I blogged in 2011 on my Predictive Index (PI) taken while at ThoughtWorks. I often try to explain it in an elevator pitch format. Here attempt to do that again, for posterity.

There are four types of people you might try to sell concepts to - each buys concepts differently, and it is important to match your sales technique to match their buying style, even if that is alien to the way you yourself would buy a concept. ThoughtWorks were doing this to give you tools to use in client engagements and pursuits. It is as useful with colleagues, of course.

I was informed that I was mostly Expresser, and a wee bit of a Controller. I don’t hugely understand the latter despite being said to be a little like that. See below.

Four Stereotypes

So, here is an attempt to way too briefly stereotype the four types identified in a PI test. Consider a scenario where they engage in a ‘Coke versus Pepsi’ challenge, and have to say which is better.


Given the choice of Coke or Diet Pepsi, and a requirement to choose one for everyone, the collaborator will ask a number of trusted lieutenants what their preference is and why. The collaborator can progress towards decision themselves that way. Indeed, the collaborator may be able to repeat the rationales put forward by the trusted lieutenants, even though they are new to the arguments.

Note: Paralysis can happen for collaborators, if the trusted lieutenants are not in a clear majority on the choice between the options.


The analyst will put together a spreadsheet with a bunch of rows for criteria and one column for each of the choices. They’ll measure (and score) effervescence, color, pH, caffeine, etc, before making a decision between the two. Other PI types might feel they take too much time to decide.


The expresser might not even open the cans of diet coke/pepsi in the extreme case. They might merely look at the ingredients and be able to make a decision based on past exposure to similar things, and bunch of reading they may have done up to that moment. For example, they could rank based on the prominence of controversial sweeteners. Other PI types might wonder how in the hell that person can make a decision without opening the can.


Makes a decision, or held an opinion already - is kinda motivated to ensure that everyone in the hierarchy below them agrees, and might get annoyed when people in the reporting structure below them don’t agree.

Argh I got stuck there. Pretty lame.

Ross Pettit on Controllers

Being stuck on the stereotype definition of ‘controller’, I asked Ross Pettit (who also received the training) to weigh in. Actually, I’ve been haranguing him for years to participate in this article. Ross has gone to town with an extreme, but delightful, stereotype definition of a controller:

In the 1970s, Pepsi introduced “the Pepsi challenge” to try to establish that people preferred Pepsi to Coke in blind taste tests, and by a wide margin. Thus they were trying to show (a) if you would only try our product you’d like it more (appealing to the analytical), and (b) because so many of you clearly prefer our product there’s strength in numbers so you don’t have to feel ashamed (appealing, somewhat, to the collaborator).

Yet Pepsi sales volumes didn’t rise to reflect the alleged taste preference of those surveyed.

Suppose a controller opts in for a taste test, and finds they prefer Pepsi over Coke. They’ve always bought Coke, they have fond memories as a kid in summer time running around and cooling off with an ice cold Coke with their buddies. Every time they drink a Coke they’re re-living that moment of their lives if only fleetingly. Where’s your taste test for that, Pepsi? You wanna go all Joe Stalin on history and rewrite my memories to say I was drinkin’ Pepsi as a kid?

So, it seems to me that the controller personality, who had already made up his or her mind to buy Coke, would set up vendor day as a blind test for all the staff. In the event, it is shown they prefer Pepsi. The controller would ignore the data and buy Coke anyway. If forced, would stock the shelves half with Coke and half with Pepsi and see what actual consumption patterns looked like. If necessary, the controller would manipulate the appearance of consumption to show that the Coke was being depleted at a far faster rate (by slightly under stocking, taking a can of Coke home every night, or surreptitiously restocking small amounts of Pepsi early in the morning).

The wife and I tasted both the Pepsi and the Coke in the build up to the second pic. We don’t really like either any more, versus Diet Coke/Pepsi - which we disagree on.

Additional Thoughts

Really the science of PI, is to:

  1. Understand the other types
  2. Realize that all types have blind spots in decision making
  3. Be able to sell to types other than your own, when the situation requires
  4. Not stuff your organization with a single type
  5. Generally have empathy for other types


January 11th, 2016