Development Is Cheap. Production Is Not.

This is a line from a blog post written by my brother in law Jerry Solomon. He is talking about film production, specifically short form videos. But the point is true of all projects, from designing a building, a home, a film, an art project, a hardware project, a software project, or whatever.

The design process is relatively inexpensive. The build process is not.

In the kinds of companies we invest in the “development” work comes from the product organization. The “production” work comes from the engineering team.

I have seen engineering teams spin their wheels and burn through countless hours of writing code that ends up getting tossed out just because the design process was not right or not specific enough or not thought through enough.

While we might think these issues and challenges are unique to the world of tech, software, internet, and mobile, the truth is these issues pervade everywhere you are making something.

This is not so applicable to a startup trying to find product market fit. But it becomes very relevant once your company starts to scale. A commitment to thinking things through, getting it right at the start, and being efficient in the “production” process is something all great companies figure out how to do. It’s really important.


Comments (Archived):

  1. Barry Nolan

    True – but only few new products survive their first engagement with production. And when Founders and VCs both worship at the trough of traction, there’s often no line between between MVP and the servers are burning.

  2. awaldstein

    I’m contrary this morning Fred.Go to market from a pure resource perspective is invariably the most expensive thing in my experience as it continues forever. Whether it be 15% or less of sales, over a brand’s lifecycle, it has to top initial investments.I am amazed of just the opposite–of how truly great products with great plans and an understanding of their market get the plan wrong.Capitalizing CitiBike here at home.Capitalizing any food product where duh–no matter how smart you are and crack open the bottle of Prosecco when you get the big deal, you realize the massive capital drain to deliver on it on the other end.You get my drift.To not have a plan is insane. To believe that it will hold is even moreso.I probably need more coffee.

    1. Avi Deitcher

      Contrary is good, makes life more interesting.I think he was including “go to market” as part of production.

      1. awaldstein

        You may be right but on a budget Marketing & Sales never sit under production as a line item.

        1. Avi Deitcher

          True that. But it is about producing, isn’t it?

          1. awaldstein

            On one level sure.But if it doesn’t fall under the dollar line item and in almost every instance, doesn’t fall under the same leadership exec, and invariably, the skills don’t rest in one person–its related but different.Apples and oranges are fruit but–you know where I’m going.

          2. Avi Deitcher

            I does, indeed. I can see the CEO jumping up and down saying, “this is all under production and has to work together,” but what you pay for and what you budget and who owns it speak more loudly…

          3. awaldstein

            btw–realized that i’m having this conversation with you and don’t have any context to who you are or what you do so been over at Atomic Energy.Is that the best link to get to know you a bit better?

          4. Avi Deitcher

            I meet the best people via this forum.Yes, the link from there to my blog is the best path. Send me a lInkedin as well, and we can share emails that way.

      2. Anne Libby

        And I was lumping all of the rest of “execution” under production…

    2. Brandon Burns

      Fred’s so off on this. Like, so so off. I don’t even know what to say.

      1. fredwilson

        See my reply to Arnold

        1. Brandon Burns

          I’m not sure the stage of a business has any effect on the core design-build process of making a product. I think that the tech industry just gets this wrong time and time again.And it goes all the way up to the Facebooks and Googles of the world, who hire agencies and design shops to design products for them (I’ve worked on products for both from within various agencies), and then the tech companies make the design shops, who save their asses, sign secrecy contracts so no one will find out that’s what really goes down. But these companies are only hiding the fact that the way the tech industry thinks about and builds products is fundamentally broken, and refusing to learn from the same repeated mistakes and do better.And most of it links back to design, and the misunderstanding of what it is and how to incorporate it.

          1. LE

            And it goes all the way up to the Facebooks and Googles of the world, who hire agencies and design shops to design products for themIf I understand what you are saying then I agree. The reason being that if you don’t have a seat of the pants feel for something you aren’t in a position to even evaluate the correct resource or manage that resource to accomplish that goal.All this flows from the top of a company and infests everyone they hire.That’s one of the reasons Microsoft’s marketing presence (down to the colors that they choose) suck so much and why Google thinks “Android” is a good name for an operating system. Also why the majority of both of those companies products (adding Amazon AWS into the mix) totally fail the puny brain test.My guess is that Andy Grove’s brain and presence played a role in Intel going with “Intel Inside” to great success.

          2. Brandon Burns

            “if you don’t have a seat of the pants feel for something you aren’t in a position to even evaluate the correct resource”Exactly. And also why we need more design voices leading the way to help people better understand how to incorporate design thinking and planning into their team structures.

          3. LE

            we need more design voicesUnfortunately (and strictly from my observation) most “artists” are typically not very good at promoting themselves or their point of view in the best way possible.Funny I was watching CNBC “The Profit” with Marcus Lemonis the other night. (I’ve done business with him). He was thinking of investing in a small sign company (he never did in the end). I marveled at how he completely missed the point of the fact that they problem in a company like that is not getting sales it’s having good people who actually show up on time, are creative and get the job done. With no bullshit. The sales part is easy. [1] I know this because I was in a similar business and also a bit in that industry as well. He (as a result of never being in that business) did not know that like with a restaurant the customers are there if the product is good. Now try to get a good chef and good servers and not have, what I used to call (this is true btw) “wheel of fuckups!”. That’s the hard part.[1] Part of the reason ironically is that over time when you are starting out the easiest new customer to get is the one whose job just got fucked up by a competitor. (The second easiest is someone who didn’t pay your competitor..)

      2. LE

        Fred’s so off on this. Like, so so off. I don’t even know what to say.If someone is wrong then point out why you think that they are wrong.

    3. fredwilson

      I should have been more clear (and rewrote a bit of it just now). This applies once you have product market fit and are scaling. It’s not really applicable to the raw startup

      1. rikardlinde

        Hm, I’ll have to re-read. Thank you.

  3. Avi Deitcher

    I would agree very strongly, **iff** you include production as everything from post-design forward. Yes, engineering teams spin their wheels because of poorly defined product/market needs (and even more probably because the product team didn’t clearly articulate the “why” so engineers can make intelligent decisions in process). But delivery to the customer, support, shipment (even in mobile – 30% of revenue to Apple?), you name it, they are all very expensive and slow.On the product side, to my mind, the top 2 things a good product manager/head/VP needs to have are: solid product/market knowledge; firm understanding that no one knows what the hell is going on inside is head. It is kind of like marriage…. 🙂

  4. Seth Godin

    The thing is, development takes guts, the guts to say, “here, I designed this, here, I think this is the right answer, here, let’s spend the time and money to build this.”And that’s an easy fear to avoid, so many organizations just punt the hard decisions and instead say, “I’ll know it when I see it,” conflating production and development.The most efficient movies are storyboarded before the camera rolls.This post belongs on the wall of every conference room.

    1. awaldstein

      Greatness comes from both sides–efficient and not.Woody Allen or Clint Eastwood efficiency sure–perfect on storyboards– but that doesn’t make for an Apocalypse Now.

      1. Vivek Kumar

        Agreed, I have always thought of Apple and Google being on the opposite end of spectrum and being very successful.* Apple: understand customer, build perfection.* Google: Make production really cheap, build lots of stuff, measure and adapt on customer feedback

        1. awaldstein

          BingoApple is a, no the, consumer electronics company of my lifetime.Google is a tech company.Both great, very different.

        2. Susan Rubinsky

          Apple is what I call a “Visionary-led” company as opposed to a “Data-led” company like Google. The only reason Apple exists is because they had Steve. Otherwise they would be dead. It remains to be seen if Apple can successfully transition from a “Visionary-led” company to some other format.

      2. jason wright

        didn’t Coppola lose his vinyard over that film?

        1. awaldstein

          to my knowledge which is about zero the film predated his purchase.regardless that is not germane in my mind to whether efficiency is a prerequisate or even possible for greatness to occur.

          1. jason wright

            i don’t disagree.four years hacking around the philippines was not efficient. the outcome was spectacular. it isn’t a standard model.he’s a maverick, a contrarian. they need patient investors.

