I’ve been wading in statistical waters recently with one of our new intranet clients, the Office for National Statistics, and by studying a MOOC on Machine Learning by the Department of Statistics at the University of Washington.
ONS intranet
The ONS recently blogged about making a modern intranet and I’m really pleased to be working with the @ONSdigital team, who have already completed a huge chunk of user research and have been using WordPress and the GovIntranet theme in alpha.
These guys literally rented a bit of web space, installed WordPress and followed instructions on the GovIntranetters community site to work up a functioning prototype.
http://govintranetters.helpfulclients.com/task/setup-guide-instructions/govintranet-setup-video-tutorial/
The current ONS intranet is managed in Tridion, a CMS with a hefty annual price tag and difficult to support. The contract ends early next year and we plan to have the new intranet in place by the end of February.
There’s a team of five on this project, covering technical, content and policy. The content strategy and policy are already in final draft and these lay solid foundations for the ongoing quality and appropriateness of content. There are very clear guidelines for what goes on the intranet and what is expected in terms of ongoing content production and housekeeping. The plan is to export content from Tridion and use this as a basis for loading the new intranet with rewritten content.
It’s encouraging to see thought going into ongoing content management, not just concerning publication and quality, but also housekeeping and evaluation. Great to see policies around content review and retention; there’s a clear message that pressing the Publish button doesn’t complete the job. Stale content needs to be removed. Important guides and policies need to be checked for accuracy and currency. Content that hasn’t been viewed needs to be questioned.
The team have been using the working name of “Reggie” for the new intranet but are running a naming exercise for staff to determine the final name. Some of the suggestions so far topically include Yoda, C3PO and Anakin 🙂
The tone of the alpha site is fresh, friendly and open. The placeholder text in the search box reads “Reggie reporting for duty” and the team have been dual-publishing news and blog posts from their existing intranet, making the alpha site feel current, up to date and alive with photos and avatars.
Zero Hits Monitor
As part of the work that I’m doing with the team, I’ve developed a new plugin to display an exception report, highlighting content that hasn’t been viewed.
Google Analytics doesn’t know about pages that exist on the intranet but which have not been accessed. So you can’t simply generate a report from your GA dashboard. The Zero Hits Monitor plugin adds another layer to your analytics so that you can see what’s missing.
As part of a background job, the plugin creates a map of all intranet pages and then queries Google Analytics for pageviews over each month of the past year for each of the pages. The dashboard widget highlights pages with no views over the last 6 months or the last year.
Testing this on a new intranet isn’t useful as there is no history of analytics. But I’ve been able to plug this into test areas on other client intranets, running against live analytics. Interesting results. Top of the list, consistently across the handful of other intranets that I tested: equality and diversity, information assurance and records management guidance. Does that mean that this content is ripe for chopping? Probably not. But it does indicate that this content is up for review.
Machine Learning
On Coursera, I’ve started a set of modules on Machine Learning. The foundation module has been a fun-filled introduction into the statistical theory and practical mechanics of things such as how Amazon and Netflix use machine learning and huge sets of data to provide recommendations.
I’ve been coding since my teens, and to this day I work with databases, but calculus, quadratic equations and regression are topics that I left behind in my teens. I’m exercising parts of my brain that haven’t seen the light of day in decades, but it’s astonishing how quickly one gets back into the routine of solving equations and finding derivatives 😉
Today’s intranets are a great source of data, containing tagged and categorised content, detailed usage analytics and staff interactions in the form of comments, likes, forum posts and profiles.
My mind is spinning with ideas for how we can use this data.
I’m picturing an awfully helpful content bot that issues a warning when HR content editors hit publish:
The Zero Hits daemon predicts that nobody will view your page!