Tim Berners-Lee: You Can’t Quantify Potential Influence

Me and Tim Berners-Lee 2

The highlight of my SxSW experience this weekend was meeting Tim Berners-Lee, best known as the inventor of the World Wide Web. I took the opportunity to ask Sir Berners-Lee what he thought of social influence metrics like Klout, Kred and PeerIndex.

His response was remarkable, but before I share it with you let me frame the scene.

Sir Berners-Lee is clearly a savant. He is so brilliant he struggled with the bloggy attention he received at the IEEE SxSW reception. When he talk, he gestured somewhat wildly, and was clearly aware of the surrounding cameras. It was exactly how I imagine Einstein would function in this 21st century world of cameras, tweets, and instant access… Like a brilliant wild person forced to live in a zoo.

I immediately recognized time would be short with this man, that he would move on quickly. So I listened intently.

Sir Berners-Lee responded (I am paraphrasing), “Of course, you can look online and quantify everyone in this room by what’s said, and then line them up in a numerical order of magnitude. But how can this number possibly predict all of the things that these people will do, and which ones will have the most impact? You can’t!

“I am very much against these metrics. I don’t see how they make a difference for us.”

He was clearly a bit annoyed by the idea, and his facial expression turned from intense and serious to a bit sour, and even a tad angry.

The statement was brilliantly concise and spot on. Potential is an empty bucket, and past successes — successes that are currently defined and measured by attention online according to these metrics — cannot predict the future.

In fact, we see many people enjoy a singular moment of public success, and then slowly fade to black. Consistent successes over a period of time distinguish true masters in any field. Replicating excellence remains one of the hardest things for any person to achieve.

So that was my minute with Sir Tim Berners-Lee. And just like that he was off to chat with the next adoring geek.

What do you think about Tim Berners-Lee’s view of social scoring metrics?


  • I agree that if you are going to quantify influence you had better be thoughtful and specific about how you define it. That being said, I don’t think anyone defines it as “all of the things that these people will do, and which ones will have the most impact.” Also, there is a difference between predicting the future and saying what is likely to happen. I can say, for example, with reasonable certainty that if Oprah recommends a book on her show, a bunch of people will buy it. This bolsters your “track record” argument and indicates that, at least in some situations, we can estimate the impact the actions of one will have on many.
    Also, I miss you.

    • That’s true, past performance is an indicator. It’s why athletes get those big contracts, right? And some earn their money through and through. Then there is A-Rod. LOL!

      In all seriousness though, if you look at pro sports the most quantifiable of any business out there, I do think you see how time impacts the ability to deliver consistently year in year out. And we always see new players rise up. That’s the core issue, I think. Who falls, who rises.

  • I love his concise statement. But that is equivalent to suggesting that infrared spectroscopy is worthless because it takes what is a 3 to 4D set of information and melds it into a 2D value. IR Spectroscopy is of the same ilk as “measuring influence” because it provides a quantity for the magnitude of particular frequencies. We lose the phase of the vibrating molecules and the dimensions of temperature and pressure are also lost.

    But that has not stopped chemists and scientists from deriving understanding of the physical realm from this blunt instrument. Heck, even the rover Curiosity is using techniques like this to understand Mars. And the potential — well, influence can be built up over time. The potential reactions of one chemical with another can be also discovered as well.

    And just like the use of IR, you use the information at your own strength or peril.

    • Exactly. Kind of like baseball stats. WAR, does it mean anything? We don’t really know, it’s a bit of an intangible, especially comparing pitchers and catchers with the same stat, yet the baseball pros love it! At their own strength or peril.

  • Ironically, it’s pretty much similar to the thoughts over on the post from our book blog that went up last week:


    Today’s “influence metrics” – social scoring like the three sites you mentioned – are simply starters of the process and bigger conversation. There’s absolutely no way you can quantify anyone based on a generic score that’s calculated with very limited information.

    Instead, you need to go much deeper, and look at way more complex data that we know about them from historical interaction, as well as predictive interaction. The predictive part isn’t 100% either – yet – but solutions like TrendSpottr.com and Tellagence.com are bringing us closer to that information.

