Learning Library

← Back to Library

The Costliest Mistake with Engineering Teams

Key Points

  • The common belief that shipping an imperfect product is the most wasteful use of engineering time is a myth; delivering a deliberately imperfect MVP can be strategically valuable.
  • Conversely, the idea that over‑polishing a product is always the biggest mistake is also a myth; some contexts demand high quality and safety before release.
  • Effective product planning should treat engineering effort as an “elastic time constraint,” matching the amount of work to the specific risk‑reward profile of the initiative.
  • High‑stakes domains—such as drone delivery—illustrate cases where absolute reliability is non‑negotiable, requiring engineers to solve critical problems fully before launch.

Full Transcript

# The Costliest Mistake with Engineering Teams **Source:** [https://www.youtube.com/watch?v=CEDXz_M-5BE](https://www.youtube.com/watch?v=CEDXz_M-5BE) **Duration:** 00:11:07 ## Summary - The common belief that shipping an imperfect product is the most wasteful use of engineering time is a myth; delivering a deliberately imperfect MVP can be strategically valuable. - Conversely, the idea that over‑polishing a product is always the biggest mistake is also a myth; some contexts demand high quality and safety before release. - Effective product planning should treat engineering effort as an “elastic time constraint,” matching the amount of work to the specific risk‑reward profile of the initiative. - High‑stakes domains—such as drone delivery—illustrate cases where absolute reliability is non‑negotiable, requiring engineers to solve critical problems fully before launch. ## Sections - [00:00:00](https://www.youtube.com/watch?v=CEDXz_M-5BE&t=0s) **Debunking the Imperfect Release Myth** - The speaker argues that shipping an imperfect product isn’t the most expensive mistake for engineering teams, emphasizing flexible time constraints and proper MVP trade‑offs instead. ## Full Transcript
0:00what is the number one most costly thing 0:03that you can do with your engineers if 0:05you're in product management or frankly 0:07if you're an engineering leader what's 0:08the most costly thing that you can do 0:10with your engineering team we're going 0:11to talk about that today and we're going 0:13to dispel some common myths that people 0:16spread pretty widely around Tech around 0:18what is the most expensive uh or costly 0:20thing that you can do that wastes 0:22engineering team time and 0:24talent number one this is a we're going 0:26through the myths first we'll get to the 0:28thing that that is the actual cost C at 0:30the end so the first 0:32myth it's it's incorrect to say that 0:36building and releasing an imperfect 0:39version of your product is the most 0:41costly mistake that you can make with 0:43your engineering team that's just not 0:45true and part of the reason why that's 0:47not true is that oftentimes you want to 0:49be deliberately imperfect so that you 0:51can trade off and get to Market faster 0:54so there's that whole conversation 0:55around what a minimum viable product 0:57looks like and how minimum is to minimum 1:00I'm not going to get into that whole 1:01discussion today but I did want to call 1:02out if you're thinking about the highest 1:04and best use for your engineering team 1:08you want to be thinking about products 1:11in terms of an elastic time constraint 1:13so that you can say this is the correct 1:15amount of time to spend on this product 1:17for the risk and the reward that we're 1:19looking at and that can include shipping 1:21imperfect product that's not necessarily 1:23a waste of your engineering teams time 1:24and if you hear otherwise that's a myth 1:26it's myth number one myth number two is 1:29the opposite it's it's a two perfect 1:30version and I do hear people who love 1:32the idea of an MVP saying that the worst 1:35thing you can do with your engineering 1:37team is to spend too much time on the 1:40product and to make it too perfect and 1:41too polished I don't think that's always 1:43true I think it's sometimes true just 1:45like a lot of things in product it 1:46depends on the context for example if 1:49you're trying to solve a problem that 1:51needs to be all the way solved before 1:52it's solved at all you're going to need 1:54to polish the edges uh a great example 1:57of that is the iPhone was released in a 2:00fairly polished State even in version 2:02one and I know we look back on that now 2:04from the perspective of like our iPhone 2:0615 over here or whatever it is and we 2:09say I cannot even imagine being in 2007 2:12and calling that phone polished but at 2:15the time it was polished I'll give you a 2:17different example that's not Apple if 2:19you are trying to solve the problem of 2:22uh drone delivery I heard a good talk on 2:24this actually when I was at Amazon you 2:26need to be sure that your drones will 2:28never ever ever fall out of the Sky 2:30because the cost of that would be 2:32killing a customer and that's way too 2:34expensive