          2. awaldstein

            I only worked in the film industry for a short while.And with all levels from tech to the studios to distribution from newcomers to James Cameron.Nothing involved in getting funding is efficient. These are bets not investments.I stand in awe of people who can actually drive forth films on astronomical self belief when everything around them says no and no minute after minute from the first meeting to the release.

      3. laurie kalmanson

        the tracks are laid down before the cameras roll

    2. Brandon Burns

      Storyboards represent the design process, not the build process.Design = the blueprints. Production = the build. Skimp on the blueprints and watch your house fall down.You seem just as confused as Fred.

      1. ShanaC

        web products do both together.

        1. laurie kalmanson

          exactly; especially now that web products more and more are software

    3. LE

      The most efficient movies are storyboarded before the camera rolls.You mean “creative” not “efficient”, right? If movies that matter are typically blockbusters does it really matter if they are done efficiently? The only argument I could give for efficiency is that it allows you to iterate more movies and thereby have a potential hit.When I worked for two companies in SV many years ago I marveled at how they pissed away money. But the truth is the companies didn’t fail because they pissed away money they failed because of other reasons (Apple changing their strategy being one) that had little to do with whether they were efficient, thrifty, whatever. If the low hanging fruit continued they would be fine with how inefficient they were (this was not a low margin business or a fast food restaurant..)

    4. laurie kalmanson

      Smartest startup I worked at after Seth’s shop was run by a former film director; he knew from long experience that $ spent on design/storyboarding/planning = $ X 10 saved on shooting; the digital shop he runs today is one of the best.

    1. Kasi Viswanathan Agilandam

      Yes….I was also thinking about Edison when i read the article.He invented 998 ways of how not to make a light bulb.

      1. jason wright

        i like the word ‘nimble’. you know it?

        1. Kasi Viswanathan Agilandam


  5. Kasi Viswanathan Agilandam

    Prototyping Is Cheap. Product Making Is Not…

  6. Vivek Kumar

    Not only pervade everywhere, but even get amplified – the cost of production in tech is typically is typically lower than other industries.That being said product organization typically underestimate the production cost including distribution/sales and over-rely on product feedback instead of due diligence and hard choices.

  7. Brandon Burns

    What??!!!!”The design process is relatively inexpensive. The build process is not.”Oh? Really?”I have seen engineering teams spin their wheels and burn through countless hours of writing code that ends up getting tossed out just because the design process was not right or not specific enough or not thought through enough.”I’ve seen this too. This is what happens when you don’t invest in design.”A commitment to thinking things through, getting it right at the start, and being efficient in the “production” process is something all great companies figure out how to do. It’s really important.”That is the very definition of design – its everything that leads up to production. Design = the blueprints. Production = the build. Skimp on the blueprints and watch your house fall down.I just simply can’t understand why the vast majority of folks in the tech industry still don’t get this.

    1. Bruce Warila

      Great response. After 5 or 6 fifteen degree pivots (“spinning through countless hours of coding”), the engineering team is burnt out on trying to deliver a product based upon designs that always seem to be half baked. There’s SO much that you can find out from customers without writing a single line of code.

      1. JamesHRH

        I have seen fried engineering teams too.No one can get revved to 8000RPM and then pop the clutch, go 100′ & hit a brick wall without losing faith in the marketing leads.

    2. Yinka!

      1 reason the design process is seen as relatively inexpensive (vs the build process) is because the former is viewed as a thought only process that takes little time. Understanding the problem properly, presence of and ability to tap into relevant experience and connect it with technical skills, tools used, practical experiments (e.g. field analysis, prototyping, customer dev, etc) run during the design process, etc are not recognized or accounted for.Ironically, this may be exacerbated in contexts where the designers are good: if they work efficiently (understand the problem and frame solution properly in less time), then the build team gets better blueprints, which leads to a more productive build. But a surface glance could make some see it as design being a minor factor.Of course design doesn’t end when build starts; it is ideally an inclusive, iterative process based on feedback that flows through all touchpoints including the market.

      1. JamesHRH

        Low headcount and lack of asset requirements makes design less capital intensive.

        1. Yinka!

          That depends on context; same could be argued for production in a scenario where the product is non-physical and “builder” headcount is low. And either design or production would be less capital intensive relative to the entirety of company operations (design, production, marketing, financial, etc).That said, in an environment focused only on scaling, production would include increased marketing operations, etc., which would make it more capital intensive.

    3. falicon

      I think you’re being a bit too emotional about the use of the word ‘design’ here…I *think* the point was that execution is where the costs start to scale…it’s important to make sure you are executing on the right thing…and so if you commit a bit more resources to those decisions before you commit and start to execute on them, it will be well worth it (and is still cheaper in the long run than executing on the wrong stuff).So I actually took the overall point to mean, spend more time on design/planning and fact-checking your reality throughout the whole process…which I *think* is what you are arguing for as well…no?

      1. Brandon Burns

        You’re probably right that I’m being emotional. I do that with design stuff.And, yes, Fred’s main point of “plan before doing” is a great point.But you can only plan well if you set things up correctly. And the system of how people work with each other in most tech companies is broken. And as long as that remains true, the planning part will continue to be inefficient.

        1. LE

          You’re probably right that I’m being emotional. I do that with design stuff.Exactly. Creativity is driven by emotion. That’s why you can’t just retreat somewhere and be creative. You have to be inspired by something and it just happens. And it can’t be assigned. Imagine if someone said “your job today is to make 5 comments on”.

    4. Susan Rubinsky

      I totally agree, Brandon. And Yinka and Bruce have points as well. Yinka hit the nail on the head when she mentioned that it’s a collaborative/iterative process. I love your original analogy to the construction industry because in that industry, the architects (the Designers) lead and the engineers (the Producers) follow. However, most major construction projects are pitched by a joint “Design-Build” team; these teams are typically separate companies altogether, yet they self-select complimentary partnerships with each other for productivity and scale. Ideally, that’s how tech should work too but, unfortunately, it’s often not so. In my experience, design is often considered an afterthought in tech.

      1. Brandon Burns

        The world of architecture and construction is an old one. Much older than tech. They’ve learned things over the years that hopefully, eventually, the tech industry will learn, too.

        1. karen_e

          Some things:1. Architects are under a lot of pressure to let construction lead the way in many sectors (clients types, building types). This makes architects convulse, panic, and/or faint.2. The profession of architecture is also under pressure from high tech to automate design. 3. Computer-aided design has created a huge swath of people who don’t know much about building actual buildings perhaps because they don’t get the chance very much to step out from behind the monitor.4. Even older than the divided professions of architect and contractor is the idea of the master builder. 5. Design-build, iteration, yes.

    5. PhilipSugar

      You are cranky today :-)This post resonated with me.To me what it meant was don’t cheap out, skimp, or not care about design because production costs will kill you.This is so true.It doesn’t mean don’t care about design. It means, just the opposite. Not spending money on design is just plain stupid because it is relatively inexpensive.I’ll use your example, being the owner of “this old house” I can tell you this is so true. Sometimes my wife just wants to “get something done”. But you really need to plan things out, because getting a wall plastered and painted doesn’t mean shit when you have to tear it out to put in new romex wiring or pex plumbing.So figure out the design, and respect the fact that production is expensive.

      1. Brandon Burns

        You’re right. Fred is making an excellent point about “planning before doing.” I’m reacting to the details of how to do it, because I don’t think they lead to the end goal of actually doing the planning part well.

      2. LE

        One of the first lessons I learned in business (early 20’s) was when I hired an electrician to put up a panel for some new equipment we were installing. About a year later we had more equipment and he had to come back. The panel wasn’t big enough. So he had to rip it out and install a new one. When if he had put in the bigger panel at that time we’d have room to grow. With a cost that wouldn’t have been that much more. He wasn’t gaming me either. He just saw that one panel was $400 and we only needed the $250 one and so that’s how he quoted the job. (Figures are made up to illustrate but that’s the general idea.)Hence forth I’ve realized that you can rarely go wrong by spending a little more and getting the larger hard disk, faster processor or bigger LCD display ($$ permitting of course) as a generality because the downside is that you’ve spend a little more money.I’m actually proactively spent money maintenance wise at my house this year for things that I don’t even absolutely need. My wife is fine with that. At a commercial rental property that I have I installed a new HVAC (roof of building needed a crane) so I wouldn’t have a new tenant move in and have issues. It’s 25 years old and it’s going to break soon so why wait? Get it done between tenants. [1] Better price as well when it’s not a rush job. Or what happens if it breaks in the winter and there is a foot of snow on the flat roof? You are SOL.[1] When I bought the building after negotiating the price I spend $125 to send someone up on the roof to tell me the age of the HVAC. I then went back to the owner and got them to (and with a bit of effort and strategy) reduce the price by the cost of a new HVAC. Best $125 I ever spent.Point being thinking in advance definitely pays off I agree.

