Hillwatch Web Analytic Brief: The Public Sector�s Problem With Web Traffic Software

Scott Proudfoot, Principal & Alex Langshur, Principal

Every government site has web traffic software but does it help public officials determine if program objective are met, stakeholders �take up� online content and the site has an measurable impact?

The old adage is �you can�t manage what you can�t measure�. Public sector web site managers need tools and techniques that yield meaningful metrics if they are to manage their sites.

Currently, standard web traffic reports are replete with arcane metrics, dense with data and short on insight. Reporting formats are not clear and intuitive. They do little to help managers make strategic decisions about the allocation of limited resources.

Extensive research indicates senior public sector managers rarely read, much less are engaged by, the results from web traffic reports. When they do view such reports, the tendency is to focus on a few basic statistics that seem understandable, such as �hits� or �page views�.

Even site operators who acquire literacy in web traffic metrics are discouraged by the difficulty of using the data to drive meaningful site improvements.

There is a clear disconnect between the data received on web site performance and the data that public sector managers require to strategically manage this new, resource-intensive and increasingly important class of communications.

Simply stated, standard web reporting software does not assist web operators to evaluate which organizational program objectives are being realized online. They do not tell you if you are reaching your stakeholders, whether stakeholders are reading your key content, nor if you are having any measurable impact.

The lack of strategic web site performance information is an issue Hillwatch has documented in many organizations.  Our research indicates the issue is related to one or more of the following aspects:

By technicians, for technicians: Web traffic software has been largely developed by technologists to answer technical performance questions.  While it has evolved substantially and now provides increasingly sophisticated tools to decipher key data about web traffic, it is still implemented, configured and largely managed by technicians.  The result is that these tools inevitably produce technical reports � not strategic management reports.

Cookie Cutter approach: In the race to provide a low cost mass-market product, the web traffic software industry has provided automated cookie cutter standardized reporting and off-loaded the customization to its customers. Most public sector and private sector customers do not have the in-house capability to customize the reports to the proper degree and, therefore, simply take the data at face (i.e. very limited) value.

Privacy issues: In order to maximize the value of their investment in web traffic analysis software, commercial e-commerce entities use multiple tools such as persistent cookies, page tagging (beacon technology) or packet sniffing devices, and parse the log data through highly customized filters based on user segmentation, path analysis and behavioural modeling.  Successful and meaningful interpretation of the data requires intelligent, experienced analysis by knowledgeable individuals.  It takes considerable skill and resources to bring these techniques together successfully. 

Federal Government departments and agencies are specifically prohibited from using persistent cookies. Effectively, government organizations lacks access to tools required to mitigate the key performance deficiencies of web traffic software.

For Business, Not Government: Primarily, web traffic software has been designed to track commercial sales and transactions. The concepts of profits and revenues do not translate well to government entities. Most government web sites are not primarily transaction driven and transactions often take a different shape and character with government. Governments need measurement tools that recognize their unique requirements.

Skewed Data and Phantom Visitors: Finally, without considerable customization, web traffic software under-reports the actual impact of automated web site visitors such as link checkers, search engine bots, spiders and e-mail harvesters. There are literally thousands of these automated visitors crawling the web. These entities generate tremendous traffic volume and can significantly distort site usage statistics.

For example, we have documented many leading web traffic software applications (for sale by US-based companies) do not count visits by French language search bots. We have also conclusively demonstrated these French language search bots routinely visit and index Canadian government sites that have French language content.

AOL is the dominant ISP on the web and it uses proxy servers. This means a single visitor may be accessing different parts of your site from different servers. To the web analysis software it looks like you have multiple visitors accessing single pages. Given the large number of AOL customers this can lead to a significant distortion of actual visitor patterns.

By not filtering out the visits by such bots or the AOL proxy servers, the number of visitors to a web site is significantly overstated and can be 30%-40% less than stated in your web traffic reports.

Whether spiders or AOL proxy servers, false data skews such key metrics such as repeat visitors, most popular pages, number of downloads, pages viewed, etc. More to the point, useful management reports are traditionally built on trend and ratio analysis. Such analysis is meaningless with inaccurate data.  The rule of garbage-in, garbage out applies.

Why?

Organizations tolerate the lack of clarity and utility from web traffic software due to:
- The relative newness of the Internet;
- Lack of web metrics literacy;
- Perceptions web site is not a core strategic asset
- No clear consensus across Government about the proper tools, methodologies and standards to measure the effectiveness of web assets.

Readers who enjoyed this Brief may like to visit our Web Analytics Briefs home page.