Measuring influenceYesterday we wrote about a new app called The Social List, which has been launched by The Sunday Times and is boxing in a similar area as Klout and PeerIndex.

Here’s how it is being positioned: “The Sunday Times Social List is set to become the definitive measure of the most influential people within the social space.”

Tricky. The trouble with measuring ‘influence’ is that it is incredibly difficult to do so in an automated way. I think it’s virtually impossible to make any real statements about who is and isn’t influential without some form of human analysis. 

These tools are of course works in progress, and as they stand they are certainly indicators of something, but I’m not sure they’re indicators of true influence.

Define: influence

A common definition of influence is that it reflects the ability of a person (or a thing) to change behaviour, or mindsets, or to persuade someone into taking some form of action. Influence is relative, and it is not the same thing as reach. 

Activity

I attended the launch party for The Social List last Friday and the first thing that struck me is that the algorithm appears to be heavily linked to activity (yours, and that of your followers, which is in part determined by how many status updates you push out).

If you go on holiday for a week your score will definitely drop. This is no indicator of influence. The genuinely influential do not become any less influential following a week of lazing around on a beach.

Gaming

We have already proved that Klout scores can be gamed by bots, and until these apps can apply the right kind of weightings and ranking factors they will be at risk of spam. And spam is one of the biggest issues they face. They are no different from search engines in this respect.

Take the Econsultancy Twitter feed as an example. We see spammers on a daily basis, who retweet our articles in batch (sometimes we’ll see 50 retweets coming through in a matter of seconds, all from related spam accounts, which we block and report). Twitter needs to deal with these losers, and so do the ‘influence’ measurement apps. This kind of activity should not be be reflected positively in any scoring of @econsultancy. 

Follower activity is where it’s at

Well, kind of. It isn’t your activity that these apps should be interested in, but rather the activity of your followers (and their followers, and the followers of those people, etc). 

Influence is partly about persuading people to take some form of action, which on Twitter might crudely be measured as a retweet or @reply. Note that some retweets are more valuable than others. This is something that these activity-based algorithms need to better understand. 

A retweet isn’t always a recommendation. An @reply can sometimes consist of a “You suck” message, or worse (after reading such a message a human observer might suggest that the recipient has lost influence, whereas an algorithm might assign points for the @reply).

Yet both retweets and @replies are somewhat nano in scope… what about the more macro actions of the influenced, way beyond the realms of Twitter (and the other social networks that are tracked)? 

Reach

If somebody within your social circles retweets your tweet then maybe this is indeed a sign of influence. A surer sign would be if you are retweeted by somebody outside of your circle. 

But other factors come into play, such as reach. Aren’t people with a lot of reach are perhaps more likely to attract more retweets from strangers? In absolute terms, yes they are, such is power of the network effect.

Take Lady Gaga as an example. She recently surpassed the 10m follower mark on Twitter. This makes her the most popular person on Twitter. But popularity is not the strongest sign of influence, only of reach. I have no doubt that Gaga is influential, but it is perhaps too early to make sense of the precise nature of her influence. Also, she never seems to @reply or retweet anybody, which isn’t a great sign for automated measurement algorithms. 

Audience

Who are you influencing, and how? This is, for me, one of the key things to consider. Consider the fate of the Velvet Underground, who while active as a band only sold a handful of records and had almost no reach. But they had huge influence and deeply inspired the likes of David Bowie, Roxy Music, The Fall, Pixies, Kraftwerk, My Bloody Valentine and so many other (influential) artists.

The same thing is said of the Sex Pistols gig at the Manchester Free Trade Hall. Everybody who was there apparently went on to form a band (The Smiths, Joy Division, etc). And yet there were less than 40 people in the audience. 

Influence and reach are hugely confused. To properly understand how influential somebody (or something) is you need to measure the activity of the people they have influenced, and the nature and scope of that activity. It’s not a numbers game in absolute terms, only in percentage terms. And some actions are much more powerful than others.

Short vs long term

There’s something to be said for having a lasting influence, rather than being influential for half a summer (not that brief flash in the pan trends are to be entirely discounted). Bebo was hot for a while. So was Debbie Gibson. Are they still considered influential? Were they ever?

Some definitions of influence from the crowd

Earlier today I invited my cherished Twitter followers to define influence, in 140 characters or less. Here’s what they suggested… there are some common themes:

@LordManley The extent or capacity to effect changes in behaviour of others.

@krisdoubleyou The act altering of another person’s opinion, either deliberately or subconsciously.

@Florina_B The ability to have a message with a high impact, depending on the channel you use, looking fwd to that blogpost.

@mari7305 @Econsultancy @lakey impacting another’s thinking, choices, and preferences?

@BenjMartin Having an altering effect on other peoples actions, thoughts or behaviour through own actions, thoughts or beavhiour.

@_RosieT Worth listening to.

@tlonuqbar Being thought worthy of attention on a given topic by informed insiders.

@timcaynes It’s what those that don’t respond to definition requests have.

@whdigital Galvanize or sway perception, opinion or action. How? Size of network, strength of medium or sheer determination.

@chummers69 The ability to create interest and the effect others.

@MarkJWBoyd The capacity to have an effect on the character, development, or behavior of someone or something, or the effect itself.

@TomOCrowley Inspiration leading to action.

@vysia Ability to cause action.

@rishil The ability to nudge peoples perceptions away from or towards preconceived ideas through ones own reputation.

@PiponSolutions Influence is the ability to exert and maintain power on the actions and behaviour of others.

@katyhowell Power to impact opinion, ideas, habits, behaviours, motivations, actions, and attitudes.

@Trendshed “Influence”: Small fish bosses a large pond 😉

What do you think? Is it possible to measure influence in an automated way? 

[Image by Funkdooby via Flickr, various rights reserved]