On 7th March I started a 21-day free trial with Contextly – “the most powerful content recommendation system for publishers,” used by sites I enjoy like Adafruit and Spacenews.
One of the main things I like about Contextly is that in addition to (apparently) making great recommendations using only internal links they promise helpful reports and other customisation options – and are completely non-spammy.
Context: DriverlessCrocodile.com
Driverlesscroc is tiny – it averages about 50 views a day – but traffic has been steadily increasing from 11 views per day a year ago, and I was curious to see if better recommendations from Contextly would give things another poke in the right direction. I can’t afford the $9 a month fee to run it permanently, so this is very much a short-term experiment.
The Experiment
I’m interested in two things:
- Do Contextly recommendations lead to a greater number of views-per-visitor for the second and third weeks of the trial? (I’m not counting the first week because with a regular reader I clicked on quite a few links to see how everything was working.)
- What else can I learn from their reports and recommendations?
User Experience of Contextly so far
In short: Contextly does what it does on the tin. It’s a simple couple of clicks to install the plug-in in WordPress, then you go to plugins > contextly > settings (or settings > contextly) and in another couple of clicks set up your trial and generate a code to get connected to Contextly’s recommendation engine.
Then all you have to do is click the helpfully titled “Big Settings Button” and you’re whisked off to Contextly’s own site to tweak the settings Everything happens off-site to reduce the impact on post load-times. You can customise everything that matters (well, that mattered to me) there in a matter of minutes, and you’re off to the races.
Options
I’ve opted for “Related Posts” and “Explore DriverlessCrocodile” recommendations at the bottom of the page only, text only (I only use images rarely so the alternative was a lot of thumbnails of my logo). This is probably the least attractive option for readers, but it’s unintrusive and suits me.
Early Results
It’s too early to tell what impact if any Contextly is having on my views-per-visitor, so I’ll save that for a follow-up post in a couple of weeks. What I would say is:
- The recommendations do seem to be much better than the previous recommendations plug-in I was using;
- The daily and weekly reports are excellent: a very succinct summary of what impact Contextly claims to have had (number of clicks on recommendations, click-through-rate etc) and some really helpful information about average percent-scroll on each post, traffic sources, and information about the performance of different tags and categories. This is a great resource for finding out more about what people are doing on your site.
Early Summary
To conclude for now, this experiment has already been worth doing, and I would recommend Contextly. My feeling is that my low traffic volume and the idiosyncratic nature of DriverlessCroc is going to limit the effectiveness of Contextly’s AI (and the measurement of any impact) – but time will tell.
More in a couple of weeks.