        1. PhilipSugar

          Agreed. The only thing I will disagree on is electronics. When you buy the total top end what happens is you pay twice as much. Three years later you can buy what three years before was the top end for half as much. I.e. you now have two items for the price of one.But when it comes to other things…..agree completely. I had a HUGE argument when I put in my deck (really big porch). They thought I was an idiot. 6×6’s instead of 4×4’s, no hanging no ledger board on the house, 2×12’s instead of 1×10’s, 5/6 ipe deck not pressure treated crap. Extra cost of material?? 25% which was 25% of the job. So for an extra 5% my deck was 10 times better.

          1. LE

            The only thing I will disagree on is electronics.For sure no disagreement on that.Example from today. I really wanted to buy this Sony 4k camera that I just saw last week at the Sony Store that just came out. But I have no real need for it right now. Then I thought “oh but I can use it to film my step son’s bar mitzvah!!!. So maybe I’ll buy it!!Then about 1 hour ago I thought “bad rationalization that’s not for at least another year or so and by then there will be a better one that I can buy for the same money”.Otoh, I failed to get the option in a car that I just bought that does the automatic cruise control sonar driving (or whatever it’s called). I thought “I don’t really do that much long distance driving”. I immediately realized I had made a mistake when I saw how many times in local driving it would be helpful. I had a HUGE argument when I put in my deck (really big porch). They thought I was an idiot.Whose “they” in that sentence? Father? Brother, Wife? Contractor? Neighbor? My dad back in the day used to criticize the equipment that I bought. Old school “do you have the work for that machine?”. He wasn’t in a production business so he didn’t understand all he knew was merchandise. Likewise your critics I’m guessing don’t think like you or have a feel for it. The guy who wanted to install the concrete for a pad we had done told me wire meshing was $100 more and he recommend it. I didn’t care if he was ripping me off for $100 just do it. He did a great job he deserves to get paid is my feeling.A saying that I developed when I was in my first business was “right tool for the job” very important concept.I just bought a Dewalt Boroscope so I could diagnose a leak prior to calling the plumber. I didn’t want him to arrive and then spend his time figuring it out. Hardest part was figuring out the best one to buy on Amazon. Took a long time. My wife doesn’t ever question that type of thing either. In a way that’s a lot I guess like managing people. You have to give them some leeway even if you don’t agree or don’t understand.5% more is a no brainer to me.

          2. PhilipSugar

            Argument is a strong word, but literally everybody. Architect could not understand. Contractor was dismayed. Since I buy the materials the lumber shop couldn’t understand (they take the plans).Even the building inspector. I was there for the inspection. He kept looking and looking. I said is there an issue??? He said are you the owner? Yup. What are you planning to do on this park cars or land helicopters?

          3. LE

            I find this discussion fascinating on several levels. First as I’ve found and observed in business (and not saying this is what happened here btw.) if the woman tells you she wants a fur coat you don’t tell her she can keep just as warm with a cloth coat. You sell her a fur coat because in her mind that will make her happy and besides she has the money. And she has perceived some benefit in her brain from having a fur coat. I don’t wear a watch but I do understand that some men like having a fancy watch. I wish I liked watches I would buy one. But my brain doesn’t care. What fun I would have if I cared.The building inspector doesn’t even count at all btw. He probably drinks and smokes and is waiting to retire so he can sit and do nothing all day. If he had any skill he’d be at least doing construction and if he had more skills he’d be Toll Brother.Nor does the lumber yard. That’s why they are doing those jobs (my assumption is you aren’t dealing with the owner or perhaps to him it’s just merchandise who knows).Contractor just wants to get the job done. What happens after that isn’t a concern of his.Now when you sell that house you can use that subliminally with the realtor and the potential buyers to reinforce that whatever you did (that was hidden that they can’t see) was “top qualify” and done right or better than right.

          4. PhilipSugar

            I think its from understanding the price of everything and the value of nothing.

    6. Matt Cassity

      … This is precisely the point being made in the post.Building a product that hasn’t been properly designed is a waste of production time.”Inexpensive” and “Design” may be confusing words here:Inexpensive: Compared to and building and re-building based on “I’ll know it when I see it” — using an iterative design process to prototype and test assumptions is a much better value.Design: The iterative process leading up to production. A thoughtful look at that process done correctly:

      1. Brandon Burns

        Yes, the main point is a good one and I feel bad for not highlighting that and being nitpicky instead.Still, the processes needed to achieve the goal of making production efficient matters.I checked out your Diamond theory. I’m not sure I agree. There’s all this “planning” before “execution” but “design” is in “execution.”Yes, designs need to be “executed.” But the design process is one of planning. The first thing I do as a designer is research and data analytics. All of that informs the design, which is the blueprint for the developers.When you have “product people” and other non-designers leading the planning part of building a product, you run into problems because they do not have, to the degree that a designer does, the ability to mold their ideas into something that will actually be functional because, well, they don’t design or build things that are functional.However, most startups are founded by business people, so they trumpet theories and processes that make sense for them. But that doesn’t make it the best option.

    7. JamesHRH

      No no no B2.Fred is right. A small group of the right people getting the Product & Positioning right is still the most valuable portion of any go to market process…..but it is also still the one that requires the least capital.Production captures the value defined by design. Produciton is, in all types of endeavours, the capital intensive part of the process.

    8. SubstrateUndertow

      Yes !It took a lot more time and effort to get the human DNA blueprint right than it does to execute quality production duplicates.Millions go years vs 9 months 🙂

    9. mohamed aliem yacoob

      Think Obama Care. I never heard of a scope document for that project, operation was doomed to be problematic. Design is crucial.

    10. Ashish Patil

      It’s essentially defining your value chain as a startup, which includes a strong design process. Apple benefitted immenesly by having their design process being an integral part of their value chain.I think the world of design is misunderstood by business execs & technologists alike. They see it as important, but so many different types of design talent need to come together to get the function & aesthetics right.I’ve become fascinated with design processes (being a UX designer myself) and actually wrote some blog posts around the design processes with my most recent one here:….

  8. JimHirshfield

    Making babies is easy.Raising them is hard.

    1. Tom Labus

      Now you tell me!

      1. JimHirshfield

        Yeah. Too many people think it’s just plug n play.

    2. Richard

      Getting to into the right bed, with the right partner, with the right mindset, at the right time, for the right reason is super hard, after that raising kids is easy.

      1. sachmo

        LOL… yes, I think this is the right analogy.Development might be ‘cheap’, but getting it right is super hard.

    3. JamesHRH

      Never heard that before. Love it.

    4. jason wright

      too easy (barrier to entry is so low), and why it’s hard.

  9. SD

    I see development and planning as high-leverage activities, where small changes can have big effects. In the best cases, production is able to provide an honest view (NOT sandbagged, not overly optimistic) of the risks of a particular path.I remember a case where we knew a technically complex feature would make the product significantly more appealing (i.e. – higher potential sales). But because it was hard to do, it added an uncertain number of months to product launch timing. The production team helped the planners understand very specifically what could add time to the process, and as a result, the planners felt ok taking the risk, fully understanding the consequences.Ultimately, it didn’t work out – we tried to produce, but we couldn’t add the feature within the time allotted. It would have pushed the launch back by many more months. But because there was clear dialog, it was possible to dust ourselves and move on.

  10. Tom Labus

    It’s worse on the business side, Management never commits in case they need to cover their tail.

  11. ErikSchwartz

    Learn to iterate in the prototyping phase rather than the engineering phase. The deliverable to engineering is a functional prototype.This is why PMs must be able to code.

