Kool-aid != Dogfood, and Certainty is no substitute for knowledge.

Palko quotes from this news article by Shirin Ghaffary quoting Marcelo Claure, the new executive chairman of the recent Ponzi scheme WeWork.

Here’s Claure, in a speech to WeWork employees:

[W]e got to drink our own Kool-aid, we got to make sure that if we’re selling this magic to others, we got to have this same magic in our spaces, in our first-floor employee workforce. So you can rest assured that what works stays and what doesn’t work, you know, we’re going to change it and we’re going to innovate to make sure that we have a very high-satisfaction workforce. And we’re going to measure that because it’s easy to say you have a happy workforce, easy to say you have a great culture, but we’re going to measure it and we’re going to be very honest with each other.

The bold font came from Palko, who was struck by Claure framing as a benefit something that is typically thought of as a problem, which is people believing their own hype.

Kool-aid != Dogfood

What struck me in Claure’s quote above was that he said, “drink our own Kool-aid,” rather than the more familiar (to me) phrase, “eat our own dogfood.”

Before going on, let me clarify the difference between drinking your own Kool-aid and eating your own dogfood. Kool-aid is a reference to the people in Guyana who committed mass suicide by being forced to drink poisoned Kool-aid. If you’re drinking your own Kool-aid, you’re committing suicide. In contrast, when you’re eating your own dogfood, you’re using your own products in your work. I guess in the WeWork context, that would mean using WeWork spaces for WeWork events. If there are statisticians at Microsoft doing their data analysis using Excel, they’d be dogfooding it. Using Excel would seem to be a poor choice for statistical analysis; on the other hand, if the top researchers at Microsoft were forced to use Excel for all their analyses, this out of necessity would probably motivate them to improve the product. Or I guess they could argue that Excel isn’t intended for serious work. Similarly, if the business is actual dog food, then dogfooding it would be feeding the product to your own dogs, not eating it for lunch.

OK, we digress. The point is that drinking your own Kool-aid is not the same as eating your own dogfood. If you’re drinking your own Kool-aid, you’re either suicidal or you believe your own hype.

But, in most aspects of life, you can do fine drinking your own Kool-aid, because most decisions do not have life-and-death consequences. Drinking your own Kool-aid is typically associated only with overconfidence and some loss of efficiency. For example, suppose that disgraced monkey researchers drinks his own Kool-aid and believes his own falsified data: so what? This has no real-world consequences. Or if that ESP researcher from Cornell drinks his own Kool-aid and uses his purported powers or precognition to decide whether to bring an umbrella to work . . . so what? Not much is lost compared to the more science-based approach of looking up the National Weather Service’s forecasts. Or if that sociologist from London drinks his own Kool-aid and makes reproductive choices based on his statistically unsupported theories of what predicts a boy or girl baby, or if that disgraced food researcher from Cornell drinks his own Kool-aid and serves himself food in small plates as a dieting aid, or if that celebrity linguist from Harvard lives his life according to the discredited “positivity ratio” theory . . . so what? No harm in any case; indeed all these people may well be happier by living their lives in consistency with their scientific ideologies (just as, for example, a socialist might be happier living with fewer personal possessions or a libertarian might be happier foregoing some government benefit, even if in some direct sense it might seem to make their lives easier).

On the other hand, if the goal is to run a big company, “drinking your own Kool-aid” could cause real harm. The tricky thing is, as Palko has discussed in other contexts, once lies and propaganda flow into a system, it’s hard to keep them separate from legitimate communications.

To put it another way: the big issue with WeWork has been the principal-agent problem, also known as the Other People’s Money problem: executives get billions of dollars to play with, and they (the executives) get to keep some of it. This creates a motivation to increase cash flow as then you can siphon off more for yourself. From that perspective, so much of the story is trying to fool other people. So there’s no real reason to think that Claure has any intention of drinking the Kool-aid; he might well just want to serve it to others. But he’s willing to tell people he’s drinking the Kool-aid, which is already a bit off.

The problem with a “what works” approach to decision making: Certainty is no substitute for knowledge

Now let’s look at the other bit from the above-quoted speech:

So you can rest assured that what works stays and what doesn’t work, you know, we’re going to change it and we’re going to innovate to make sure that we have a very high-satisfaction workforce. And we’re going to measure that because it’s easy to say you have a happy workforce, easy to say you have a great culture, but we’re going to measure it and we’re going to be very honest with each other.

