Vague Requirements

Speaker A: 00:00:00

Hello and welcome to the Cognitive Engineering Podcast produced by me, Fraser McGruer, for Aleph Insights. In this series of podcasts, we take a look at interesting topics and discuss what we think they tell us about analysis and decision-making. I'm here with Nick Hare and Chris Wragg of Aleph Insights, and this week we're discussing bad haircuts. So, Chris, looking at you, I'm guessing you had a bad haircut.

Speaker B: 00:00:26

That's right, yes. Well, I went to my usual hairdressers that I've been to, I don't know, perhaps 30 or 40 times previously, and I went in, and normally I have this Welsh woman who cuts my hair, but, like, there's always this weird thing in hairdressers where you sit there and there's a number of barbers who could do your hair, and it depends where you come out in the order and you feel a bit awkward saying, no, I don't want that person, I want the person who always does my hair. So, anyway, I popped out and there was this young, sort of, hipster chap, was the hairdresser who... And the Jaws theme started playing. Yeah, exactly. And, anyway, so I thought, oh, well, you know, I won't, well, I did prejudge him, but I, you know, I won't let it affect my mood. And so I gave him the usual instruction that I give to my hairdresser, which is to say a grade four at the back and the sides, blend it into the top and take about an inch off the top. And those words have worked, you know, consistently well. Get Danny LaRue over here, bloody hell. 50 years, yeah, well, not 50 years, probably about the same haircut for about 20 years. So, yeah, those were the sets of instructions I gave. And as you can see, the result was, well, it's grown a bit since then. How long ago was this? Yeah, this was sort of two weeks ago, so it's sort of recovered. But anyway, as far as I was concerned, it was a bit of a disaster. What went wrong? Well, this was what got me thinking, what went wrong? And I thought, initially, I thought, well, look, this set of instructions I give, they're reasonably specific, right? There's some quantification in there that grade four is a universal standard, right? So there you go. About an inch off the top and sort of blending it is kind of standard procedure. But little did I know how much room there was for interpretation within that set of requirements. So it got me thinking about things like haircuts and instructions you give to people and how we take for granted mutual understanding that you've built up with a hairdresser over years and how much information is implicit in what we say to one another and how it can result in sort of

Speaker A: 00:03:08

mistakes. Before we continue, just give me some specifics. I mean, I know it's been two weeks,

Speaker B: 00:03:13

your hair looks okay to me. Yeah, well, what went wrong? It was like there's some weird stuff going on around the back. So normally, like I, you know, it's kind of nice and straight at the back. But this was all sort of fluffy feathered at the back. I normally, you know, I didn't mention my side parting because I thought that was sort of, you know, it was there. But it sort of did this kind of weird, like Caesar type effect on the hair coming forward. And anyway, so yeah, there were some things that weren't within the instructions, but neither were they sort of forbidden. They weren't explicitly excluded. They weren't excluded within the instructions. And just one other thing before we continue.

Speaker C: 00:04:02

Fraser's trying to vicariously enjoy the idea of having hair. I am, I am. So he's really going into a lot of detail.

Speaker A: 00:04:07

It's been a long time since I was a hairdresser. So at the end of it, and he did the whole mirror thing at the back, I suppose. Did you bring him up on his errors? Or did you go, Oh, that's great. Thank you very much. Well, of course I did. Of course I didn't.

Speaker B: 00:04:25

I'm English for God's sake. Exactly. No, I thanked him for his time. Did you give him a tip? Do you know what? I did give him less of a tip than I give the Welshman. But he wouldn't know that that was a bad thing. So feedback for him was probably positive.

Speaker C: 00:04:43

rtainly been around since the: 1960

Speaker A: 00:05:36

Well, look, I'm sure we'll circle back to all this at the end. And certainly your hair needs further discussion.

Speaker C: 00:05:41

But my hair, the hairdresser I had at university at Edinburgh once was giving me, clearly, despite me saying I want it to kind of look the same, and I've got a parting, she started removing the parting and turning it into a kind of fringe, like a bowl haircut fringe thing. Very, very strange why she would have assumed I'd have wanted that. But I had to tell her no, and stop, hang on, are you giving me a fringe? Because the problem is you can't really tell halfway through, can you? You can't really tell where it's going.

Speaker A: 00:06:08

And you think, oh, they know what they're doing.

Speaker C: 00:06:10

Yeah. And often they'll be combing it in weird directions.

Speaker B: 00:06:12

One of the key things about a haircut is you can't really see it's gone wrong until it's too late. So I was having a think about the economics of this. Yeah, let's get on to our actual, what we want to talk about.