    1. Bruce Warila

      Yes, PMs should be able to code, but the time spent coding (for a PM) is not time well spent. Average coders are a commodity. PMs that talk to customers (a lot), that have great listening skills, and that can actually distil product vision from customer gibberish are rare, and probably shouldn’t be coding. I would rather see four hours invested into OmniGraffle comps + ten hours of customer feedback, versus three days invested into writing throwaway code.

      1. Steven Cohn

        Agreed. Time spent for a PM coding a prototype is a waste. You can build a great prototype using a tool without code and then spend your time validating that prototype with customers.

        1. ErikSchwartz

          That very much depends on what kind of products you want to prototype. There are many products with little or no visible UX (more and more as data and embedded systems eat the world). Sure you can make simple cognitive walkthrough prototypes (and those are useful) but you can’t prototype a complex state driven UX without a system in which complex logic can be specified.It is very difficult to prototype a data driven product if your prototype can’t access data.When your product is an API it still needs to be prototyped.

          1. Steven Cohn

            agreed. the tools i mentioned are for UX driven products.

          2. ShanaC

            actually, the worst is if the UX is displaying or dealing with the manipulation of data/does something where the logic in the back end is data intensive. And you’d be surprised the amount of products where this happens and this is implemented poorly from both a UX point of view and a data/product point of view because most PMs and engineers do not realize they are dealing with complex math problems.EG: I can choose a bunch of audiences to target based on previously taken actions. But I would like to link those actions to the probability of taking new actions. Should I therefore show all possible previously taken actions to target against that I can track? Should I show none? Should I show as a general choice to target against previously taken actions and let the system figure out what previously taken actions matter?(all of these are different options, all of these require different knowledges of users and math exposure of what is possible, many pms won’t even know to ask the question)

          3. Twain Twain

            This is SO true. Pretty much all the websites in Web 1.0 were badly designed from a data extraction and analytics perspective. So along came Web 2.0 with its structured meta-tags, APIs and JSON, OAuth and wikis, and mobile.Yet doing data is still such a messy patchwork and like “wrestling with an octopus”.There’s a prevailing philosophy in both Web and mobile design of “Just put a text box there, some 5-stars, push buttons and click links” and collect the data. We can figure out how to do the Data Science later………..As an example, what a lot of PMs and Engineers don’t realize is that the input in the text box is supposed to go into the {i,j,k} cell of the Data Science matrix as an action insight — not just the # button pushes and links clicked.Instead, the text inputs are treated as separate from the other actions in the loopback. An engineer somewhere scripts some Python to measure the word count and maybe do a bit of sentiment analysis on it; sentiment analysis which is invariably fuzzy rather than reliable.

          4. ShanaC

            it is really not though. That’s the shocking thing. If you can explain the problem clearly enough, you could actually figure out how the matrix of the data affects the UX. Most people just don’t bother to learn about the questions/way to explain.

      2. PhilipSugar

        Average coders are a commodity. Great coders which literally are 100 times better are not.If I had to say what one thing has made me successful, it is my belief in this.

        1. ErikSchwartz

          My team’s engineers are too valuable (and the problems we are tackling are too hard) to waste their cycles on iterating broadly scoped consumer UX.We fired our only commodity programmer.

      3. ShanaC

        depends on the PM and the product – I know at least one PM in one major NY startup that is PMing core math architecture of a data science product, so something not necessarily customer facing. He doesn’t have a good enough grounding in math to tell if what he is building is right, because if you go up and ask why he is doing it way a, he can’t explain it (ironically, his wife, who is in a math field, can). A mathematician who could listen would actually be a better pm in this case.(and this pm can code)

    2. Steven Cohn

      Prototyping does not equal coding. You can build amazingly interactive prototypes without writing a line of code. Check out Axure or UXPin.

      1. ErikSchwartz

        The definition of coding has changed. I do most of my prototypes in Livecode. The hard pard of programming is not the syntax of the language, the syntax can be visual or textual, makes no difference. The hard part is understanding the flow of how an application solves the problem. If you can do that easily you can spend your time figuring out if the problem is the right problem.My point is that the PM should be able to iterate a prototype quickly without having to depend on outside engineering resources. It’s about increasing the cycle rate of iterations.

        1. Steven Cohn

          Hi Erik. I understand your point. And I am countering by saying a PM can iterate very quickly on a prototype without writing a line of code by using Axure or UXPin.

    3. Twain Twain

      I went to several Product Group meetups in NYC and there were some really heated debates about whether PMs should learn how to code and whether Engineers should learn more about business models.One PM said: “That’s all wrong! PMs don’t need to know how to code! The roles are clear. We shouldn’t mess with them! Engineers have no idea what they’re supposed to code. WE TELL THEM WHAT TO MAKE. They just do what WE tell them to do.”I disagreed with that PM so much.The PM-Engineer interaction should be symbiotic. When a PM knows how to code, it can help the team because that PM doesn’t end up “top-down ordering” the engineers to build things that simply don’t make sense.Plus when a PM shares their insights into product-market dynamics and business, it empowers the engineers to be more end-user centric and to develop their careers (especially if they want to become PMs).

      1. LE

        Engineers should learn more about business models.Business, like programming is something that you learn over time. I don’t think you can have more than a rudimentary feel for business models unless it is your full time job and full time thought process. In fact you have to live it and breathe it and like it 24/7. Like the same way some people obsess over sports. If not, you are someone whose mind is elsewhere and phoning it in.The argument I’d give for engineers not learning business models would be simple. “A little knowledge can be dangerous”. Business is analog engineering is digital. Different brain structure. If you know just a little or you don’t keep on top of things you know jack squat. Would be like me trying to learn about sports something which I know practically nothing about. (Did read an interesting article about the Sixers and draft picks the other day..but I don’t have the seat of the pants feel to know how right or wrong that strategy is..)When a PM knows how to code, it can help the team because that PM doesn’t end up “top-down ordering” the engineers to build things that simply don’t make sense.When do they find the time to learn how to code and also understand enough to actually have skills in that area that is their responsibility? I’m sure there are examples of some people pulling this off but I’d hardly call it something that most people can do. Enough to be helpful. I could be wrong of course (hey that’s my gut feel).I mean in theory if you decide that you only want product managers that either know how to code or are willing and able to learn how to code you will end up with a smaller pool of talent. Seems to be best to stick with someone who has the most talent in the area that you are interested in (as a generality and taking the other side of this discussion.)Would Fred be a better VC if he knew how to code in modern languages? I’d argue that that would mean he’d be doing less meetings and travel in theory and that would result in a lower investment return, not a higher return.

        1. awaldstein

          Product managers that don’t have a sense of the business are a dying breed.Out in the world i think you will find that your view of pms is not how it is any longer.i sure don’t find that and certainly don’t look or it in my hires.

          1. LE

            Product managers that don’t have a sense of the business are a dying breed.Huh? Whose talking about product managers not knowing about business? I’m talking about engineers using a little knowledge about business and thinking they know about business. That was my point.

        2. Twain Twain

          Taylor’s Scientific Management model where there’s a distinct, hierarchical separation of roles happened because of Adam Smith’s “division of labor” principles. In this way, expertise builds up in an individual and they focus on one specialty (e.g. perfecting how to screw a bolt in place whilst someone else knows how to polish rough edges or paint the product a nice color).Taylor’s model worked for a lot of companies in industrial production where efficiency was paramount: car, food, fmcg etc. and when management didn’t want the costs involved in training employees to be multi-skilled.We can map Taylor as being equivalent to the “Waterfall” methodology in technology builds and the “Push-Broadcast” of media practices. It’s linear, hierarchical, sequential.Over time, Taylor was superseded by Six Sigma where roles were more synergistic. People still had their specialisms but the feedback loop mechanisms between their roles was more two-way.Six Sigma is equivalent to “Lean Agile” and the two-way street of social media interaction.The observation about “different brain structures” is an interesting one as is whether deep knowledge in one sector is better than distributed knowledge on a multi-sector basis.The reason is because in AI there’s an emerging argument that to solve some of the harder problems, it needs to be done on a multi-disciplinary basis and beyond its historical limits of a Mathematics and Computing focus.

