CaseStudiesOnline.com

Measuring ROI and The Problem with Outliers

(5)

Jaremy Rich at Viralogy.com has posted a great article about social media ROI and makes the point that social media is, indeed, measurable — the problem, though, can lie in using that measurement model twice. For instance:

If one major influencer picks up your social media campaign and promotes it to their followers, you have a big outlier problem (not the worst problem to have, to be sure). Sure, you can measure and report on the results of that individual campaign, but moving forward, you’re going to have issues duplicating and predicting future success.

The problem is with data extrapolation. If one user with millions of followers retweets your post, it changes everything. It’s just like if a new post or viral video you’ve made lands on the Digg front page – an event that provides an exponential boost in traffic.

Though you can measure that one event, it will be extremely difficult to account for that when predicting future events. So if you’re running a contest that costs $5,000, it’s certainly something to keep in mind. If you consider each impression on Twitter to be worth one tenth of a cent ($0.001 – due the relatively small number of people that actually *read* each tweet), a retweet by @aplusk would put you immediately at break-even when your campaign might have otherwise lost thousands of dollars

It’s a great read.