This sounds good, but the problem is that we typically don’t know what works. Phrases such as “you can rest assured” and “to make sure” are nothing but cheap talk. Certainty is no substitute for knowledge.

P.S. Also this from Claure: “I’ve never been introduced by somebody that’s taller than me. (audience laughs) That was like, that’s good. It makes you humble.” He suffers from tall person syndrome! Only a tall person would think, even as a joke, that standing next to someone taller than you should “make you humble.”

P.P.S. Bonus points: Claure is a bullshitter! Check out this quote:

I come from a country, from a tiny, small country, Bolivia. . . . I was part of a team, of our Bolivian national team, not playing, but part of the management that took Bolivia to the World Cup. To put things in perspective, Bolivia to the World Cup is like a middle school team winning the World Cup. It’s not supposed to happen. It’s like the high school NBA winning the World Series, never happened. It only happened once.

Heyyyy, is that really true? Let’s go onto Wikipedia and check which Latin American and Caribbean countries were in recent World Cups. We’ve got Costa Rica, Panama, Honduras, Paraguay, Ecuador, Trinidad . . . Trinidad! OK, fine, it’s easier to get in from Central America and the Caribbean, so let’s not count them. But Paraguay and Ecuador: they’re small countries too. Good for Bolivia to have made it to the World Cup finals in 1994, but it’s not that big a deal. Nor is it true that “it only happened once”—it also happened in 1950 and 1934. No big deal, but . . . jeez, what a bullshitter. What is it with these rich guys—they just think they can make up their own reality?

30 thoughts on “Kool-aid != Dogfood, and Certainty is no substitute for knowledge.

  1. The problem is Wansink was head of government food programs and by publishing Gus research got other people to take him seriously. Similarly, Ben, by publishing in a reputable journal, at best wasted other people’s time by misleading them.

  2. According to Marcelo Claure’s wikipedia page, he graduated from Bentley College in Waltham, Massachusetts in 1993. Now, the qualifiers for the World Cup take place over the four years prior to the World Cup. So, if he meant that he was involved in the management of the last Bolivian national team that qualified for a World Cup in 1994, then it really is an impressive accomplishment that he did this while in the United States in college. At this point, these business guys need to learn to go full Trump. He should have just said, “Last year, I lead my tiny country of Bolivia to victory in World Cup. I’m going to bring the same great spirit of confabulation to WeWorks. Please don’t look to hard at our financial statements.”

  3. Um, Bolivia did not make it “to the World Cup finals in 1994,” which really would have required one of the most spectacular runs of upsets in sports history. I think you’re confusing Bolivia with Brazil.

    • Phil:

      From Wikipedia: “The tournament consists of two parts, the qualification phase and the final phase (officially called the World Cup Finals). . . . The World Cup Finals is the most widely viewed sporting event in the world . . . Bolivia have qualified for the Finals on three occasions, in 1930, 1950 and 1994. They have played in six matches at the Finals, but have lost five and drawn one, with their only goal coming against Spain in 1994.”

      • Any actual soccer fans who say the “World Cup Finals” most likely means the final game whose winner is the winner of the world cup. So, this notion of the finals is a kind of jargon. Sure, they made it past the qualifying rounds so that they could “play in the world cup” (which the US didn’t even do last time) but it’s jargon because everyone will expect “The World Cup Finals” to mean *the final game*, contrast for example “the semifinals” which is the round before the finals where there are 4 teams playing 2 games to determine “who’s in the finals”.

        In 1994, Bolivia made it into group C but didn’t leave the group stages, as Germany and Spain advanced and Bolivia and South Korea went home. https://en.wikipedia.org/wiki/1994_FIFA_World_Cup#Group_C

        In 1994 in the Final, it was Brazil vs Italy 0-0 and decided by penalty kicks 3-2 in favor of Brazil. https://en.wikipedia.org/wiki/1994_FIFA_World_Cup#Final

        It’s kind of like saying “I ate breakfast today” when you slept in until 5PM and then had steak and broccoli. technically this meal was “breaking your fast” but it’s incongruous with what everyone expects from the statement “I had breakfast”

        • Daniel:

          Believe it or not, I thought about this issue when writing the post! I used the Wikipedia formulation because it figured that must be the standard way to write about it.

          In any case, I think we can all agree that Claure was bullshitting.

