I typically don’t veer too far from web analytics on this blog, but a couple of news stories have caught my eye this week. One has the headline Ex-Google employee says Google+ has ruined the company (Source: CNN) and Goldman exec quits, calling firm ‘toxic’ (Source: CNN).
What is remarkable about these stories is that they are the opinions of two single individuals, yet somehow they were able to make national headlines. I suspect a lot has to do with whether we agree with those opinions or not. If we are confirmed in our beliefs then it must be true, even though all we have is one data point. Actually, it’s not even “data” – it’s an opinion.
I am reminded of a memorable passage in Daniel Kahneman’s book Thinking, Fast and Slow:
“Subjects’ unwillingness to deduce the particular from the general was matched only by their willingness to infer the general from the particular.“
Kahneman further states that there is a “deep gap between our thinking of statistics and our thinking about individual cases”. Isn’t this what is going on with these news stories? People are of course entitled to their opinions, but are we too quick to assume that they are true and general?
I actually think that we as outside observers have basically no relevant information at all to form a decent opinion. We don’t even know what we don’t know about any of these companies. It’s like being able to only see a single visit with a 100% bounce rate in your web analytics and then inferring that all visits must have a bad bounce rate.
Thankfully we have more data!
Follow me on 
{ 3 comments… read them below or add one }
Hi Michael
Whilst there’s obviously an element of inference of the general from the particular, I think there’s another potentially more important element at play in both these stories.
Both companies carefully, some would say – obsessively, limit their staffs’ ability to speak publicly about the company; this has the effect of making unfettered comment on the companies a scarce commodity; scare commodities tend to be regarded as more valuable.
Time will tell whether the views of both individuals were prescient…
I agree with Gideon. I think the fact that we only have single data points in each case still leave either probabilities: Google/Goldman are fine, or, they have both become enamored with money. Should we believe them 100%? No, but as you said: “We don’t even know what we don’t know about any of these companies,” meaning, we may not know that there are others who feel the same way.
Thanks Gideon and Adrian for taking the time to add another great perspective to this. These are companies that we love to hear about and we are drawn to that.
You bring up another good point with “Time will tell…”. Not only do we only have two data points, they are also at one specific moment in time. So we can’t even detect trends – are things getting better or worse? Gideon, I disagree with you using the word “prescient” in this case. Both individuals have a 50/50 chance of being “right”. If we actually bother to come back to check their original “prediction” it’s just chance. There is no causality, merely a lucky correlation perhaps. We are extremely good at explaining the past and look for causal links in our stories.
I submit to you that these individuals have their own biases (which we don’t know about). We all do of course. I am biased whether I like it or not. These individuals could have left their positions for a number of subconscious reasons, but then rationalize their decision after the fact. Same thing happens with voting. We’d like to think we pick candidates based on facts, but maybe it’s more about whether we “like them”.
Cheers,
Michael