          1. LE

            The reason is because in AI there’s an emerging argument that to solve some of the harder problems, it needs to be done on a multi-disciplinar basisWas the basis for how Bell Labs and the Manhattan Project worked iirc.The observation about “different brain structures” is an interesting one as is whether deep knowledge in one sector is better than distributed knowledge on a multi-sector basis.No doubt depends on the task at hand. If brain surgery the best brain surgeon with the highest level of skill in that area.. Otoh, if hiring a divorce lawyer, legal and strategy knowledge is better than just the best “legal” mind. And so on.

          2. Twain Twain

            Exactly. Everything is relative. No one wants to be operated on by a pharmacist who happens to be interested in neurosurgery but hasn’t been through formal training and is therefore neither qualified nor experienced.Re Manhattan Project, look at what Google is trying to do in AI:*…They’re tapping into Neuroscientists:* http://www.technologyreview…They also have the child psychologists and designers working on versions of Google search, YouTube etc for the Under-13s:*…Is and will Google be “running the table” in AI and all our data online? Who knows? Maybe we’ll read an AVC post in a few years about that.@fredwilson:disqus

          3. LE

            Thanks for pointing those out.From one of those links:Diwanji says that watching those kids tinker reminds her that a child’s-eye-view of, say, the Google search engine isn’t remotely the same as an adult’s. That fact was brought home by her younger daughter, who after Googling “trains” was stunned to see a list of Amtrak train schedules pop up.”She came to me and said, ‘Mommy, you should tell Google about Thomas the Tank Engine, because Google obviously doesn’t know about him,'” Diwanji says, laughing.Step in the right direction. Right idea but wrong implementation. I mean think about i. What could possibly not be representative of “kids” more than the kids of children of google employees? I would guess that those kids would tend to be way higher in aptitude than average kids.I had a girlfriend a few years ago who had a 10 year old daughter at the time. I used to take her out and we would have lunch. I was amazed at her level of comprehension at that age. I knew she was special just in how she could keep up with things that I was saying and the questions that I would ask. No surprise that I find out (we broke up when the girl was maybe 12) she is now at the top technical high school in Manhattan or Brooklyn wherever it is located. In other words she is atypical of 10 year olds when I met her she was very advanced. Likewise my current step daughter is very advanced as well.They also have the child psychologists and designers working on versions of Google searchThat’s great if they are doing that but a startup doesn’t need to spend google money on this. They can easily setup their own tests and just observe. This is not a limitation of money.Charlie doesn’t need to hire a consultant to figure out helpful minutia with his Lancaster bread on the shelves. He can just go in and observe and interact at different stores and talk to customers. [1] That will get him 85% in the direction of where he needs to be with his product (if needed I’m still waiting for it at my local store..)[1] I’m sure he’s done that.

          4. ShanaC

            it is had, because tracking children gives the ethical oogy boogies to many many people.

          5. LE

            Hard is this context is good though not bad.Things that are easy are also easy for your competition to do. Things that are hard keep out the pain in the ass bystanders. Create barriers.I hate that kids are treated like sacred cows. I feel weird every time I’m at Starbucks and see some kid with their parent and I want to make some comment to them. I still do that but always feel as if I’m crossing some kind of line.

          6. ShanaC

            I wonder if agile is actually making this harder or easier. It is really hard to tell, because sometimes the mixture of production/development makes a better product, but then other times you are spinning wheels, for basically the reasons describes.

          7. Twain Twain

            Agile can seem like spinning wheels because it can be “production by consensus” where each member of the team is pivoting and adapting but not necessarily in the same direction and so they counter-act each other’s pivots.Meanwhile, waterfall can be “production by dictatorship” from the top.As with everything, it’s about people and individual working styles.A product lead can be persuasive and encourage others to pivot in the same direction and facilitate the flow of the ball (the product) or they can micro-manage and only pass the ball to people already facing in the same direction as they are.

  12. Brandon Burns

    “In the kinds of companies we invest in the “development” work comes from the product organization. The “production” work comes from the engineering team.”I don’t get this, either. Why is there even a “product” team? This absurdity is widespread, and is the dumbest part about the tech industry.Newsflash: EVERYONE WORKS ON THE PRODUCT!There is no product team. Designers make the blueprints. Developers build. The result is the product. The end.

    1. ChuckEats

      who talks to customers? to sales? to marketing? to legal? to BD? who decides on the feature set? who cuts features? who determines the strategy for rolling out? for determining success? who makes the final call when someone disagrees? who is the owner?effective PMs can be engineering-oriented, design-oriented, metrics-oriented, or even psychology-oriented, depending on the organization and the product; but there has to be a multi-discipline person that owns the product.or else nothing gets done. this has been proven endlessly.

      1. Brandon Burns

        that person should be of a design background.why? because the person who decides on features needs to know how to make blueprints for them. otherwise you get people who may have good ideas in theory, but its a crapshoot if they’ll work out in practice.

        1. ChuckEats

          might work in some cases. from your logic, you could easily then argue that person needs an engineering background – so they envision something that is build-able.effective PMs in large orgs have breadth of knowledge that spans the entire organization. i’d argue design should be reserved for a specialist, if resources allow.but nothing should be built in a silo. effective PMs are constantly communicating w/ their teams; so they should not be dreaming up something that is a “crapshoot” – it’s something everyone has already agreed to.

  13. Steven Cohn

    This is literally the reason why I started As a multiple time entrepreneur, I was frustrated by the cost and delay in “Build-Measure-Learn”. I wanted to measure and learn before code. I couldn’t have written a more perfect post for the thesis behind Validately.

    1. Bruce Warila

      Steven, your product looks interesting. When I watch intro videos I love to see screenshots, and screenshots, and more screenshots. Will explore more later. Best of luck with your latest venture.

      1. Steven Cohn

        Thanks Bruce. I understand your point about adding screen shots in the video. The challenge with that is that the product was in constant iteration. So the video would be out of date quickly. That’s why we went with an animated video that hopefully excites you enough to sign up and see the live product. Happy to give you a demo if you want to learn more. Just email me, steven @

    2. Susan Rubinsky

      Unlike Bruce, I think your intro video works wonderfully. It highlights the major concepts quickly and clearly. Screenshots are for demos, not overviews. A video like this is something I could easily show one of my non-tech-savvy clients to get buy-in. If I showed my clients screenshots they would just yawn and kill the project.

      1. Steven Cohn

        Thank you Susan. Agreed, I wanted to make the teaser video more high level.

      2. Bruce Warila

        when you have to onboard nerds, screenshots are the roadmaps that tell us what we are getting into 🙂

        1. Susan Rubinsky

          Oh, I totally understand that. but when you’re selling through to clients or management, they just want a nice concise overview.

  14. William Mougayar

    I’m with you, and I think the contrarians are getting caught in the semantics of your title (which is confusing btw). This is the nugget that is correct, IMO:”The design process is relatively inexpensive. The build process is not.”

    1. awaldstein

      Contrarian here.Please share with me an example of a build process that got it right for the market at anywhere near 50% on the first go round.Point–the design process is never over and is part of build in all but exceptional cases.

      1. William Mougayar

        Fred has clarified that this applies to the scaling stages, and post product/market fit iterations. If you don’t get the design right, nothing in production will help.I’m not sure what we are arguing about. You’re saying that the go-to-market is expensive, and I agree. But that’s beyond the engineering/production part that Fred talks about.

        1. awaldstein

          I’m not arguing.What’s true is that the importance of capital and the large percent of time that people need to dedicate to raising it, has changed the vernacular of business.Changed the language honestly from a top down perspective.To most analysts and investor and the big company executive it is all top down as that is a point of view that enables them to think of the impact and value of their capital and resourcesTo the operator, as you may remember my friend, it is nothing of the sort.Its a maelstrom and a constant stumbling forward with poise and to the best, the ability to forecast capital needs way in advance and raise them.It’s that disconnect between raising capital and building a business that make the startup CEO such a crazy and wonderful conundrum.

    2. SubstrateUndertow

      Seems to me there is some degree of conflating the relative iteration costs of the design phase vs production phase with the relative importance of each phase?Zooming out on the overall discussion I’m hearing mostly agreement that much attention to the design-phase-iterations is of great important specifically because those iterations tend to cost less and seriously cut down on the number of more costly production-phase iterations.

