I was fortunate to be able to spend four days last weekend down in Austin, TX at South By Southwest’s (SXSW) Interactive Conference. In addition to meeting a ton of great people from around the world, there were amazing sessions on every topic under the sun, from user experience to mobile, from copyright to managing Gen Y. One thing that was obviously missing from the conference though was a discussion on analytics. Of the several hundred panels within the interactive schedule, there was just one panel on analytics. Today, now that I’m recovered and back in the swing of things, I checked the panel picker, where interactive fans and attendees can vote on which topics will be presented at the conference, and of the 2400 submitted ideas, only 20 had anything to do with analytics (some very loosely). Analytics is clearly underrepresented. Next year, I think I’ll need to do my part to try and fix that!
Just One Panel?
The one panel that made it onto the conference was “What Are Analytics? A Guide To Practical Data”, which had a slightly misleading title, since it was 100% about social media analytics. Margaret Francis and Blake Robinson lead a packed room of over 200 attendees and did a nice overview of the top things to be measuring within social media campaigns. Some of their key points included a focus on social media sentiment, comparing your brand/company to the competition and considering the tangential relationships that can be found by going down the rabbit hole of tag clouds and keyword usage. The panelists admitted that we are very much in the infancy of social media measurement. While tools like PostRank.com, TwitterCounter.com and Nielsen's BlogPulse are useful, they are not at all equivalent to the tools that we have to measure website interactions today. Geographic and demographic profiles on social media are barely directional at this point, but everything will continue to evolve. Their key takeaways were: make data visible, set goals and metrics, track performance over time.
Mobile Tracking
As part of the conference, DigiKnow had the opportunity to work on a mobile project with one of the event sponsors (who I can’t mention by name). I wanted to share a little of this experience, since it was our most focused effort in measuring mobile. For this site, we initially setup Google Analytics with image based mobile tracking, figuring that would cover both smart phones that could run the JavaScript tracking and feature phones that might not be able to. On Sunday (half way through the conference), our data seemed low and we wondered if there was a problem, so we added in a second profile running the JavaScript tracking. As it turned out, the image tracking was only getting about half of the data that the JavaScript tracking revealed. Today, well after the fact, I also ran the site’s logs through a log-based tool and it came back with data that was very similar to the JavaScript tracking. Given the fact that the typical SXSW crowd are smartphone users, this wasn’t too surprising.
Key takeaways:
-
For mobile, if you can use multiple collection methods (image, JavaScript, logs) you will get the most complete picture, but if you can only do one, I recommend sticking with standard JavaScript tracking
-
Any time you are using a new tool or data collection method, test it thoroughly before launching it. We needed more test data to reveal the deficiencies in the image tracking for mobile in our project.
