Comment by asark | Hacker News Reader

Comment by asark | Hacker News Reader

Parent Full thread

My experience is that most people in business are 1) terrible at constructing trials, tests, or experiments, of any sort whatsoever, 2) terrible at interpreting the results of same, and 3) have a very limited tolerance for discussion of how bad they are at it and why their various tests are meaningless, no matter how gently it's presented.

I've not seen a place where this isn't broadly true, no matter where the business folks come from (top-tier B-school analysts, Bob's discount MBA emporium, doesn't matter) or how good the reputation of the firm they're at. This also extends to management and administration in the public sector—schools are rife with obvious pseudo-science bullshit, bad attempts to replicate results while skipping half the measures that were taken, et c., but dumb-ass superintendents (you would not believe it, seriously, "dumb-ass" was not chosen lightly) and principals eat it up.

[EDIT] which is to say I'm not at all surprised design "science" is full of BS, because it's sold to people who almost all suck at evaluating those kinds of things. Any "science" that largely exists to sell stuff to middle- and upper-management or "stakeholders" probably tends to be awful, because it doesn't need to be good.

Remember that academia has a longer history as an accreditation vehicle for any concept you can come up with as long as you write convincingly enough, than as a platform of science since academia predates science by centuries if not millenias.

I think this tradition of authoritarian bullshit has some traction in the modern world still. Which in part causes all the silly pompous pseudoscientific bullshit.

Completely agree. In my experience every client I've worked with that wanted 'User Testing' or 'A/B Testing' would have lofty goals of having an awesome product backed by science, but at the end of the day were unwilling or unable to actually execute it in a way that did anything other than smush the data into the shape they wanted. They already have results in mind, they just want some data to support their conclusion so they can put it up on a slide for 10 seconds in a meeting.

I would go much further and say almost all "science" is done poorly.

The scientific method, when applied rigorously (!), is deeply uneconomical in the vast majority of cases, because the market has already eliminated most glaring inefficiencies by pure trial and error. Any possible gains are likely marginal anyway.

Doing a bunch of half-assed and meaningless experiments just to slap the prestigious "science" label on something on the other hand is very economical. So that's what people do instead.

Also, if you've studied human psychology in depth, you realize that the scientific method is a mirror of a very specific set of traits "owned" by a subset of humans.

If you're on the outside of that subset, it turns out that not only is it easier just to slap the label on, but your ideal customer (relative to your psychology) doesn't really care to do much more than a simple label check anyway.

From zero to 'rm -rf science-budget' in one marketing exercise.

A sad truth but coming around to it can help us understand why scientists need to work on social policy and outreach messaging as well. Scientific values simply will not sell themselves the way we think they should, outside of our little sphere.

Fortunately our cultural messaging has vastly improved in this domain recently. Bridge psychologies like that of Bill Nye are great examples of how this can work.

I'm also interested in what motivates this behaviour. Anecdotally I feel that the current incentive structure is entirely financial. People may get into a field to do good, but they have to justify an ROI on everything they do. Society fetishizes financial success and universally condemns financial failure.

Live in that world as a struggling student scientist, juggling the prospects of a "marketable" thesis versus an ethical one, as debts pile up and staring down the barrel of your entire life amounting to being yet another mediocre drone / living on the streets / not being a scientist at all.

The prospect of being completely broke, unable to afford rent, unable to afford food, being an embarrassment to friends and family, these factors exist and must surely inform the judgement and behaviours of students in a position to have a study funded.

I know of phd's still struggling in their careers against the same thing, where an institute is more interested in publicity than the actual science.

The symptom may be half-assed and meaningless experiments but that is not to say the individuals responsible aren't capable or willing of so much more if the appropriate supporting social structures existed to allow ethical scientific thinkers to exist, or if institutions stopped taking in students for the money and instead tested applicants for profit.

It's easy to point out the problem but how can it be solved? To me, I think there needs to be publicly funded institutes that have no profit motive and exist purely to intake the best academic minds, this might seem anti capitalist but as a model it is an investment in the future of a society that is calculated in quality rather than quantity.

By and large, profit does represent societal good, this whole "anti-profit" attitude is misguided.

On the other hand, it is often possible to profit unethically and sometimes that is the most expedient way. There is a cost to this, which is reputational. That's why it's important to call out bullshit.

The problem with research is that the business risk is high so the potential rewards need to be equally high. So there is an argument to be made for large entities (up to and including the state) to fund research, but there need to be checks and balances. The ones we have may not be perfect, but that doesn't mean they're inadequate. It's all just overhead as far as I am concerned.

Knowledge is a product, we should not only be a geek and leave the profit to the others voluntarily. All scientists should stand up to learn how to raise funds, market, sell, and refine. In fact, the business side of the effort is also very difficult and professional. The effort can further help reshape the research. Having a small team can reduce the bureaucratic requirements of prestige pseudo-scientific stamp but enforce the true scientific methology to solve the problems.

The business relationship among people is becoming flatter instead of deeper. Tooling is going to help individuals achieving more independently. At least it's the reality I would like to live in.

a depressingly accurate assessment.

> Any "science" that largely exists to sell stuff to middle- and upper-management or "stakeholders" probably tends to be awful, because it doesn't need to be good.

as long as it's reeaally expensive. that's really the only way to asses the accuracy and importance of these things.

[flagged]

So... my business school covered all of that. Probability, statistics, linear algebra, diffeq... and yes, boolean algebra in the context of gradient descent. Even Markov chains and neural nets. Needless to say, there was a lot of math.

Interviews at top business schools are nothing but partitioning and mathematical analysis. Google “business school case interviews”. Modern business is analytical and quantitative. Over 50% of my classmates came from software development or engineering backgrounds. This mix is not at all unusual.

I think your reaction may be to the PHB types — who aren’t good at any of that but they have soft skills (aka people like them) so they get promoted. Honestly the biggest truism in business is that it’s better to be liked than to be good.

>So... my business school covered all of that. Probability, statistics, linear algebra, diffeq... and yes, boolean algebra in the context of gradient descent. Even Markov chains and neural nets. Needless to say, there was a lot of math.

That's an incredibly rare thing. To date, the only systematic exception to the above rule I've encountered have been from certain European schools (mostly French: HEC, ESSEC, etc) that have a math-heavy prep-school curriculum to get in.

From what I've seen though, this is very much exceptional.

A subset of business schools are considered “quant” schools. These schools tend to be pretty high up in the rankings as their faculty often end up winning Nobel prizes.

What school?

What are one-dimensional business people actually good at? Besides just being a piece in the system that soaks up resources?

The good ones are good at convincing, negotiating, charming, manipulating, uniting people, turning people against each other etc. All very important skills out in the real world of human interactions and they translate to $$ quite directly.

Learning and exploiting unwritten rules of a game referred to as business.

https://en.wikipedia.org/wiki/Gamesmanship

Making deals with similar people at other companies.

They're good at talking to people and convincing them of things, which is enormously valuable in business.

Reply on news.ycombinator.com