Speaker C: 00:06:25

Well, I mean, there's a number that I think we want to probably want to talk about this issue of what do you do about vague requirements? But I was trying to think about the kinds of goods which fall into this category, or goods and services. My first instinct was, well, this seems to be an issue about goods versus services. With the good, there it is, you can touch, you can pick it up. With the service, it seems a bit more intangible. And so it's probably harder to measure. But actually, I don't think that is quite the issue. Because there are some goods, some services, like having a car washed or something, where it's just they've either done it properly or they haven't. And it's quite straightforward. So I think the key categories here are, first of all, homogeneity. So is the thing you're pretty much exactly the same as another thing like it. Now, in the case of, let's say, going to a restaurant or something, your meal is going to be pretty similar to someone else's meal. So you can gather information about that from other people. And the meal you're going to get tomorrow is probably pretty similar to the one you're going to get today. And I think in the case of haircuts, the reason why we like continuity in our hairdressers is because then you have homogeneity, because you are getting the same haircut you got last time from the same person. So you've got homogeneity over time. The problem is where you have something which is really one off. And obviously, getting a haircut with a new person, it's a one off. And things like getting a conservatory or something, there's no other house like yours. There's probably no other conservatory like yours. So non-homogeneity, that's a problem. Then we have the issue of durability, which again, is to do with, this is really whether or not you can almost try it out. So do you consume it by using it, essentially? Can you undo the thing you've received? Now, if I want to try out a new hammer, I can go into a shop, hold it in my hand, bash a nail with it. And if I don't like it, I'll give it back to the shop. But in the case of a pair of pants or a Cornish pasty, well, pair of pants, you could. With a Cornish pasty, I can't really try it out and then give it back to the shop if I don't like it. I don't think they'd be talking about the pants either. But anyway, go on. With a haircut, it is clearly non-durable. You cannot undo a haircut. A haircut happens once. And once it's happened, you cannot unconsume that haircut. Can't stick your hair back on again. No, exactly. So already a problem. Then we have the issue of sort of how complex the thing is. Even if it's a one off, even if it's non-durable, well, it still might not be complex, right? It might be something relatively simple, like washing a car or something, or someone cleaning your windows. The only windows in the world which light yours are yours, right? And yet it's a fairly simple matter for someone to come and wash your windows. And so it's not really a problem if the thing is simple and whether it's been done properly is simple. But then we get on to the issue of sort of is it objective or subjective? So in other words, and this is what Chris was saying about the the problem of sort of vague instructions, is can you specify exactly what it is that makes something good? And first of all, it might be that the kind of interaction that the good has with you or with humans in general is just very complex. So for example, it's very hard to say what makes a film good or what makes food tasty. So that complexity issue is another one. And then finally, the final category I think is interesting is does it exist, right? Is the consumption and the creation of the thing simultaneous? So with a haircut, for example, you can't look at the haircut beforehand and then say, oh yes, I'll take that one. Whereas you can to an extent-

Speaker B: 00:10:12

Unless you're in North Korea where you can order them.

Speaker C: 00:10:13

Yeah, exactly, or order them online. So I think there's all of those- Did you come up with those categories? Yeah. Well, no, I mean, a lot of them are sort of standard economic categories, but I was trying to think which of these things apply. And I think a haircut is a good, it is right in the nexus of those things. The only difference, but I think the key thing, I would actually say getting a home extension or something is probably the absolute worst. Because it's high cost. Well, it's also, it's high cost, but it also, yes, I mean, you have this sort of durability, but you're never going to get it done twice. So that's the only difference. With the hairdressers, at least you can go to the same guy if he does a good job with a home extension or something, you really can only get it done once.

Speaker A: 00:11:01

But there was something in there, because when they're doing a home extension, they should be working off a plan, right? And one of your categories, you couldn't have a plan or something, wasn't it? Or- Yeah, that's true. I mean- Objectivity. So this should make it objective. But I have had an extension to my house, and the builders should have been on a TV program. They were that bad.

Speaker C: 00:11:22

Well, Peter had a similar experience.

Speaker A: 00:11:23

Right. That's right. And one of the things that used to really frustrate me was, every day I'd be just going, look at the bloody plan. Just do what the plan says, and you'll be all right. And they just couldn't, or didn't, or wouldn't. I don't know.

Speaker C: 00:11:34

I mean, again, that's a complex, because you're getting not the plan as such, but also the construction of that plan. And there are going to be things in there which aren't minutely specified, you know, the tightness of the bolts.