  15. Robert Heiblim

    Your point is very well taken. Not only product development but the overall lifecycle of a product and service need to be considered and those stakeholders with parts of it included at the very beginning of planning. All too often market realities, sourcing or logistics, sales, customer cycles or other factors impinge on the product or service specifications, prices, timing, etc. I see far to often these things being done in a serial manner handing off team to team. This usually leads to revisions, delays all things that cost. Even worse this is often “adjusted” in the market at the highest possible cost both in dollars and often brand. Developers and marketers need to consider how to get the whole team on board.

  16. Stephen Bradley

    Prehistoric carpenter’s blog says it a bit more simply…Measure twice, cut once.

    1. Susan Rubinsky

      Yeah, but if the design was wrong from the beginning, the measurement is correct but was wrong for the product/service. Carpenters are those guys who build hideous home additions for half the cost. Just drive around any suburban neighborhood and you’ll be sure to see a few. Sure, they’re functional but they don’t really work with the overall structure and surroundings. In the long run, those products (houses) are worth less than well-designed ones.

  17. Salt Shaker

    The design process can be “relatively inexpensive” (or “reasonably efficient”) to the extent a start up engages and invests in the proper due diligence and validation, inclusive of concept and executional pre-testing (e.g., product positioning, attribute/benefit sort and conjoint, price elasticity/copy/conversion testing, etc.). The tech world is bereft of classical product development fundamentals, inclusive of concept and executional testing, iteration and validation.Efficiency (cost) and effectiveness (performance) are not mutually exclusive….They should always be joined at the hip.

  18. Twain Twain

    This is why the Producer-Director-Master Architect role is vital. The startup should have a Product Visionary who happens to also have design and engineering sensibilities.The design process isn’t inexpensive if we’re talking about teams like Jony Ive-Marc Newson at Apple, the storyboard animators at Disney / Pixar and the industrial designers at Ferrari. Their R&D budgets are huge.Conversely, the engineering process isn’t expensive if the Product person knows how to hybridize tools like Cloudability to do cost management on whatever databases the engineering team’s spun up in AWS or other cloud provider.I completely agree that thinking things through (and synergizing all the moving parts) is really important.It’s how I personally approach problem-solving.(1.) At UBS, I created the e-Intelligence platform. My blueprint and roadmap was so simple and clear, the Global Head of eCommerce signed off on it within 2 minutes and the small crack team of engineers had it up and running on an expedited basis across all the silos within the bank (on Intranet, on P2P IM and in email with feedback loops).It was one of the lowest-cost, highest-value productions which contributed to my promotions within the bank.(2.) In 2014, at 4 different hackathons with 4 different teams, my teams won prizes.Lots of teams have been larger and technically stronger than my teams. They all coded A LOT but most of their code was what we can call “fat code” rather than “lean code”.At Angelhack SV, one competing team came to me for advice after they’d spent 12 hours overnight trying to hack a food-ordering and delivery app. I took at look at their UI and their Github stack. Within 2 minutes I’d sketched how their user flow SHOULD have worked to save them about 10 hours of coding.Afterwards, one of their engineers, who had 12+ years experience, came up to me and said, “I don’t know what your background is but you’re a really really strong Product person. You really know your stuff. You took apart our code and saw what we should have done…just like that (snapped his fingers).”Meanwhile, a designer from another team came up to me and said, “When we met I thought you were a designer too. But now I realize YOU DO EVERYTHING!”My team of 3 won a 3-month incubation at that hackathon. Only 1 of us did any of the development work; he was the only one with an Android and we’d decided to code the app in Python. It took him about 8 hours, based on my wireframes.I later re-coded the app for my iPhone which took 2 hours.(3.) This month, I’ve been asked to consult on a global Data Science and Machine Intelligence project. The team’s been working on the project for most of 2014.I took 2 days out from my day job to get a handle on what they’re trying to do. [Day job is a perceptions platform I’ve built which will go live in 2015.]When I showed their senior people my blueprint for solving their Data Science problem, they were astounded. They didn’t think I’d get up the curve on their sector (I have 0 experience in it) so quickly or be that forensic in understanding the critical moving parts.One of the things about your Dad teaching you how to play chess when you’re 5 is that you understand about moving pieces of the puzzle and the importance of thinking things through………quickly and effectively and………And playing all the way through to the game’s end and the next game(s).

    1. Susan Rubinsky

      You are a rare breed. What I call right-brain/left-brain. Hope you charge the big bucks when you consult, because otherwise you’re devaluing the rarity of your brain. Whenever I try to replace myself on my team, I have to hire three people to do what one me does. Charge accordingly.

      1. Twain Twain

        Thanks for that great advice, Susan.Unfortunately, we women often under-value ourselves and we don’t negotiate:*…*…Part of it may be social conditioning: “Good girls are supposed to help others out of the goodness of our hearts rather than ask for payment” or something like that.Progressively, I look at the fee schedules for consultants across 4 dimensions:* Strategic management (product & financing)* Design * IT (engineering and database architecture)* Legal (securing trademarks from bluechips and patent process)And I inform the client company I’m bringing my proprietary systems into the project.

        1. LE

          Unfortunately, we womenInteresting I have been reading your comments for some time and never would have guessed from your writing style that you were a woman. I can usually tell that by the way someone writes. .Unfortunately, we women often under-value ourselves and we don’t negotiate:I run into this with my wife who doesn’t want me to open my mouth at the restaurant to complain about something that I don’t like. It bothers her to no end. It’s literally almost the only thing we disagree on. We booked a hotel and the rate dropped. I told her to call the hotel to give us a lower rate she said “no way that’s your job I can’t do that”. (She’s a Physician by the way and she barks things over the phone to interns, residents, nurses and ER docs all the time when she is on call from home). But she won’t complain to the hotel! I didn’t know what the big deal was. She said also “they react better to a man’s voice”. That may actually be true there was an article in the WSJ this week about charisma that dealt with some of that (will try and find the link). Her boss sent her a contract renewal letter with no raise I strategized and gave her the moves to get them to give her a raise and a better contract term. When I continued to push her to get even more she didn’t want to do that she didn’t feel comfortable being “to pushy”. My mom is giving me a hard time currently she doesn’t want me to negotiate on her behalf for the life care community that she is going into. She is worried that her son will embarrass her or something with the person in the sales office! Imagine that. It’s funny and sad. I have to knock some sense into her if I can find the time. Those places are super expensive and everything is negotiable.That said while it might be easy to hide behind the “I don’t have a deep voice” I do a large amount of negotiating by email where my voice is never heard. But I do think there is truth to that.The reason is when a strong and/or angry voice of a man is heard the other party is more likely to be fearful and that fear sometimes makes the other party give in to avoid what they subliminally feel is conflict which they want to avoid. Women can do this as well though but it is harder.

          1. Twain Twain

            Haha, yes, I’m a woman. I learnt in my 20s at UBS that when people see my writing they often think I’m a man. I was board observer on 20+ investments and responsible for reporting on our portfolio to all the Global Heads and senior MDs so it was all good practice.My manager sent me to a Negotiation Skills course and, afterwards, I secured the bank’s board seat at against 25 other financial institutions and their MDs who were twice my age and higher up in the corporate ladder than I was.The number of times people did double-takes when I entered the room and sat down at the table…….Instead of being the PA who was there to pour the coffee…….Luckily, my brilliant manager would set them straight: “This is Twain. She’s the one you’ve been emailing and she’s the one who wrote everything in these slides. So if you have any questions, please ask her.”

          2. LE

            I think you’re a perfect example of why people shouldn’t whine about some “shortcoming” they should just work harder if they have some real or perceived disadvantage over others. I’d like to be taller (and better looking I mean who wouldn’t, right?) but I don’t use that as an excuse and complain that it shouldn’t matter to the world. Because it might in some cases. I just figure out a way around that when I have to. Actually that’s a good example as a result of that “disadvantage” I developed a way to run circles around the bigger animals which in a sense is what you were able to do. The bigger stronger animals tend to use their size and presence to intimidate and get what they want [1] whereas the smaller animal needs to use their brain and creativity to solve the problem.[1] Which ends up resulting in a lazy brain.

