Through a glitch in the Matrix, intrepid blogger Gonzo was able to secure access to South by Southwest Interactive (SXSWi) badge and airfare down to sunny Austin, TX. Before he was tackled by security, tagged & bagged and hauled off the county jail he was able to capture the following insights:

SXSW Day One
What are Analytics? A Guide to Practical Data

Margaret Francis and Blake Robinson led the discussion. Some key learnings:

There’s a lot more to it than volume. What about your marketshare vs. your competition’s? What about sentiment? Looking at a bunch of numbers without context is a waste of time. You need to establish a baseline and start measuring change from there. Also, look at how your competitors are doing, relative to your brand. Sites like Compete, Alexa and Quantcast can be useful for this, but remember that these numbers are not reliable since they aren’t based on actual server logs.

There is no “full Twitter data” because of limitations that Twitter has placed on API calls. Data usually only goes back about 10 days (lately), meaning older tweets have essentially disappeared down the Memory Hole. This data may eventually become available as Twitter matures and its rapid growth flatlines, but until then you can use tricks this to capture data:

First, it’s helpful to start monitoring beforehand. Failing that, jump on it as quickly as possible. Use Twitter-Search’s RSS feature and pull that feed into a reader, like Google Reader. Boom! You’ve got a backup of all tweets using the #sxsw hashtag, for instance, and if you used Google Reader it’s all easily searchable. Hopefully this hack won’t always be necessary, but for the moment it is.

Here’s a quick rundown on useful tools for measurement:

  • Use postrank to measure engagement on blogs
  • TwitterCounter for measuring follower growth
  • ScoutLabs for insight into related keywords (panelist Francis works for SL)
  • Google Analytics still reigns supreme
  • Using a bit.ly link can help you track links in closed networks like Facebook

Other key takeaways:

  • Engagement and influence matter more for B2B clients than B2C clients because the space is so small that even one influencer can have tremendous impact
  • There’s no silver bullet for weighting comments/posts
  • The panelists still lean on spreadsheets like Excel to crunch their data
  • Time spent on site, not page views, could be most important metric for many sites (each site is different and you need to decide on what’s important to your brand/client)

One dirty little secret in analytics is that a given person will agree with another given person on the sentiment of a post only 85% of the time (based on their research). Get computers to do it? Good luck. Another thing that clients don’t like to hear is that they need to take the lead on measuring ROI. Since they know the Return (if any) on the Investment, they are in the best position to judge. Vendors and agencies don’t have access to sales data so clients will need to have their own homegrown measurement capabilities so they can decide how successful campaigns are from their perspective.