        • No way, the final game is the world cup final, not finals. No non-American would ever say ‘world cup finals’. When you get through qualifying you qualify for the ‘world cup’, not the ‘world cup finals’. That’s what FIFA call it, but nobody else.

          The semi-finals is correct if you’re referring to both games collectively, otherwise you get through to the semi-final (singular), pronounced ‘semmy’, not ‘SEM-I’ obviously.

        • Most Americans would find “World Cup Finals” awkward as well… They’d say “the final” or “the final game”, but you might infer that if you won the final game you also must have won the quarter finals and semi finals round as well… so you “won the finals”. And saying “went to the finals” would probably make most americans think you at least made it to the quarter-finals or semi-finals games.

          In any case, I totally agree with you when you say: ” When you get through qualifying you qualify for the ‘world cup’, not the ‘world cup finals’. That’s what FIFA call it, but nobody else.”

          but obviously it’s what FIFA *calls* it ;-) you Brits and your collective plurals

      • Well, I certainly stand corrected. This will make a great bar bet — did Bolivia play in the Finals of the World Cup in 1994? I will never have to pay for my own drinks again!

        Bob makes a good point with ‘Finals’ vs ‘Final’ but at least in America (or at least in the U.S.) I agree with Daniel that very nearly 100% of people here will think you mean ‘the championship game’ if you ask about the World Cup Final(s), whether the s is on there or not.

  4. Bolivia did not win a game in the World Cup. Soccer fans don’t think of the qualifiers as a stage of the World Cup. We think about the group stage,and if you get out of your group and get into the elimination round that’s the big accomplishment. Bolivia didn’t get out of its group, and I doubt that Marcelo Claure could have been involved in the management of the team while he was in college in the states.

  5. “the problem is that we typically don’t know what works.”

    No, that’s not true! :)

    Most of the time problem is that we don’t like the answer. We’re still searching for a way to have our cake and eat it too. We just don’t know what works for that! :)

    Let’s illustrate why we can know “what works” and not do it. Take Seattle, San Fran, LA, and some other cities with a homelessness problem.

    They’re purportedly looking for the “what works” to solve the homelessness issue. The citizens of LA are building appts for homeless people that *cost* more than market rate appts rent for. Does that work? No. Do we know that doesn’t work? Yes. Alex Tabarrok (marginal revolution) has repeatedly pointed out that the rate of home construction in the Bay area is lower than the rate of job growth. Does that work? No. Do we know it doesn’t work? Yes.

    So what works? Why, it MIGHT work to build housing at a faster rate than the rate of population or job growth, ya think? :)

    So a lot of the time we actually ****do***** know what works, but powerful people don’t want to do that because what works for the general population doesn’t work for certain people to increase their political power.

    • Yes, for the most part this. We know what works to achieve lots of things, we just don’t know what works to get past the political and power barriers to making that happen. Eventually if it gets bad enough, you wind up with revolutions…

      • Yes, we DO know what works. One reason I have stayed on the periphery of some issues is b/c I haven’t liked academic politics much. Plus I find the whole exercise of speaking before audiences a means to maintain the gatekeeper role when it is clear that the several crises of knowledge are due career incentives and disincentives. The reality is that domains are expanded by daring eclectic intellectuals, a point over which there is quite a bit of cognitive dissonance.

        However, there are some exceptional thinkers even today that have opportunities to solidify their influence in a very crowded market of ideas. Some practical humility has embedded in some circles.

  6. “Kool-aid is a reference to the people in Guyana who committed mass suicide by being forced to drink poisoned Kool-aid.” It is my understanding that the People’s Temple members who died in Guyana voluntarily drank the Kool-Aid (though they might have been dubious that it actually contained poison), and that “drink the Kool-aid” does not refer to suicide in general, but rather to mindlessly buying into an ideology.

      • If my memory is correct, the people doing the “forcing” or “pressuring” drank the Kool-Aid themselves. The purpose of that figure of speech, properly deployed, is to evoke the horror of collective self-annihilating action by a group blindly following a delusional yet charismatic leader.

        • Brent:

          Yes, this sounds about right. Claure’s statement, “[W]e got to drink our own Kool-aid, we got to make sure that if we’re selling this magic to others, we got to have this same magic in our spaces, in our first-floor employee workforce,” is potentially disastrous if “this magic” is not real.

        • I thought it might be helpful, or at least kind, to give Claure the benefit of the doubt that when he said “drink our own Kool-aid” what he meant was “practice what we preach.” Then again, Jim Jones was a preacher….