          3. Twain Twain

            Thanks, LE. My parents enabled us to put our brains to hard work from the earliest age so that’s what I do. We were at school all 7 days and every evening was extra-curriculars (art, computing, music and sports).I wasn’t one of those kids who made excuses like “Dog ate my dinner”, lol.

          4. ShanaC

            I didn’t know either….

        2. Susan Rubinsky

          Do that then multiply by three 🙂

    2. ShanaC

      how did you get this way?

      1. Twain Twain

        It’s all due to my family and teachers.My grandparents separately discovered I was left+right brained when I was 2, my mother likes to remind me. Given that my parents were multi-lingual and technical, it’s likely I benefitted from their influences early and, unlike Sheryl Sandberg’s experiences, no one ever called me “bossy” or made me feel bad for being bright and sharing my brightness.At school, my first-ever report said I had a particular interest in two subjects: maths and art. Then when I was 12, I passed some IQ tests and the Royal Society invited me to their maths masterclasses where I learnt about Turing.Throughout school I was an all-rounder: straight A’s, Captain of Sports, Head Girl. I’ve always attended mixed schools and collaborated/competed with the opposite sex in exams, so that acclimatized me to gender dynamics very early on. Comp Sci was compulsory from 11-14 and I later chose it as an exam elective along with Physics, Chemistry and a handful of other subjects.I graduated high school with the highest exam marks in my year; including being the only girl to get A’s in Maths and Comp Sci. Plus I put my art portfolio together in 3 months instead of the 2 years the other students had. Art wasn’t one of my electives because of a timetable clash but the Head of Art decided to enter me into the national exams based on a doodle I did.Meanwhile, my languages teacher was hopeful I’d go on to study Modern Languages and become a UN translator. My childhood languages were: English, Chinese (Mandarin and Cantonese), French, Italian, German and Spanish.From that knowledge base is how I became this way.

  19. pointsnfigures

    Dev Ops, the unsung hero of growth.

    1. awaldstein

      so so true

    2. Andrew Kennedy


    3. Emily Merkle

      long live DevOps

  20. Rob Underwood

    Great post. I’m sharing this with our leadership team here at Relay. Very applicable.

  21. BillMcNeely

    The plan not so important. The Planning Process is awesome. 2 slide post conventional fight Iraq anyone?

  22. BillMcNeely

    even raw startup should have planning. Lean Business canvas gives you the start point to test/theory get to scalability

  23. Twain Twain

    The “Is the value/cost in the design or in the production” question isn’t exclusive to the world of tech, software, Internet and mobile.My teenage years were spent in the development labs of the world’s second largest aroma chemicals company.Basically, if we don’t think through and write down all the ingredients, equipment needed, how to make it and who’s going to drink it BEFORE the product’s made there’d be:(1.) The risk of accidentally poisoning yourself and other people or making something so unappealing that no one would drink it.(2.) A lot of really expensive chemicals wasted and poured down the drain.Chemicals can be even more expensive than code to make.(3.) Getting the mix and ratios wrong which would have made it impossible to scale 6 prototype drinks up to 600 million drinks.There are drinks my team made which you folks all have in your fridges today.

  24. Jan Schultink

    Traditional management in big corporates is to lay out the objectives, and get a bunch of highly motivated people to run with it, iterating as you go along.I discovered over the past year that designing a software product V1.0 is different. Extraordinary attention to detail, nit picking, until – you think – you get it right.It goes right against team work, delegation, and all the other stuff they teach in business school or the boxes you see in big corporate annual performance reviews.My guess that things are more back to normal again when you hit V2.0.

    1. LE

      Traditional management in big corporates is to lay out the objectives, and get a bunch of highly motivated people to run with itPeople who end up rising within corporations are luckier than they are good. Statistically some will end up being good enough that every bet that they make or situation that they end up in works to their advantage in some way. I thank my lucky stars that I don’t have to work for a large/medium/small arbitrary corporation who is only able to reward success and doesn’t and isn’t able take into account specifics in determining why something didn’t work. Just the bottom line of evaluating “got the job done” or “didn’t get the job done”. And all the politics that go along with that as well. At least as an entrepreneur you can choose what you want to get involved in. In a corporation you usually don’t get that choice.With entrepreneurship you can actually make some mistakes and still end up in a great place in the end. You don’t have to be right all the time just enough to end up putting a dollar in your pocket. You don’t have to make decisions to cover your ass just decisions that in your judgement overall end up being more right than wrong. There is leeway and that leeway (and this is important) prevents you from having to mentally stutter when you make creative decisions. Because you know nobody is going to judge you on each and every thing that you do.

  25. oh7dp

    Actually, I find it’s the opposite– especially if you can use something like Kickstarter to presell the results. If you can guarantee N sales before tooling up, you can usually get the mass production costs to be manageable.The design takes time and it’s expensive.

  26. Chavi

    In my opinion, development and production are the same thing. And design should go hand in hand with the development.

  27. Terry J Leach

    It is easier to scale production systems with the ever increasing platform as a service systems and tools. I suspect over time the production process will be the commoditized part of the scaling a startup, the real “work” will be the “Thinking through process”. My thinking through process production occurs even during the product market-fit phase, but I have to temper it.

  28. Brett Topche

    Isn’t this basically the old construction adage “measure twice, cut once”?

    1. PhilipSugar

      Great analogy.

  29. LE

    Not surprising that Jerry’s writing style is very similar to Joanne’s. Short terse punctuated sentences. Totally reminded me of reading a GothamGal post.

  30. LE

    From the blog post:They blew me off for a bidder who hit the number with no caveats. I told the producer following this methodology with that vendor would lead to bad mediocre work. It wasn’t that my approach would assure something better but I do believe it would at least give a chance at greatness.Sounds like a person who was referred to me recently who needed some help and said “can you get this job done for us”. My answer sounded something like “sorry I don’t give yes or no answers” to questions like that. This isn’t physics or programming it’s a creative process that involves plenty of analog decisions along the way. I can’t predict what the counter party will do how they will react I can only minimize the impact and try to flatten them based upon my experience. (Not the words that I said I obviously don’t talk like that..)I’m getting bored of saying “you can only be as honest as your competition” [1] which in this case essentially means it’s difficult and perhaps pointless to try and do the right thing when a competitor is willing to take a chance that you aren’t willing to take. And make promises that you don’t feel comfortable with. [2]In the end I have no patience for people who want to absolve any responsibility for the minute but extremely important details and most importantly nuance of how things actually work in the real world. The competitor can sleep at night which is why they can make the promises that they do. You can’t be them. I don’t try to be them and I’ve lost plenty of business as a result over the years as a result. However along the same lines my wife doesn’t try to be me in what I am able to pull off and feel comfortable with that she can’t. Everyone is different and has a different line depending on the situation and circumstances..Super nice films in the work section… btw.[1] I might also use Springsteen’s “it’s hard to be a Saint when you’re just a boy out on the streets”.[2] Reminds me of my ex father in law who would promise to install 3 alarms systems on the same day knowing he’d never be able to do more than 1 job. But he could sleep at night doing that type of thing while the more honest competitor couldn’t. Guess who got all the jobs? And when he died he had a boatload of monthly residual monitoring accounts to show for it. I remember thinking “I could never do that it wouldn’t sit well with me”. I’ve always found that type of “integrity” to be a limiting factor that is “caring to much”. It’s a difficult balance to keep.

    1. ShanaC

      so should you be honest? (you in the general sense)

      1. LE

        How honest totally depends on the circumstances and the specific situation. Upside gain and downside risk. And each person has a different comfort level and financial situation and competitors.The point of my “saying” which is “you can only be as honest as the competition…” is that if your competitors are lying you will then have a hard time not matching them to a certain extent. Anyone who tells you different is either lying themselves (trying to appear above the fray and holier than thou) or has never been in this type of situation or is deluded into thinking that honesty always wins in the end. It doesn’t. Sometimes honesty buys loyalty and sometimes it loses the sale. I got my first big contract by telling the buyer “if you want to make the right choice choose Xerox and not me…”. That was a hail mary and psychology play since I knew I couldn’t give a good reason to choose me other than that (it worked). I didn’t do it because I was honest. Otoh I hired people to come in and fake that they were working for me when it was time to be inspected by the purchasing department. With that I wasn’t honest (also worked). Otoh I rarely promised things that I couldn’t deliver. Not because I’m so honest but because the way I was raised that would make it hard for me to sleep. In all honesty! Note how honest I am in relaying how honest I’m not! Or am, but for specific reasons.I guess it’s probably a subject that binary thinkers have a really hard time with because they can’t see the nuance of different situations and how to adjust their approach.