Speaker B: 00:11:46

And I think that's the bit that I'm sort of particularly interested in, is how you, under these circumstances where you've got these kinds of service product, um, uh, somebody conducting activity on your behalf, how do you articulate, uh, something clearly, especially like in a dynamic situation. So when they're building, they might go, well, you know, the architect didn't realize that you've only got, you know, soil under here rather than concrete or something. So we've got to, we've got to go off piste and, and, and adapt. And, um, you know, one of the, one of the, the key areas where communication of, you know, um, of what you require is essential and a matter of life and death is obviously in the military, um, the military sphere and military orders. And I was thinking actually, you know, giving, giving instruction for a haircut is a little bit like issuing a military order. Uh, and the, the, you know, um, well, I was thinking, because I said, you know, about half an inch off the top. And, and, and, and I thought afterwards I was thinking, well, about half an inch is one of the, or about an inch off the top. That's one of those figurative terms like, oh, it's, it's about a mile or, and then I started thinking of half a league, half a league, uh, half a league onwards, you know, the, the, um, charge of the light brigade and, and I'm half an inch off woods. Yeah, exactly. Not comparing my, my hair to the, to the death and wounding of a few hundred men and horses. But, um, but you know, there was a, there was a great example of that, um, sort of, you know, somebody giving requirements that they thought were specific. I mean, the exact, uh, the exact order was, um, uh, Lord Raglan wishes the cavalry to advance rapidly to the front, follow the enemy, carrying away the guns, uh, troop horse artillery may accompany French cavalry is on your left immediate. So Lord Raglan, the commanding officer of the day probably thought that was, that was quite explicit. The problem came when the captain turning up and delivering hand, delivering the order was asked by the person about to interpret the order and pass it on to, to Lord Cardigan. Um, which guns do you mean? And they were supposed to have gone for the ones on the hill up by the redoubt. And instead they went for the ones that he sort of waved vaguely, you know, and, um, that kind of, uh, um, misinterpretation of something, something potentially specific can obviously have calamitous, um, uh, requirements. And yet you, you can't under these circumstances and for these kinds of products, you can't specify to the Nth degree exactly what you want because you, you need to

Speaker A: 00:14:31

have human interpretation. So, so two things here. First of all, yeah, one of the, one of the things you're saying there way around that, uh, when you're going to get your haircut is to go in there with super detailed instructions, written down pictures, et cetera, et cetera. Yeah. But there's a kind of, uh, um, a tipping point where that starts to become a bit weird if you do that. So you, you, you come away with your perfect haircut every time, but they, you're probably not very popular in that hairdressers and yeah. And it's an over-investment in something that doesn't have huge consequences. And I guess then the other thing sort of, I mean, it seems to me, as you were saying, Nick, you know, the hair, hair, hair cuts are right at the nexus of all this, but what else is similarly complicated? Because I know we were talking about, um, um, military operations there, but that was one specific example of, of, of where they got wrong. Famously stuff goes wrong in, in military, uh, operations all the time. What else is there? Um, apart from haircuts and military disasters, what else? Well, I think, I think the, the, in

Speaker C: 00:15:26

terms of everyday life, you know, is, I know we've done a podcast about dodgy, dodgy builders and stuff, but I think that's where the biggest problems, uh, always seem to arise is with people doing something in a sort of fairly permanent way to your home or your body. Now I'm also, and I'm also a subscriber to, um. You've had an extension to your body. No, I, well, I not unless you count my glasses, but I, I, uh, have, I'm a subscriber to, um, a, uh, uh, uh, uh, uh, read the subreddit called awful taste, but great execution. And, uh, that's, uh, there's a various other ones, a bit like that about, um, about, and, and a lot, a very common category is, is bad tattoos. And, uh, yeah, I, in fact, I used to work with someone who, um, who had got in their twenties, a tattoo of a, of a, of a dolphin doing a thumbs up or a massive, like, and as a result, she, she had to, she had, she always wore long sleeved clothes, but then had to, had to go through quite a lengthy removal procedure. I think it was, she'd managed to turn it into something quite tasteful. Um, uh, but I, I mean, so I would guess that that's, you know, tattoos is another one that people commonly, uh, come across this problem, but in obviously in a much more permanent way, you know, we've all heard the stories about people who get, uh, you know, think they're getting a, uh, the Japanese word for wisdom and it turns out to mean toilets or something, you know? And, uh, yeah. So I, so I think there's, I think it's not uncommon for us to encounter this problem, but I, but I think what I was going to say is that in, certainly in our business, you know, sort of when we're, we're providing, um, effectively analytical services quite a lot of the time to, uh, customers like, you know, like the ministry of defense, um, we encounter this problem of, you know, particularly where, um, the customer has an idea about what their solution might be and, and, you know, that what they think the solution might be might not be actually what we think the solution ought to be often. Uh, and, um, you know, when, and so they'll specify a requirement for something that they want, but it's very hard to do that without pushing the supplier to, to, to some kind of solution, you know, they've got an idea what it will look like. And then, and then of course, often you don't even know how many, how much, how difficult something is going to be to provide or, you know, what the pitfalls are going to be and those kinds of projects. And I guess, you know, thinking of any kind of large procurement project in government, procurement projects always go wrong. And often that's, that's one of the drivers, you know, is actually what you, what you think you need isn't, isn't, it's hard to specify because actually what you need is kind of an effect. So you end up specifying a system or, or a solution. And that often isn't actually going to deliver the thing you wanted. So yeah, it's something we do encounter quite, quite a lot. Yeah. Yeah. Sorry. Go on Chris. Well, I was just going to