you can't ever afford that 2:35that's never 2:37acceptable and so you have to solve the 2:39problem of safe flight and safe package 2:42delivery completely before you release 2:44it all and so it would be incorrect in 2:46that context to say you're wasting your 2:49energy by spending too much time 2:50perfecting the product because you're 2:52not wasting it because that's the bar of 2:53the product demands I'm going to make a 2:56second point though here that is rarely 2:58made and that I want to stand behind it 3:00is also true that the most wasteful 3:02thing you can do with your engineering 3:03team is not is not to over perfect a 3:07product that you could have released 3:08sooner that's still not a great thing to 3:11do so if you are in a world where you 3:12can release a scrappy minimum viable 3:14product and those things do exist not 3:16everything is high stakes like drone 3:18delivery great maybe you're releasing a 3:21sort of a small proof of concept piece 3:22of software it maybe you spent too much 3:24time on it with your engineers that's 3:26not a great choice it's still not your 3:28worst choice and we're here today 3:30looking for your worst Choice as an 3:31engineering manager or engineering 3:32leader because I think understanding 3:34that helps you understand where you can 3:35put your time most and this goes for 3:37product okay so if you if you spend your 3:41time on creating imperfect versions of 3:44things that's not the worst thing you 3:46can do if you spend your time creating 3:47two perfect versions of things that's 3:49myth number two that's still not the 3:51worst thing you can do and the reason 3:53why is neither of those things play with 3:56the core risk factor involved those are 3:58all playing with engineering time as the 3:59key risk factor and that is a false 4:02Assumption of risk your biggest expense 4:05here is not the cost of your engineers 4:08and people assume it is and so when they 4:09talk about you built it to imperfectly 4:12now the engineers will have to go back 4:13and do it again it still translates to 4:15engineering time as the basic unit of 4:16economics if it's too perfect it goes 4:19back to engineering time as the basic 4:20unit unit of 4:22Economics it's based on a full flaw 4:24assumption because at the end of the day 4:27engineering cost is not the biggest risk 4:29risk factor for your 4:31business and I'm going to give you one 4:33more example on the engineering cost 4:34fallacy so the third one that people 4:37often do is they say Engineers need to 4:40maximize the time they spend on keyboard 4:43and not planning and thinking right 4:45there's this assumption if the engineers 4:47aren't producing code they're not doing 4:49work and anyone who's worked in 4:51engineering will tell you that that 4:53doesn't make any sense at all but 4:55Business Leaders sometimes make that 4:56assumption and sadly product managers 4:59will sometimes along with that is 5:01absolutely not a mistake a problem 5:04something that needs to be corrected 5:05engineering thinking and planning before 5:08they go into coding is actually a great 5:10choice and so this is one of those 5:12things where there's not really in my 5:13mind much of a debate you want to make 5:14sure you have time for your engineers to 5:16think and plan carefully but people do 5:19because they think of engineering cost 5:21as time it takes to produce lines of 5:24code assume that you have to have your 5:26engineer sort of maximizing the number 5:28of lines of code out and any engineer 5:31will tell you it's actually not the 5:32number of lines that you want to 5:34maximize it's the efficiency and 5:35elegance of the system and the way it's 5:36able to solve problem and so we have 5:39this factory production mindset that has 5:41crept in from frankly the 19th century 5:43and the early 20th century when we did 5:45assembly lines where we assume that 5:48Engineers are workers on assembly lines 5:50mentally and therefore they need to be 5:51working to make code all the time and 5:53the core cost that we're dealing with is 5:55like how quickly can they turn out the 5:57code and is the code the right quality 5:59and that factory mindset underlies all 6:03three of these that's what we talk about 6:05and think about when we say this time 6:08unit cost of engineering as our biggest 6:10risk factor I want you to take that 6:12whole mental model and just put it to 6:13the side it's not super helpful in 6:16actually understanding the real cost 6:17because at the end of the day that is 6:19the wrong mental model to understand 6:20software software is not a factory you 6:23do not make software on an assembly line 6:25like that instead software is a series 6:29of bets that you're playing on how you 6:31can assess product Market fit using 6:33software and if you get the BET right 6:35software gives you tremendous leverage 6:38to unlock that value farther and that's 6:42why software isn't like factories 6:44because it's not like you're making five 6:45different kinds of cars and you can make 6:48a thousand