Your day can change pretty fast in Lisbon. One minute I was sitting down at a coffee shop, talking with Pedro Garcia through Facebook, the other I was writing down some thoughts on social media metrics and reports to share with his class at EDIT.

Not surprisingly, some of what I wrote down came from what I learned with James and Sabrina at Webnographer, other bits from the reporting models I built at Fullsix.

The one I feel works best uses Output, Outtake and Outcome.

Output will refer to everything we are able to produce and publish, it is also a map of where we are putting our social media efforts.

Outtake on the other hand will be the first step we get from that output. It can be interactions on each social media channel, clicks or visits to the website. Metrics we fit into this category will be a half-way step towards the communication goals and business objectives.

Outcomes should be clear indicators that our strategy is working or failing.

There are of course other ways to look at our efforts, such as AIDA or a conversion funnel.

One of the questions that came up was which source we should use for the reports. Either each channel's own export or a more industrial solution like Buffer or Social Bakers. Exporting raw data directly from the source will no doubt be safer. Take buffer for example, one disclaimer on their analytics is that there may be discrepancies on linkedin data because of the way it un-shortens links.

And for an end-analysis, the client should provide data for business goals. At Fullsix, some brands went as far as running survey studies to get a perspective of the effect their social media strategy was having on an offline context.

”Social Media is always changing” is repeated every so often, and even if we don't agree with it, it is important to keep some things in mind when the landscape changes and we need to adapt reports.

  • Make sure everyone can understand and replicate the data
  • Always link communication objectives to business goals
  • Prefer metrics that show actual behavior (a pageview is not really a reflected act, but sharing a piece of content is)
  • Avoid compound metrics

This is by no means a comprehensive list, but after such a great time talking with Pedro and his class I couldn't just let it die there. Feel free to add to it or correct in the comments!