Speaker B: 00:18:13

say that, you know, the way we try to mitigate the, the problem of turning up with something that is, is not what the customer requires is, is through two things, I think, you know, one is a regular process of feedback so that you don't get too far along, you know, so, so, you know, when you're cutting the hair, that would be sort of, is that right for the fringe? Yes. Okay. Now I'll move on to the back bitch, you know, so that you don't turn up with something, you know, so, so that's what we, you know, we try and encounter sort of break points where we can readjust and get their input and so on. And the other one really is you know, because, because they it's not right to fully specify something because you want to often, you know, you, you don't know exactly what it is that you want, you know, something like commissions to do with design or something like that, you know, where you want an interior designer to come in, you don't know what you want, that's why you've got an interior designer, but you've got some, some sense of, of what it is. And there, I think it's about, you know, properly giving license to the, to the person you've commissioned, but describing the, the kind of objective function properly, you know, what, what, what is the, the thing that they're trying to maximise, you know, and, you know, mission command in, in, in the military is a way of doing that way where the commander doesn't tell everybody exactly what to do, but they do say, you know, like Nelson used this in Trafalgar, you know, the aim was to kill as many French and Spanish sailors as possible, but how that happened was left down to the, the, you know, interpretation of the, of the people concerned.

Speaker A: 00:20:02

I mean, this happens to me quite a lot in my daily life, in my sort of normal day-to-day work, which is essentially making films for people, because I'm the, the, I'm the tool by which I'm the tool by which a film is made. And typically what happens is most of the time I'm working with people who don't really understand the filmmaking process and they just know they want a film and they kind of, they've got a vague idea of what they want. And then when they start sort of saying, yeah, we want this and we want that, it starts to become apparent to me that, okay, well, that's just not going to work. And what I find, you know, I wouldn't like to think I'm the equivalent of a, of a bad hairdresser in filmmaking. And, and, and so what's key is all, is just going along with people from all the way along. And at the very beginning, be quite clear, well, look, if you want that, that can't happen for various reasons. And then all the way along and almost as, as much as possible, lock it down. But then one of the things that happens is that you can lose spontaneity. Anyway, this is what happens to me a lot. Yeah. And there is also

Speaker C: 00:21:07

that fear of trying to tell people how to do their jobs, which I think is partly an issue, possibly worse for British people. I don't know. But I, I think, you know, why don't we march into hairdressers with a photo of what we had last time and say, this is what I want. And then, you know, every two minutes say, right, tell me what you're doing and you know, how long is it? And you know, certainly when, when things have gone wrong with sort of home improvements, often it's, you know, you have this kind of nagging sense of, well, that doesn't look quite right, but I'm sure they know what they're doing. And then they, they, they don't, you know. So I think that, but how do we, yeah, I mean, obviously that's the problem. That's part of the problem. So getting good people is, is, is a set. Getting good people and people who are willing to have you question what they're doing, you know, I mean, are able to explain what they're doing and, and to explain why that relates to the thing you want. Yeah. Is there anywhere, there's something I want to finish

Speaker A: 00:21:59

off on, but is there any anywhere else we want to go with this for a moment? Anything you need to

Speaker B: 00:22:03

round this off with? Well, I mean, the only thing I was going to sort of talk about a bit was, you know, the main, usually the main mechanism, other than for things like a building where you might have some blueprints, which have got, you know, clear quantified measurements and materials specified and so on. But most of the time we're commissioning these things, we do it through words. We do it through natural language. And I, you know, while that is brilliant, right, because you can sort of say things like just blend it and that carries loads of information, you know, natural language is very evocative. It also means that there's, there's range for people to misinterpret what, what you, what you say. And so I think, you know, we don't want to go down the route of having in a military situation, they do have controlled language for things. They give grid references for places and they have specific unit names and so on. But generally we don't want to use that kind of controlled language. We want to maintain the flexibility of, of being able to talk to people naturally, but give them the space to interpret it. But it's how do you make sure that interpretation is good and not bad? Yeah. And I think the, you know,