    Like the picture, too. :)

    • The predictive part will never be 100%. Life and people are unpredictable. And influence extends well beyond data, to the human psyche and the conditions that trigger responses and actions from abiological perspective.

      Thus forcing an algorithm on the human to predict future behavior is like trying to slam a square peg in a round hole, over and over again. Mark my words, none of these solutions will ever achieve complete success. At best they will come as close to the science of pollsters.

      • Completely agree with the human psyche part as well as the predictive – you’ll find none of the quality platforms ever say they’ll get 100% accuracy, and the CEO of Traackr admits you need part machine, part human to analyze the data.

        I do believe there will be a time, perhaps not even that far off – maybe 18-24 months – where we’ll be at the stage of being able to dig much deeper into transactional relationships and how they can help predict the purchase cycle of someone. We’re not close yet, because there are too many variables, but the technology is catching up with the theory. Of course, then you add another level of permission marketing into the mix, which opens up another conversation altogether.

        Fun times!

        *EDITED: Just saw this link, quite an interesting study on 58,000 participants, and how accurate predictions can be.


        • I personally wouldn’t describe this all as adding up to “fun times” – I rather shudder at Jeremiah Owyang’s forecast, in a Six Pixels interview with Mitch Joel, of someday (soon) being able to see anyone’s Klout score floating above their head as your Google Glass facially scans them upon approach. I do appreciate the emphasis here on the predictive limits of any kind of scoring, any historical assessment. It all reminds me of the standard disclaimers in so many of the investment ads I created in my ill-spent youth “Past performance is no guarantee of future return”

          • That’s a great point, Chuck, although it could also be looked at as the one that finally introduces some privacy options into the equation.

            Currently Klout et al make the information about you public, and you have to go through various hoops to remove. But at least it’s contained to the online space.

            If we were to move to a vision like Jeremiah’s, I foresee far stricter rules being brought into play, and perhaps even regulatory boards set up to monitor this particular area.

            Something which would be welcome, given that employers are now starting to use these half-complete scores.

    • And, thank you for commenting. My colleague Erin McCahill took the pic. You can see him looking at the camera… He was so buggy about pics!!!

  • Sometimes I look at my children and wonder what they could do if they had unlimited resources.

    Would they become the next Steve Jobs or Einstein?

    I don’t know and then I wonder what resources do or not do for people because there are so many different variables and factors that are involved in all of these things.

    Sometimes the real utility of a tool depends upon who is using it. A hammer can be used to destroy just as easily as it can be used to create.

    • Well said, the strength of the tool really depends on the strategist. WHat I worry about resources like this is the same as I worry about SATs and my daughter. If she performs poorly does she think she is stupid, thus creating a downward script…. Such is the problem with all standardized statistics.

  • I couldn’t agree more. These rating services annoy me.
    I confess I did go and look at Klout and was bemused to see that it rated me for topics like pizza and camping, as well as what I actually share online.
    N.B. I think the camping one was because I’ve been writing abut unconferences, which often have ‘camp’ in the title, e.g. BlueLightCamp: http://BlueLightCamp.org.uk but to the best of my knowledge I’ve never mentioned pizza online (Damn, just did…)

    • Mark, I get all of my pizza advice from Chris Brogan, but I’d really appreciate it if you never quit doling out that excellent camping advice!

  • Thanks for the great conversation starter, Geoff. The biggest problem we have with these online statistics is that we’re too lazy to measure many of the things that could actually be measured, and we settle for a single number — Klout score, follower count, likes, connections — to summarize far more than is possibly realistic. Then we proceed to devote our energies on improving that number, rather than improving the things that really matter.

  • Geoff,

    When I get a surfboard in the mail compliments of Klout, I might start to care about it.

  • Pingback:Why on earth sponsor an Unconference? | markbraggins

    […] a brilliant short blog post from Geoff Livingston in which he manages an ever-so-brief conversation with the inventor of the Word Wide Web, Sir Tim […]

  • Pingback:Why on earth sponsor an Unconference? | BlueLightCamp

    […] a brilliant short blog post from Geoff Livingston in which he manages an ever-so-brief conversation with the inventor of the Word Wide Web, Sir Tim […]

Comments are closed