  31. Emil Sotirov

    The argument can be flipped the other way – spend more (time/money) on design & development if you don’t want to have expensive production. The proportion depends, of course, on the character of the product – novel vs conventional, simple vs complex, etc…BTW, in architecture and large urban projects – production is almost always way, way less expensive than the consequences of bad design.

  32. Medicalquack

    This is a good post and I have to say I agree with this as I have watched some get promoted that don’t have a good model from the start and we have this duping effect out there with a proof of concept that doesn’t work in the real world, and yet there’s been so much money put into development the flawed model gets pushed upon consumers anyway, again because of the development cost and folks stuck on stats if you will. Numbers can fool you and bring in the physicists to get back to reality:) Big Gulp anyone? There’s a perfect very public example with a lot of power, money, etc. devoted and look what happened, the old proof of concept just didn’t work in the real world:) Bloomberg will have that dupe tied to him forever but hey we all get fooled.http://ducknetweb.blogspot….In going a little further with this I borrowed a phrase from Emanuel Derman for this post “People Don’t Work That Way” and again it’s addressing a world of flawed models and it’s not easy out there to see what works in the real world versus the proof of concept today and I don’t think anyone bats a 1000 at this, I never did when I wrote code and a developer can get fooled with their own code, can happen to every one of us, it’s the development game, but key here is to “stop” when you see it’s not going to work and don’t shove broken models at consumers, create studies and reports telling consumers how dumb there are a well, to substantiate as such. I see that out there as well as broken models that don’t work as they don’t change them to update for the times too, so bit of both.http://ducknetweb.blogspot….Lot of wisdom the the book written by Emanuel Derman “Models Behaving Badly” and I recommend it. Read that and I think things will make a little more sense. Again I used the Big Gulp as an example here as everyone can relate as we watched it all evolve and then fall on it’s face, as again “people don’t work that way”. If you watch the Quants of Wall Street video at the Killer Algorithm page, Emanuel Derman and Paul Wilmott, both with physics backgrounds tell you the story flat out and go into the Modelers Hippocratic Code of Ethics they created. Mind you this is financial but it applies everywhere. There’s some other good videos over there as well and most at layman levels. In another area I also wrote about a big problem we have today too which is the result of some of this with marketing spins added and that’s the fact that people confused virtual world values with the real world. None better than Larry Ellison addressed that back in February in stating “be careful with pieces of intelligent software that’s smarter than you, there’s only one LeBron in the real world, where in the virtual world, everybody gets to be LeBron”.…So as I see it, development and implementation costs and potential profits do sometimes get on a crash course with what really works “in the real world” and not just the virtual world. I see it even all the way up to the White House too with this confusion with emails that quote stats that nobody can predict, so a bit of Algo Duping 101 if you will. We want stuff that works and is not pushed on a flawed model simply due to the money spent on both development and implementation:)

  33. Nick Devane

    Great PM + Suitable Dev/Design > Suitable PM + Great Dev/Design

  34. Pete Griffiths

    The first lesson of quality management. The higher up the production line you find the problem the cheaper it is to fix.

  35. SubstrateUndertow

    What if we were to zoom out and apply that concept, of thinking it through, of getting it right, of applying an optimally flexible framing from the out set, not just to any particular project but to our larger contemporary reality, namely, our collectively global project of designing “software to eat the world”.I like to demystify “software is eating the world” into its more grounded/operationalized Mcluhanesk equivalence.As McLuhan frames it, all forms of technology are simply extensions to our three core biological powers. All technologies fundamentally extend/coordinate our biological powers of Awareness – Analysis – Action. These biological-extensions not only amplify our core survival-powers, they give those powers transcendence over both Space & Time. (go ahead do the set theory)We are now adding to all those previous Biological-Extensions a powerful new supercharging Meta Biological-Extension technology, namely, the “Ubiquitous-Network-Effect” thus facilitates unlimited integrative synchronization of all our previous Biological-Extension technologies.When I read “software is eating the world”,I hear,unlimited authoring of storable, recombinant and reusable webs of organic social/economic synchronicity are now available to colonized and remix all our previous Biological-Extension technologies.The power or the self-organizing dynamic that has perviously been availably only to our biological substrate has now been dropped like an organic/nuclear time-bomb into the human organizational arena.Still it seems we are rushing into, that “software is eating the world”, that “Ubiquitous-Network-Effect”, that software-colonization of our complete inventor of historically accrued Biological-Extension technologies with little to no collective introspection for thinking it through, for getting it right, for applying an optimally flexible framing from the out set.Case in point !We are busy building out our present version of that potentially “unlimited authoring of storable, recombinant and reusable webs of organic social/economic synchronicity” otherwise known as the organizational potential of the “Ubiquitous-Network-Effect”, up-side down and inside out !The “up-side down” thing is self-explanatory as most players attempt to reconstitute topdown oligarchic corporate control structures, which is probably unavoidable/necessary as a short term transitional economic reality.The “inside-out” thing seems less transitionally necessary and a more far reaching/intractable longterm constraint on the evolving potential power of the “Ubiquitous-Network-Effect” as a break-through remix social-technology ?By “inside-out” I refer to the ever growing moats being built around all manner of Data/Process software-objects. Add to that the accelerating trend of subsuming App functions within more global Apps as in the Chinese-style integrated Apps now being pursued by the likes of Kik. Apple’s iOS “extensions” seem like a more middle of the road compromise at open Data/Function recombinants while still keeping the outer moat firmly in place.In the short run such self-serving corporate isolationist limitations on the potential for establishing any real up-front universally-interoperable open-standard for Data/Process software-objects will pay big dividend for those corporate players but as a global social strategy for our collective project of designing “software to eat the world” and thus colonize our potential for organic social organization, not so much !Yes… Yes… I realize we can’t get there from here !This comment is just a though experiment.That’s my pessimistic point, we have probably reached our self-interest based evolutionary hull-speed and such cognitively collective introspective social phase-changing organizational behaviours are simply not open to us and our evolutionary ilk.

  36. giffc

    It’s a mistake IMO to say that production is engineering. The best companies today are splitting their product teams into cross-functional units that encompass engineering, design and product management, and often incorporate some specialists as well. They work in tandem, not a linear process.That said, it is easier and cheaper by far to try new ideas without thinking in terms of production code, which can be slow to change. You might still code, but iit is experiment code not production code.

  37. GustavRodewald

    I disagree with his article. The economics of software development are the opposite of building house or any of the other things the author mentioned. In all of those other activities the cost of production dwarfs that of design. Of course what you pay to an architect to design your house is going to be a small fraction of what you pay in construction costs.In software, on the other hand, production costs are almost free because going to production can be automated. This is what DevOps and Continuous Delivery are all about.In software, getting the design right is very hard. Most software endusers have no idea what the want or need. Most house end users have a pretty good idea what they want and need (how many bedrooms, bathrooms, etc…). Because design is very hard in software, we use agile methodologies that exploit how inexpensive it is to take software to production to iteratively design it based on feedback we get from our users and the market.

  38. l555iu

    The key point is:creat new is hard,reuse is easy,yes?

  39. Anne Libby

    It’s also true around people process. If you want to implement performance reviews (for example) start with a spreadsheet. Understand your real requirements. And for heaven’s sake don’t build something yourself. Unless it’s a spreadsheet.

  40. LE

    that’s how I roll but it always leads to problems.Judging by all the complaints about twitter (up time wise) when they started I think they rolled like that as well.fortunately I’ve partnered with people who know betterWhat I have always liked about you Charlie. (Or at least the persona which you appear to be by your comments since we’ve never met..)

  41. lisa hickey

    One of the most useful skills I’ve learned: How to turn a spreadsheet into a to-do list.

  42. Anne Libby