Speaker C: 00:23:21

the enemy of all of this, of course, is Goodhart's Law, which is the idea that, you know, as soon as you make something a target, it ceases to be a good target. A classic example being something like, well, we know, we kind of have an idea what we want for schools to be able to deliver, but it's quite a complex thing. But, you know, we know that one of them is good exam results. So you, you, you know, you make exam results a target and you say, that's how we're going to measure your quality. And then of course, you know, what will happen is schools will teach kids to be good at doing those exams. And that's not necessarily anymore going to indicate that they're learning what you actually want them to learn, which is something that's hard to measure. So I think it is a, it's a, it's almost a universal, it's almost a, you know, a kind of inherent problem with certain kinds of things. You know, that it's just the, the output is complex. The way it interacts with you is complex and being able to specify what makes it good or

Speaker A: 00:24:11

bad is just inherently hard. So, which begs the question, Chris, what are you going to do the next time you go back to that hairdresser's? If indeed you go back to that hairdresser's, will you pass that person if they come up to you in the queue? I will, I'll probably,

Speaker B: 00:24:25

what I'll probably do is give it the one's pass, check the hipster's not there and then go in

Speaker C: 00:24:31

for my haircut. But can't you say, I need, I want, you know. Any of you can cut my haircut, that hipster right there. Yeah, or just ask for the Welsh lady, Mifanwy or whatever she's called. Yeah. I mean, I say I'd like a haircut with Ray Ruff. There's no one else can do it, you know.

Speaker A: 00:24:48

So yeah, do that. They should have that on a placard somewhere in their barbers. Yeah. I want a haircut with Ray Ruff. Yeah. No one else can do it. Okay. Look, so to finish things off, let's, you seem like quite sober, sensible people, but maybe you've had moments of creative waywardness. Have you ever chosen to have a bad haircut or what turned out to be a haircut, Chris?

Speaker B: 00:25:10

eenager in the, in the, about: 1990

Speaker A: 00:26:19

Yeah, but obviously in a magician's trick. Yeah. What about you, Nick? Have you chosen

Speaker C: 00:26:24

any tragic hairstyles? No, not really. I mean, my hair has barely changed. I mean, when I was at university, I grew it out. I had long hair, ponytail length hair. I looked a bit like an Irish footballer when I was about 19 from, from the seventies, you know. But I, I'll defend that to the death. I, I, you know, it's silly now. Youthful misadventure. Long hair was very common in the early nineties. Now we look back and it looks like a comeback. Yeah. You look, you know, you kind of think, well, it was one of those ridiculous things. Nobody has long hair now.

Speaker A: 00:26:56

Everyone has beards and we'll laugh at that in 20 years time, but you know, when, when long hair's back. But yeah. It was too late for me. It already is. I was just talking about mullets. Yeah. So in the early eighties, I decided, I was quite a fan of cricket at the time. And Ian Botham was a big thing. And I think he had a perm on top kind of, and then a mullet going on. And I thought, yeah, that looks good. And so I went for a month or two. I had that, you know, I went for a mullet. And then again, when I was about 13, do you remember, did you have this thing? You're a bit older than you. The spike. Do you remember the spike haircut? I remember people having a hair spiked and then there was the flat top as well. Yeah. So I thought that would look great. And so I, that's what I had. I had a spike done. So it was just, it involves lots of gel, but cause I've got quite fine hair. All that happened was my hair just sort of clumped together in like about seven different spikes. Like a hedgehog. Like a hedgehog. Yeah. And much the amusement of my family. So those were my two. I'm imagining you looking a bit like Keith Flint now. Who's Keith Flint? You know, from the prodigy. Yeah, no, it wasn't exactly. It was a lot like that. A lot like that. Yeah. Yeah. Fortunately, that's all a thing of the past. And now, as you can see, I've got these lustrous locks and I've gone for the long hair look. So yeah. Okay. Well, I think that's some good advice out there for anyone going out to get their haircuts. Detailed instructions and, and take a photo. Take a photo. There you go. Thank you as always for listening to the Cognitive Engineering Podcast. I'm Fraser McGruer. I've been here with Nick Hare and Chris Wragg of Aleph Insights. Until next time. Goodbye. Transcribed by https://otter.ai