of them cheaply as soon as 6:50you get one to work you're actually 6:52committing to a production run and 6:54you're making them linearly when you're 6:55dealing with physical goods and physical 6:57products I've had to do that 7:00but with software it's actually really 7:02elegant and really easy to make these uh 7:04thesis or bets and see if they fit and 7:08that brings me to the part that is the 7:10biggest cost what is the biggest cost 7:12for your engineering team in a world 7:13where you actually understand how 7:14software is made in the unit economics 7:16of software it's aiming incorrectly it 7:19is producing bad thesis or bad bets that 7:22you will then follow up and have to 7:25spend time on when you shouldn't be 7:27spending time on the BET in the first 7:28place and that is why product is such a 7:31high leverage higher because at the end 7:32of the day a product manager's job is 7:34not to make bad bets they need to make 7:36bets that are as high quality as 7:38possible that doesn't mean you expect 7:40every single one of them to work but it 7:43does mean that you expect them to be in 7:47a direction of a learning Loop really 7:49deliberate as high quality as possible 7:52grounded in customer as much as 7:54possible and that you expect the product 7:56manager to be able to learn and adapt 7:58quickly and so when you're hiring for 8:00PMS that's what you're looking for and 8:01when you're looking at the cost to your 8:03engineering team you're looking to 8:04basically say is the engineering team 8:06spending their time on the right 8:09bets and so in a sense the value of the 8:12engineering team is contingent on the 8:13value of product Direction and whether 8:15or not you get your product direction 8:17from a product manager or from an 8:18engineering leader or from your CEO 8:21that's still the fundamental question 8:25are you putting your engineering team 8:27against the right bets because each bet 8:29you start is like a little thread across 8:31an unknown landscape and if you start 8:34that thread just the motion of momentum 8:37tends to carry you forward on that 8:39thread and you will put inordinate 8:41effort into trying to prove out that 8:43thread going forward it will take you 8:45longer than you expect to stop longer 8:47than you expect to release fast follow 8:49features that's just the way software 8:51works so you better get where you want 8:53to start right and that's the key risk 8:57if you are bad at pick in where you want 9:00to start at exploring what I would call 9:02a latent space for product where you're 9:04trying to figure out whether the product 9:06fits or not if you pick incorrectly if 9:08you frame the problem wrong if you pick 9:11the choice of features for your MVP set 9:13so you're not actually learning 9:15something that's going to be a problem 9:18that's a bet that you can't take back 9:20very easily that's something that's 9:22going to be very expensive because your 9:24entire effort from your engineering team 9:26would be poorly poorly in tested and 9:30poorly chosen from day one like it 9:32doesn't even matter if it kind of works 9:34if you frame the problem badly enough it 9:36won't be 9:37scalable in a sense it is possible to 9:41get your engineering team going in a 9:42direction where it will take months to 9:44unwind whatever you do you will learn 9:46nothing and your bet will be wrong from 9:48the GetGo and I have seen that happen 9:51and that that is the actual risk of how 9:54you waste time with engineers and it's 9:55not just product managers that do that 9:56I've seen business leadership do that 9:58sometimes I've seen enging leaders do it 10:01and if you understand how software Works 10:03getting your engineers focused on the 10:05wrong bets is the number one biggest 10:07risk you can you can face so make sure 10:10that you are betting correctly and that 10:11is probably a separate video we'll do a 10:13separate conversation on how you aim or 10:15how you set up your bets correctly but I 10:17wanted to call out those myths because a 10:18lot of people think oh it's about over 10:20perfecting your product that's the most 10:21expensive thing you can do with 10:22Engineers no it's or it's about under 10:25perfecting the product you're releasing 10:26a flawed product that's the most 10:27expensive thing you can do because 10:28you'll have to to rework again no unit 10:31cost is not the problem and unit cost is 10:33also not the problem when your engineers 10:35take time to think which is the third 10:36major myth none of those are really the 10:39issue because factory models of 10:41production for software are a flawed 10:43mental model software is about bets and 10:47the core value that you have is making 10:49the right bet and it follows that the 10:51core risk that you take is making the 10:53wrong bet that is the most expensive 10:55thing you can do with your engineering 10:56team and I recommend not doing it let me 10:59know if you have any uh bad bats or good 11:01bats that you want to chat about in the 11:03comments uh cheers