Hacker News Reader

Comment by philk10 | Hacker News Reader

Reminds me of the Orange Juice test from Jerry Weinberg - https://www.intercom.com/blog/the-orange-juice-test/ - asking if a conference hotel can deal with 700 people who have to have freshly squeezed juice at 7am

View on HN

Or the brown M&Ms test from Van Halen:

https://www.snopes.com/fact-check/brown-out/

A friend of mine once had to pull all of the brown M&Ms out of a bowl for a band. I can't say that they thought it was worth it, but we also followed all of the other rules in that contract when booking that band. We hired a different band the next year because that was a real nuisance.

David Lee Roth explains quite reasonably it:

https://www.youtube.com/watch?v=_IxqdAgNJck

I don't blame him at all, it's a safety thing and the M&Ms are one way to see if everything is being done properly

I don't blame anyone for stuff like this as a canary, but if you want the M&Ms, you had better be worth the M&Ms.

if the business agreed with the contract then... that's all on them? they didn't have to take the job.

Yeah, we agreed once, but nominally you would prefer to have repeat business. I assume that there are venues that either buy pre-made no-brown-M&M mixes or just don't book bands with that clause.

Van Halen got away with it because it's Van Halen.

Van Halen had his choice of cities and venues; a test like this is perfect for weeding down to the ones who will take the all the other seemingly-random-but-actually-vitally-important requirements around stage and equipment setup seriously.

It doesn't work for the musician hired just to have someone playing.

So the Orange Juice test is just finding out if clients are willing to say they'll charge you more for special requests? I can see that being practical sometimes.

It's about understanding if a person is capable of grappling with hard problems, if they can quickly pareto down to the 20% of the part of it that's going to be tricky (big conference -> breakfast -> early morning -> fresh squeezed). Anyone can quote a big top-line number for the whole shebang, but that just tells you they haven't thought it through and any "real" hard problems that materialize are going to be met with last minute, slapped together solutions.

As a mid-career software person, one of the most important things I've recognized is that my trust for my colleagues hinges on them knowing their limits and knowing when and how to escalate hard problems. I don't want to hear about small problems— those they can solve themselves. But they need to know which problems should become a quick slack message, a whiteboard chat, a revisiting of the design doc, or even questioning the fundamentals of the whole project. A person who slaps on duct tape when they should have engaged a war room discussion is someone who ends up needing almost constant supervision.

Stretching a picture to fit in 2024, on an article about heuristics for dismissal.

> The key idea here is to propose a task you know to be extremely difficult but possible and then measure the candidate’s reaction. If they are defeatist (“That can’t be done!“) or deluded (“I’d code that in a weekend“), then that’s what you’d be hiring.

Oh man, I'm saving this because this has been my experience working with folks on the business side who seem to always optimize for the second group (the deluded gives a low quote, then the project drags on forever, spiraling out of control because they didn't think through the difficulty).

If billed by hour, that's great.

I'm looking for something insightful here but struggling. This seems like it will lead to horrible (and expensive) outcomes.

tldr: It's a test for whether the consultant will take the request seriously and give a legit estimate (which will be a huge number). You can fail it in a number of ways:

1) Reject it outright as impossible. (shows unwillingness to consider difficult projects or refusal to let the buyer make the call about what is too much)

2) Assure the buyer you can handle it with no specifics. (shows willingness to overpromise)

3) Give a too-low estimate. (Shows technical incompetence and overoptimism.)

There's also "give a too-high estimate" which tells you that they are out of your league.

Ah, okay, I missed that one. But, to be pedantic, it says that's technically not part of the Orange Juice Test:

>>If your price had been too high, we would have eliminated you, too. But that’s a different test.

Quoting: "The key idea here is to propose a task you know to be extremely difficult but possible, and then measure the candidate’s reaction. If they are defeatist (“That can’t be done!“) or deluded (“I’d code that in a weekend“) then that’s what you’d be hiring.". In this case, the request is 700 large glasses of freshly squeezed orange juice at 7am.

Reply on news.ycombinator.com