Jan
21
Consolidated UBC web analytics: better data, better metrics, better SEO
Posted by: Jens | January 21, 2012 | Leave a Comment
The UBC web infrastructure is decentralized and web properties are managed within individual units which consist of faculties, departments, programs, research centres, administrative units etc. It is estimates that more than 1400 websites exist under the UBC domain. Many more sites might reside outside this domain despite an affiliation with UBC. Sites can be hosted on UBC infrastructure or externally. Despite this fragmented landscape we all share some common goals:
- Building websites that support UBC’s and unit goals, such as student recruitment, community engagement, knowledge dissemination or donor relations
- Optimizing sites for conversions and retention
- Improving the user experience and increasing satisfaction with our UBC web experience.
The question is if optimization of individual websites will lead to the best results or if we should incorporate a more holistic approach to evaluate and analyze user behaviour and experience within our overall domain and network of sites? We can distinguish between various perspectives:
- Unit: operating one or more sites within or outside the ubc.ca domain, measuring and monitoring the site(s) from the unit perspective only, e.g. what is the user experience like on site.ubc.ca?
- Institutional: a holistic view of activities across all UBC affiliated sites, e.g. how does a visitor navigate between various UBC sites?
- User: the user most likely will be (and should be) oblivious to the organizational structure and who operates a particular site. The goal should be to optimize the user experience when interacting with UBC. Current optimization happens on the unit level only and we do not evaluate transition between sites, flow of information (e.g. repeated entries of the same personal information in forms) or similar.
At this stage UBC only measures on a unit level and completely misses out on the institutional perspective with potentially significant impacts on the overall user experience. As a university we are unable to determine among other things:
- How many web sites exist in the UBC network
- How many unique visitors UBC serves each year across all these sites
- What the biggest web properties are based on pageviews, unique visitors, attracting new visitors etc.
- How visitors enter the network and navigate through a variety of UBC sites, including where they exit or where they convert
- Analyse and optimize UBC’s search performance from an institutional perspective to avoid competition for similar keywords and losing traffic for unidentified keywords
- Investigate user experience or engagement with the UBC domain.
I submitted a discussion document to UBC IT to evaluate options for centralized web tracking at UBC. This would have several advantages, esp. to improve user experience and search engine optimization. We could for example develop UBC-wide SEO benchmarks and best practices or identify click paths through our network of sites. This could potentially highlight websites which tend to lose visitors, e.g. due to usability, design or performance issues, and would provide means to act and improve these sites. Keyword analysis could be conducted globally to identify terms that we perform strongly on in our network and surface keywords that we might miss out on completely.
Furthermore, centralized tracking seems to be a required step for better measurement of our marketing and communication activities. Campaigns typically include a landing page and action items on these pages to measure conversion rates. This provides us with basic metrics to evaluate success of campaigns. However, it would provide additional information if we could identify which other UBC sites get visited or how visitors interact with these sites.
Example: a development campaign might drive traffic to a landing page which informs a visitor about opportunities to support certain projects with an option to fill out a form for more information. We will be able to monitor who clicked a link in the campaign email, what the click through rate is and what the conversion rate for the form is as well as metrics such as bounce rate on the site etc. However, any additional follow-up that happens outside the campaign page/site will be lost. Let’s assume a visitor does not fill out the form on the campaign landing page, but is intruiged by the presented content and continues surfing other UBC sites. Maybe an action is taken on any of the other sites, such as a signup for a newsletter of a faculty or research centre. This contact should be credited to the campaign, but right now is not…
My conclusion is that we have exciting oppotunities ahead of us to utilize the available data in better ways and to learn from the valuable information at hand. We can start in the existing setup by clarifying who has responsibility for web analytics and SEO within units and making these tasks part of job descriptions. At the same time, we can evaluate if a consolidated web analytics approach is possible and how we can best organize such a setup.
Comments
You must be logged in to post a comment.