Metrics and intuition

Jun 3, 2009 by

The other week I posted a piece about social media measurement and the on-going “you can’t measure it!” debate. After all, with close to 3000 delicious results when you put in socialmedia + measurement, surely we actually have metrics coming out of our ears?

Norwegian planner and blogger Helge Tenno replied saying that it’s easy to measure…all the wrong stuff! And I was reminded of that when doing a recent client project where I looked at the influence of people online.

There were plenty of tools available for me to use, some very good ones. Unfortunately in just about every instance there was a flaw with them.

Trust, engagement, reach

My starting point is that when we measure someone’s influence, we are really looking at a mix of three things: Trust, engagement and reach. Trust – is that person generally respected in his/her community? Engagement – does s/he spark off conversations? And reach – the numbers game marketers love.

So far so good and the measurement tools I favour are all openly available (free). They include:

Twingly as opposed to Technorati. A Swedish owned search engine, Twingly gives every blog a trust score from 1-10 (in practice you are doing well if you get 5+). But unlike throwing everyone into a big international pot, Twingly scores different language blogs against each other. So Polish blogs vs other Polish blogs, German blogs vs German blogs and so on.

Twingly is a great resource….except being relatively new it has gaps, not everyone has been indexed.

Twinfluence. Twinfluence starts from the premise that with numbers you can’t compare like with like. So you have 10,000 followers. So what? Arguably 100 somebodies are more valuable than 10,000 nobodies. Twinfluence measures the influence of your followers to give you both a second order follower and social capital result.

Excellent, though of course people need to be on Twitter in the first place and Twitter also seems to have more uptake in English speaking countries.

PostRank. AdAge recently switched to making PostRank its most important metric for its Power 150 list of English speaking marketing blogs. Engagement is clearly something we don’t measure enough of, so good on AdAge for introducing this. However, PostRank suffers from a few of the problems of Twingly. From personal observation, coverage isn’t 100%, especially with non English speaking sites.

Statbrain. Takes all those Alexa, Google and Yahoo scores and reckons how many people visit a site. It’s often right, but it is ultimately only a guess. Taking this site for example, it flattered me when it came to estimating visitors!

That will do, it looks good

Let’s be honest, within agencies we often paper over these gaps taking the view that a nice tidy spreadsheet with the numbers filled in will do the job. It ticks the client boxes and it makes for an easy life.

The power of observation

But by doing that you miss a lot. It’s not an easy argument to make for clients who want measurable ROI but often there is no substitute to taking a judgement call: Using your own personal observation about what someone is doing and deciding from what you personally see whether they influence a community you are trying to reach.

In AdAge, Judy Shapiro put up an all about “the tyranny of technology in testing marketing effectiveness.” Judy’s take was that by obsessing about measurement we end up getting things wrong:

“Maybe it’s time we all start dialing back the measurability rhetoric because the promise rarely lives up to the expectations anyway. Stuff can and does go wrong — technology stuff, plan-test stuff, real-world stuff.

“Then, on top of all the things that can go wrong, there is still some amount of guesswork required to interpret results because you need to explain why certain behaviors did or did not work. No raw metric can answer that. “

Well said. It’s messy and it takes some justification, but sometimes there is no substitute for guesswork and taking your own call on things.

Enhanced by Zemanta

Related Posts

Tags

Share This

1 Comment

  1. Anonymous

    You can also try http://www.estimix.com – a free tool that provides a nice summary of the website performance.The estimation provided by estimix is the result of a complex analysis based on factors like: the age of the website, the demographic structure of the traffic, the countries where the website is popular and sources of the traffic.