Why founders shouldn't be the developers | Hacker News Reader

Why founders shouldn't be the developers

zemanta.com

143 pointsSwizec94 comments

What it looks like to go from being the first real coder to being a full time manager of a pretty large eng team over the course of 3 years in git commits per month:

  37 2009-09
  99 2009-10
 177 2009-11
 104 2009-12
 136 2010-01
 115 2010-02
 126 2010-03
  72 2010-04
  37 2010-05
  43 2010-06
  30 2010-07
  39 2010-08
  53 2010-09
  43 2010-10
  59 2010-11
  66 2010-12
  13 2011-01
  23 2011-02
  67 2011-03
  66 2011-04
  13 2011-05
  20 2011-06
   0 2011-07
  29 2011-08
  25 2011-09
   3 2011-10
   2 2011-11
   3 2011-12
   3 2012-01
  35 2012-02
   6 2012-03
  10 2012-04
   2 2012-05
   0 2012-06
   0 2012-07
  38 2012-08
   0 2012-09
   0 2012-10
  32 2012-11
   0 2012-12

For whatever it's worth I miss coding an awful lot.

I can understand missing coding, too, but your team will be able to sense that in the same way children can sense the stress of their parents. As a manager, it's important to get over that so you can take yourself and your team to the next level.

How is that bad? Knowing my manager missed coding would only make me happier.

▂▄█▅▆▅▅▃▂▂▂▂▃▂▃▃▁▁▃▃▁▁▁▂▁▁▁▁▁▂▁▁▁▁▁▂▁▁▂▁ [1]

[1] https://github.com/holman/spark/blob/master/spark via ./spark 37 99 177 104 136 115 126 72 37 43 30 39 53 43 59 66 13 23 67 66 13 20 0 29 25 3 2 3 3 35 6 10 2 0 0 38 0 0 32 0

I'm guessing there's a flag to get that printout, mind sharing?

git log --author harryh --format="%ci" | cut -f1-2 -d- | uniq -c | tail -r

(then I added in the 0 months manually)

Learned this just now via twitter btw: https://twitter.com/harryh/status/279010454376960000

Also, if your version of tail doesn't support the -r option, you can reverse the order of the lines using the tac command.

... and if you didn't get that "tac" is a command to reverse some text, then you need to realize that it's "cat" backwards.

I promise, if it weren't so annoyingly hard to implement, Linux would have command called nwodtuhs to turn the machine on. Those crazy geeks! :D

My favorite backwards command is the 'sl' command. If you have a Debian/Ubuntu command line handy, sudo apt-get install sl

I'm curious about 2012-08 and 2012-11. Pet feature, hackathon, large release with hotfixes?

Refactorings to help make our codebase less of a giant hairball and more of a DAG. Big project that a lot of people helped out on and that I cared about a lot. Also it was relatively easy for me to pitch in cause most of the work was tedious but uncomplicated.

Is there a way to get this history out of github / many repositories at once?

Great. You should miss it. It's the only way you're going to have a shot at being a decent manager, let alone good or great.

I run my own start-up and I have to do everything. I am the sole employee. Some days I wear my designer hat, others its server config, and others it's customer service. Is it a bit crazy? Sure, but I really enjoy what I do.

I don't have meetings. I don't go to conferences. I don't pitch my company to VCs. I don't try to raise money. I don't care what other companies are doing. I don't try to use the latest tech because I'm expected to.

LAMP is fine.

I concentrate on building the product, satisfying my customers, and increasing my revenue. Everything else is bullshit.

The founder should be able to do everything. Being a developer is one of those things.

While I totally understand your feeling, you should be aware that this strategy wouldn't work if you wanted to start a company like Tesla or Intel.

No offense, but what you call a startup might be more considered like a business by the HN crowd, and this is why you can operate this way.

Not every type of company can stay out of the now traditional pitching/VC game.

There is a lot of room between a 1-man show and Intel/Tesla. People really think it's all or nothing and there is so much room in between.

It doesn't mean it's a lifestyle business either, you can have a business earning over $10 million revenue p/year in this in between and with no funding. Is that really just a lifestyle business?

Of course there is a lot of room, but it doesn't invalidate my point.

