more easily done and more flexible in the former, while the later suffers from long delays and stricter IT rules. In a multi-tier approach, the presentation layer is de-coupled from the logic and data layers. A simple mindset change is required: web analytics should be considered a presentation-layer responsibility. Front-end web developers are usually much better equipped to deal with it, with the agility to make changes and increase velocity. Techniques exists to ensure the "tags" will not interfere or break the user experience, the logic or business data. Given that agility and velocity are essential ingredients to become truly analytical competitors in the online world, IT should empower marketing and support their desire to use web analytics - for the benefit of the whole organization.

What's the future of web analytics?

The history of web analytics tracks that of the web itself. We went from logs to tags, from IT-centricity to marketing-centricity. Organizations with a competitive edge today succeeded in naturally blending the online channel in all functions of the business. Online analytics isn't a matter of IT versus marketing, nor a matter of web analytics versuss business intelligence. It's a matter of cultural changes leading to the use of the most appropriate data, at the most appropriate time, that will lead to the most optimal and realistic decision in a given context.

Eventually, as web analytics grows and integrates with business data, and as marketing eventually wants to optimize the online representation of business processes, we will necessarily enter the world of business analysis and business intelligence. Today’s web analytics should be viewed as a smaller, narrower and more agile little brother of business intelligence - and as a little brother, web analytics might need some guidance and a reality check from time to time.

Additional info: