OU Conf: Tony Hirst on Google Analytics

(Was on slideshare but lost during technology snafu.  Oops.)

What would we do if we were a 2.0 company? Use web analytics to explore student/customer experience to get them to come back.

Helps us answer: What are your students doing online?

T184 Robotics and the meaning of life – 10week presentaiton, two a year, tracked for two years.

How often do they come?  Using webstats: “Most people visited: 15-25 times” over a ten week period.  Breaks down – a lot of people visiting once, but is an artefact – but most ~50% visit ~2-3 times pw.

How many course web pages do they read per visit? “7.66 pages/visit” – again artefact at 1, 2, 3 pp but basically falls off as a tail (exponential? looks more like a polynomial)

How long are they around for? “11:19 minutes average time on site” – artefact of people popping in and out – huge spike 0-10s, but second peak at 601-1800 secs (up to half an hour).

When are they doing it? Mostly daytime, peaks at 12 and mid-evening 7/8pm.

Every day the same? Fridays are bad – big bumps on Sunday and Monday – but that’s when the assessment fell – generally Tuesday seems to be good. (Eh?)

What are they reading?  Mostly the home/landing page. (Duh) But then Calendar, then Search results (to find answers to assessment), then End of Course Assessment, then Assessment landing page, then ‘What is a robot’ which contains several answers, then CMA, then rest of course content.

How long are they reading each page for? Average 2 minutes – can see per page too. Spend a bit more on content pages than other ones – probably artefact from bouncing off home page.

Where are they leaving from? Mostly from ECA – 40% read that then go.  Robotlab is a common one too (27% leave them vs 13% site average) – practical activities described in those, so obvious they read that then go away and use robot (or simulator).  A high rating might flag up that students are unhappy with it.  Outliers and changes are good to watch for.

Does pacing work?  Do they do it like we intend – week 1, 2, 3, 4 etc in order?  Lovely graphs – generally they sort-of do, with nice curves that show that happening, in order, in sort-of the right numbers.

When do students do their assessment? (meaning look at the ECA pages) CMA visits peak around when it happens, then sharp fall.  But some before, and also a little tail afterwards – perhaps search results hitting the CMA page.  ECA – very few look at it at the start! (Blimey, not what I’d have thought.)  Builds up steadily to a huge peak at the due date.

(Wow – students are less strategic than I imagined. At least these students are – as Tony says, would be good to see if this happens with old lag students too.)

Should make it easy for students to get to the assessment – there are extra click involved from the home/landing pages.  Should take a single click – so here’s some user behaviour we can learn from and do sometihng valuable.

What computers do the students use?  (Fun pic of Commodore PET – Tony’s main home machine)  Connection – Most students are on reasonably fast, but 11% on dialup.  “We should encourage dialup to die.”  Screens are getting bigger – so can design screens differently.

Ingrid – can tool tell you what students are printing out?

Tony – could track printable pages, setup not quite right.

Ingrid – evidence that students go to pages and print them out, so how long they spend there is interesting.

Tony – SiteIntelligence – OU signed up for – bigger and better and can segment by PIs and whatever – can track every single page, not just ones with a code on.  Could pull out students who printed out and compare it with other poeoples.

JohnP – Any Data Protection Issues?

Tony – Google can see the URLs, but can’t sniff the content.  Adds to what Google knows about you, could tell you’re a student and add in to their marketing mix.  Google track everything you do.

John – Anything you had to clear?

Tony – Next question!  (laughter)

Steve Godwin – Lots of skimmers, a few detailed learners – does this give you median values rather than just means, because you get very long tails?  Also artefacts for the one-page users – e.g. people putting it up as their home page.

Tony – Google Analytics doesn’t give you raw data, but does some disaggregation.  The OU system does let you do proper stats and segmenting.

Author: dougclow

Experienced project leader, data scientist, researcher, analyst, teacher, developer, educational technologist and manager. I particularly enjoy rapidly appraising new-to-me contexts, and mediating between highly technical specialisms and others, from ordinary users to senior management. After 20 years at the OU as an academic, I am now a self-employed consultant, building on my skills and experience in working with people, technology, data science, and artificial intelligence, in the education field and beyond.

2 thoughts on “OU Conf: Tony Hirst on Google Analytics”

Comments are closed.