Most people i've met seeking venture capital have zero chance of being capable of building the next giant company. Yet many have succeeded in raising capital regardless. VC's generally aren't willing to take on much risk, so you see them spending lot's on low-capital requirement companies that are +EV. And little on companies that need large amounts of capital to scale.

Nothing wrong with a "lifestyle" company and making a couple 100k for yourself. No need for all the headaches of running a public company and possibly getting kicked out of your own company by VCs, boards, etc. like Jobs did.

If his business is not a "startup" then I would rather be like him than the startup of that HN crown that you're saying.

I'm in exactly the same situation right now and I ABSOLUTELY agree with you on everything. One minor thing here though that I don't call it a 'startup', it's rather a 'small business', which grows linearly, not exponentially (not yet at least).

That's funny you say that. Over the past 6 months I started to reference more as a small business or company and less like a start-up.

I hope that you have plans to at least grow to where there's someone else that can cover for you while you take a vacation.

Amen. I'm in exactly the same situation. Took me a few years to get there but now I'm totally happy with it. I actually like the fact that I have to do everything myself. That's what keeps it interesting (and challenging once in a while).

You are over-generalizing. You are in a fortunate situation where you are able to make this work. What a blessing!

There is nothing fortunate about it. It took hardwork and over 3 years in the making.

Three years? Exactly the same here. I am a one man show and I highly enjoy the current situation. Next year I'll think about hiring one or more student to help me out on minor parts, but I still wonder if I can do without. Caring about employees might not be worth the trouble.

You are forgetting that some people work very hard on an idea for many years and it still may not pan out.

Xpose, congrats! You're an inspiration to all of us who are working at our own stuff

I founded a product called Planscope (https://planscope.io). I'm the only person involved. I develop, design, support, and handle all the marketing.

* I don't pitch 50% of the time (0% is more accurate)

* The lights are on. With more than 200 paying customers and 20% growth month-to-month, they should stay on for quite some time.

* I know most of my customers, and tailor the product around what I learn in talking with them. Thus, customer development.

* Me, myself, and I all agree about our company's vision.

* I keep tabs on my industry by asking my customer's about their business — and it helps that I eat my own dogfood.

* I do take all responsibility.

* And I handle all support

What happens when you get sick and need to take 10 days off? Family emergencies? A company with a "team" can give you that margin of safety. A One Man Band is great, but is not sustainable for long, because life happens. Startups are built to scale so that founders can step back and enjoy the fruits of their labor.

The idea that the goal should be to work for a bit and then "step back and enjoy" while other people take over is a bit of a weird aspect of silicon valley. Successful small businessmen in most fields expect to continue operating their businesses long-term, maybe with a few employees. One of my uncles ran a shop for 45 years! I don't see what's wrong with being a tech small business long-term, either. Past the bootstrapping phase you do need some trusted people to cover while you're on vacation or sick, but it doesn't have to be a huge operation to get to that.

This is also the case with many professional fields. There are plenty of lawyers, accountants, marketers, etc. who start a small firm and then run it the rest of their life without expanding much, if at all. There's nothing wrong with a lifestyle business, it's just not a "startup".

Many startup founders are overworked — I usually put in maybe 3-4 hours a week into Planscope (when I'm not actively working on any major new features.)

I'm not opposed to delegation at all, I don't want to be glued to a screen — I just haven't put all the necessary pieces in place yet. 90% of support requests can be handled via cut and paste by a virtual assistant, and having someone on-call for server issues is easy enough.

A labmate of mine recently shared his term for this [1]: "Bus Factor" -- as in, the number of people who could be hit by a bus and still have the team / company / project function.

With a single-person company, the bus factor is essentially zero... unless you're a master of automation.

[1] Though he said it was common among his peers in Austria.

Another term used in this context is "single point of failure".

For those who can be control freaks with their businesses, this is one of those show stopper risks that creep up on you.

What happens is you have your systems set up so you don't need to tend to them if you don't want to/can't. What daily business does a single individual running an internet shop really have to tend to in the age of automation?

Do you have an API? I'm working on a tool that might fit in well to part of your app's workflow.

Email me: brennan@planscope.io

I have a few customers testing out my soon-to-be public API. Would be more than happy to have you work with it.

An hour… well maybe you can fix a quick bug or two. But you won’t get anything hard done.

