Friday, 14 January 2011

Enterprise microblogging: measuring true value "is relative"

Yammer announced a new feature yesterday: Leaderboards
Leaderboards gives users access to statistics about their network activity. The Leaderboards include:
  • Most Liked Members: Top 10 users whose messages have received the most ‘Likes’
  • Most Replied to Members: Top 10 users whose messages have received the most replies
  • Members with the Most Posts: Top 10 users with the most public messages posted
  • Most Replied to Threads: Top 10 threads with the most replies
  • Threads with the Most Participants: Top ten threads with the most participants
I applaud the introduction of stats as they give insight in general behaviour and performance of enterprise microblogging users - but I think Yammer is missing an opportunity to add real value here

Absolute figures do say something, but not much really.
Threads with the most participants? That could very well be a Friday afternoon LOL chat that got out of hand.
Most replied to threads? Same there.
Members with the most posts? Most replied? Most Likes? Those will highly likely been more or less the same. If you post most, you're likely to receive the most replies, and likes - oh and posts of course

What do I propose then? Relative stats - not absolute ones

Relative stats will take into account the frequency with which you post, and the interaction that leads to. What if you want to spot pure genius? You'll be looking for someone who speaks wisely - meaning that whenever he (or she) speaks, people reply and like in great abundance.
You are not looking for someone who yams every time he arrives or leaves somewhere, has lunch, plays the guitar or engages in other omphaloskeptical activities that merely add to the chatter, and don't add value

To show you how this works, here's an example of stats from Yammer:
  • First, most Messages
Who's leading there? It's the Social Media Evangelist! Followed by the Thoughtleader and the IT architect. What do you think of that? They seem very active participants, and contribute a lot - wouldn't you say so? Are you wanting to hand out rewards already? Hang on to them please, we'll get there
  • Second, most Likes
Who's on top of all that? Hey, it's that same Social Media Evangelist again! He must be really good. And that IT architect, he's very consistent too - still at third place! And that CTO must have very interesting things to say really, he's Liked a lot. Our Thought Leader there is in the top 5 as well, yes he must be awesome too. Well let's hand the MVP award to our Social Media Evangelist then, he clearly rocks! He's leading both stats, that means he's truly awesome and great at his job, right?

Wrong. These stats show you nothing really, and you certainly can't combine the two of them just like that

The guy with the most Messages having the most Likes is just as big a surprise as the guy with the most body weight winning most arm wrestling games - very predictable and not exciting at all.
Why is it so good when the little tough lady beats the big muscly 6-foot tall brute? Because you wouldn't expect it.
You wouldn't, because there is a direct relation between body mass / muscle and strength. Why is the (original!) Karate Kid final fight so cool? Because he's the skinniest, the least-trained, the least aggressive - and still wins. And he doesn't even cheat either!

So, we have to take into consideration the relation between Messages and Likes. There is a very easy way to do so: just divide the number of Likes by the number of Messages
  • Third, most Likes per Message
This really shows value perceived by the community as a whole.
The "technology thought leader" has almost dropped off the list, the SM evangelist ranks 6th, and is beaten by that Integration Architect who was trailing in both other lists.
Better yet, the Head of Internal Communications, who was almost equally trailing as the Integration Architect, is our best man!

On top of that, this list can also show the relative relativity: if you sum all Messages and Likes, you'll see there's an average of 28% for Likes per Message: that Social media Evangelist is just scraping it there, and the Technology Thought Leader should give his messages a lot more thought before posting them, or just stop doing so. Apparently, nobody cares about them

That does shed some new light on our previous heroes there, doesn't it? Usually, you'd look at the Message list and marvel at the top three there, then look at the Likes list and think: "Wow, the Evangelist and Thoughtleader really rock! They contribute the most and get (almost) most likes as well!"
Now, you'll just look at the Likes per Message list and be able to assess their real value and contribution at one single glance: rubbish

This is how you make useful stats: take into consideration the relations between the different types you measure. This is just an example, but you can probably imagine that a "Most replied to members" list would look quite differently when taking this into account.
And of course this example is not saying everything either: it would be useful to know for how long these roles have been active in the enterprise microblogging network. Placing 100 message a year or a month does make a difference. Equally so, the size of the network matters: how big is the chance of getting a Like, or @Reply?

Yammer, why don't you beat the rest by offering real competitive value and give us stats like these? It would certainly give great insight into the value Yammer adds to a network / community!

0 reacties:

Post a Comment

Thank you for sharing your thoughts! Copy your comment before signing in...