Influence measurement is recipient-led, not creator-led
There’s nothing like a bit of nonsense to jolt me back into blogging action. The specific nonsense comes from, surprisingly, Forrester, for whom I usually have high regard.
Josh Bernoff and Augie Ray have created Peer Influence Analysis, an attempt to measure the extent of influence occurring in the online world. Brave.
They claim that 256 billion influence impressions are created every year. Wow.
Their methodology is thus : “Start by counting every instance in which a person influences another person online about a product or service. (We model this from Forrester’s 10,000 person survey, which asks how frequently they post, in what places, how many followers they have, and what products and services they post about.)”
How do they model influence? They are measuring outputs of people creating opinions and reviews online. But this has no bearing on whether anyone did something different as a result of reading it. In other words, all this output may have had zero influence.
Or, some reviews (and reviewers) may be more important than others? One negative review might have more weight that ten positive reviews. How can you tell? You can’t, based on this method.
The key weakness here is that there is a disconnect between one’s intention to influence and any actual influence created. We all know people who try to influence us but fail: they’re called bores. Social media is full of such people.
Influence is dependent on whether the recipient of content change their decision or behaviour. It is not dependent at all on the creator of that content or their intention to influence.
But surely all of these 256 billion influence impressions must have some impact?
This reminds me of a Hugh Macleod post, which explains why it ain’t necessarily so. One percent impact from 256 billion influence impressions is a lot of influence. But what if you only get one billionth of one percent? Or zero percent?
Measuring influence is really hard. Measuring output is possibly a good indicator of a medium, but it’s too coarse to provide real information. Which of the 256 billion influence impressions are really changing behavior? That’s the real question.
Well said, Duncan! I think the title sums it all up.
Any meaningful business measurement should be outcome-based and not activity-based. Measuring influence by outcome, which is what Influencer50 is doing in B2B and Pursway in B2C, is not trivial.
It’s easy to see why Forrester and many others are rushing to “analyze” influence based on tweets, posts, “followers” and “friends”. It’s the visible portion of social interaction, but it’s like looking for the lost coin under the lamppost just because that’s where the light is.
Is there influence in these interactions? Probably some. How much? Nobody knows, but there is good evidence to show that not much. Think about yourself. Where do you get your trusted advice? On Twitter or when you actually talk face to face or on the phone with your friend (the real one, not just a Facebook “friend”)? This intuitive conclusion is also backed by studies Pursway and others have done.
Is the Forrester offering nonsense? Maybe. There are some companies that don’t fit into the Influencer Marketing models of either Influencer50 or Pursway. For these companies, the Forrester offering may be the best they can do.
What we do know is that for companies that can measure influence at the outcome level, the results are extremely rewarding. I would be curious to hear what kind of results companies get using the Forrester service.