Before starting a web redevelopment project, you need to be able to answer questions like these:
- Which pages or subsites are important enough to warrant immediate attention?
- What, specifically, is wrong with those pages? (Consider both the organisation’s perspective and the end-user’s perspective — they may not match!)
- What problems can be fixed at minimal cost to the organisation?
- Once the problems are fixed, how can you prove that it worked?
- Did this situation arise because your web maintainers lack essential knowledge or technical skills?
Analysing site traffic can help to answer these questions. In this article, we will look at a site traffic report created specifically for a web redevelopment project.
The example report, Site traffic analysis for a web redevelopment project (PDF 320 kb), was produced in mid-2007 as part of the ‘evaluation’ phase of a web development project. I have (mostly) obscured the organisation’s name in this sample report, but the data and recommendations are real.
Key features in the sample report
The report covers a representative time period, based on the organisation’s annual activity cycle. The sample report focuses on the first quarter of the year. This period encapsulates the annual web usage trends at a typical Australian university:
- January is particularly quiet, because staff and students are all on holidays;
- March, when first semester begins, is the busiest month in the academic year.
For a tax consultant or a department store, the peaks and troughs would occur at different times of the year.
The report focuses on a relevant selection of web pages or subsites. The sample report was written for a project that focused on web sites intended to support the delivery of IT services for the university’s students and staff.
Visitor demographics are summarised, but not analysed in detail. This summary information can be compared with other market research, to confirm whether the right people are visiting the web site. However, site traffic data cannot tell you whether visitors are happy with your web site.
The most-viewed pages/subsites are identified. In this project, the number of pageviews for a particular subsite was not important. The key metric was the proportion of overall traffic that went to each subsite. We wanted a relative ranking of the subsites, rather than a snapshot of traffic volume.
The last section of the report is a brief audit of the technical quality of each subsite’s home page. A subsite’s home page is more likely than other pages to receive regular attention from the site’s maintainer and owner. In this organisation, the home page also tends to be the most-visited page on a subsite.
Doing the technical audit
The technical audit took about 10 minutes per page, using the Web Developer extension for Firefox. Pictures of the screens were captured using SnagIt.
Here is the checklist I used:
- What kind of site is this? eg e-commerce, information about a topic, information about an organisational unit, promotional campaign…
- What are the main pieces of content on the home page?
- Based on the content, who is likely to be the main audience for this page? What other audiences might find this page useful?
- Are there any access restrictions on the page? Is the access setting appropriate for the page’s intended audience?
- If the main audience is inside the university, is this clearly indicated by the page’s design and content?
- Does the page use valid code?
- Does the page comply with the university’s branding guidelines?
- Is the page accessible to people with disabilities?
- If the page does not use a standard university design template, does it meet expectations for multi-browser compatibility? Could a visitor make sense of the page with stylesheets, javascript or tables removed? On a mobile phone or other portable WAP device?
- Does the page make unreasonable demands on the web server? (download time, number of components required, size of individual files)
- Has the page been updated recently? Is its content current?
- Are there any obvious errors in the content, design or construction of the page?
The results are listed in a consistent manner for each page, so that comparisons are easy.
Finally, the “Recommendations” section
Because this report was part of the ‘evaluation’ phase of a web redevelopment project, its greatest value is in providing evidence-based recommendations for further action.
To be useful, the recommendations should be:
- Clearly based on the evidence presented in the rest of the report — no personal opinions and no suggestions that are based on other reports or experiences
- Practical and achievable within a specified time — I usually assume 3 or 6 months as a maximum for each task, and make a note of any resource constraints that may impede progress
- Aligned to the organisation’s business needs and strategic directions
- Aimed at producing a measurable outcome, so that you can demonstrate success
Though the sample report is far from perfect, it served its purpose well. The recommendations were added to findings from other research, and eventually organised and prioritised into a 12-month action plan.
As the action plan is implemented, the report can serve as a checklist for the people working on the various web sites. The project manager will be able to produce another report in mid-2008, using the same methodology, and (we hope!) thus be able to demonstrate quantitative and qualitative improvements in the web sites.
Further reading
Scott Berkun (2005) The Art of Project Management. O’Reilly Media Inc. ISBN-10 0 596-00786-8, ISBN-13 978-0-596-00786-7. See especially chapter 3, “How to figure out what to do”. Scott also publishes occasional essays online and runs an email coaching forum called PM Clinic.
Mike Kuniavsky (2003) Observing the User Experience: a practitioner’s guide to user research. Morgan Kaufmann Publishers, San Francisco. ISBN 1-55860-923-7.
Tags: snagit, firefox, user experience, web analytics, project management, quality audit