        • Exactly, the objective of both hype and preaching is to garner the uncritical faith of the recipient in a message or enterprise. If they have faith in the product, it follows that the product will improve the company, which will ultimately show up as better numbers. But he’s not just saying they will improve, he’s suggesting a new, internal measure of improvement to supplant external, market-based measures (stability, profitability, valuation). That mechanism, CEO judgment fed by internal measures, happens to be completely within his control. He’s selling it as inherently superior to those other metrics because his judgment has been proven somehow in the performance of a soccer team and will rely upon *honest* measures, whereas external metrics rely on superficial data. Presumably, his real objective is to distract employees and investors from the bad numbers long enough to effect actual changes outside the view of critics. Or if he drinks his own Kool-aid, maybe he’ll change nothing and wait for the external mechanisms to return to the “correct” evaluation.

    • Oh, man. This has me feeling my age. I always thought “Drinking the Kool-Aid” was a reference to Tom Wolf’s “Electric Kool Aid Acid Test” (1968). Which makes more sense as a metaphor for accepting and believing craziness.

      Jonestown was 10 years later (you whippersnappers), and was, as Sparky points out, Flavor Aid, not Kool Aid.

  7. Your comment that we don’t know what works (and the minor controversy in the comments) reminded me of the What Works Clearinghouse, a website established by the US Dept of Ed to advise educators on what constitutes evidence-based interventions. The WWC defines “working” as having a positive effect estimate for a particular population and outcome based on results from all reported RCTs and high-quality quasi-experiments. In most cases, an intervention that “works” has been evaluated only in a single study or set of studies by the same coauthors. So really, it doesn’t so much tell you “what works” as “what has a reasonable chance of working after excluding all information but evidence that meets the highest standards.” That’s a good goal, a practical goal, the kind of thing you need as a non-expert being forced to make a choice. It’s dogfood for policymakers. But the epistemological claim that these programs are what works is just Kool-aid. Or to use a different analogy, it’s using science as the feather that convinces Dumbo he can fly. All of which would be only incidental to real science except that even many scientists don’t make this distinction, particularly if they are in some way invested in the success of the intervention, and even those who do distinguish between practical advice and the gradual narrowing of uncertainty that is science are being incentivized to work and write as if they don’t.

    • “…many scientists don’t make this distinction, particularly if they are in some way invested in the success of the intervention…”

      hoooo, yeah! Especially in education. Lots of people with conclusions looking for data.

      What I love about education advocates is that they’re always telling us that teachers are so great and their effect so profound that you can’t even measure it! NO NEED TO TEST! CHILD CENTERED EDUCATION! DEAR MICROSOFT PLEASE HIRE FROM (country where testing is mandatory)!

      • I’d say there’s a spectrum among researchers from “teaching is an art” to “education can only be assessed qualitatively and/or subjectively” to “education effectiveness can be measured quantitatively.” We have a large contingent of ed researchers on the quantitative end–almost everyone who applies for grants from the Institute of Education Sciences, for example. Many of us aren’t in university education departments–some are in departments of psychology or statistics or evaluation, or at research orgs. Practitioners are more likely to fall in the first or second categories, and their voices may seem dominant because their message is aimed directly at the public, not filtered through journals and conferences (and stats blogs!). Policymakers seem to prefer the quantitative end, although it seems to me that they often define measurement without consulting quantitative researchers. Or dictate to the psychometricians they fund what measures ought to look like. Or prioritize minimizing cost and maximizing bureaucracy. (You might think cutting bureaucracy would be an easy way to reduce costs, but it’s far cheaper to fund an office of people charged with a mandate than it is to fund the mandate.)

        • “(You might think cutting bureaucracy would be an easy way to reduce costs, but it’s far cheaper to fund an office of people charged with a mandate than it is to fund the mandate.)”

          LOL! Foundational concept to the education bureaucracy: hire more administrators to reduced costs.

          I can believe, for sure, that the less quantitative oriented people are the most outspoken. They have a lot to lose.

          WRT testing: WA State, home of Microsoft and Amazon, just eliminated graduation testing for high school students, bcz certain interested groups were clamoring against “high stakes” (e.g., useful) testing. Meanwhile, MSFT, GOOG, FB, AMZN and all the tech companies here are importing people from all over the world, but notably from places that have testing.

Leave a Reply to Brent Hutto Cancel reply

Your email address will not be published. Required fields are marked *