I work full time at a startup, and I have a 6 week old daughter at home, and I'm working solo on a proof-of-concept on nights and weekends. Since I've got my hands full with the baby at home, I've learned to be very efficient with my time. One decision that made things a lot easier was deciding to go with node.js. Since the entire app is written in JavaScript, I can quickly test out ideas on my phone or tablet browser using JSFiddle, and integrate it later when I have a laptop handy. Sometimes I only have an hour of uninterrupted time, but that's usually enough to write a function or two.

That approach might be fine for small projects/webapps, but I stuggle to see that approach working for anything more complex/hard. Sometimes the only way to make effective progress is a full day of uninterrupted progress where you can hold the relevent parts of the system in your head at the same time...

Sometimes you need several weeks of working on the problem in this way to make progress! (e.g. machine learning improvements, distributed system design).

This seems like a pretty good idea; was it difficult to get used to jsFiddle on the phone?

Isn't the whole point of startups that they are usually funded (at least initially) on about enough money to buy ramen noodles for a year?

Sure, if you have enough funds to buy some rockstar programmers then go for it.

Otherwise what are you going to do? Outsource it to the lowest bidder? Then spend all that time worrying if they will produce something that matches your vision and worrying about the quality.

No, you put in the time to find a good programmer who will work for a reasonable price. Typically that means finding someone who graduated fairly recently and is looking for additional work, or someone that doesn't have a lengthy track record/portfolio. It's not easy (I've probably talked to a hundred some odd programmers in the last year, and maybe three would fit the bill), but you can find the right people if you put in the work.

It's no different than finding a good salesman, or a good PR person, or a good anything. You can find great talent at a reasonable price if you're willing to dig. Rockstars don't always come at rockstar prices, because a lot of the time they don't even know they're a rockstar.

That's dangerous, because very often they'll figure out they're a rockstar as soon as they accomplish something useful for you, and move on to firms that will pay them rockstar money. Meanwhile you're left high-and-dry with a bunch of code written by a talented-but-inexperienced coder, and you have to find a replacement - often right as people start taking interest in your product.

I worked pretty cheaply for the first year or two of my career, as did several of my friends. I'd like to think that my employers got quite a bargain. But I only stayed with them a year or two, as my $32K/year job became a $66K/year one became a six-figure income, or as my friend's $15/hour wage became $75/hour became $72K/year employment became six-figures.

The expectation would be that our revenues will grow to the point that retaining talent won't be a problem. We only just started the market entry process at the start of the month, so it's too early to say, but if I'm losing talent in a couple years because I can't afford to retain it, then our problems go way beyond our talent growing too expensive.

There are other costs besides price. Someone who graduated recently, or doesn't have a length track record etc may also need a lot of time for self learning, mentoring, accumulate a ton of hidden technical debt, not write maintainable code, have bad naming practices and code conventions, limited experience coding in a team environment, low output, etc. And getting someone to come in to clean up the mess will cost more time and money and effort. This is on top of nostrademons' the market corrects itself eventually.

And I am someone who used to outsource many technical projects to Eastern European developers, and wound up becoming a developer. You maybe getting good talent at a good price, but there's always a cost.

Maybe true for a sole founder.

I was the technical co-founder. My job was to code, work with a few other people who code, and be the overall passionate guy about the quality of the code. I find it helps to have a stake holder in this position, you end up with better code, and someone who is going to take true ownership of the code.

My co-founder was a true marketing/selling machine. So it worked out for us I think.

Well, not that I don't think you have a valid point but some very smart people have done otherwise...

"When we were working on our own startup, back in the 90s, I evolved another trick for partitioning the day. I used to program from dinner till about 3 am every day, because at night no one could interrupt me. Then I'd sleep till about 11 am, and come in and work until dinner on what I called "business stuff.""

(http://www.paulgraham.com/makersschedule.html)

I somewhat agree to an extent, my take on things is that a successful startup needs two founders; a technical founder to do the programming and a business development founder who possess's high amounts of confidence and a vision to push both the development and end goal of the startup.

Having said that, I'm currently a solo entrepreneur trying to balance work, life and startup all by my lonesome self because I don't know who or where to ask for help with my startup ideas with spending a lot of money.

"I somewhat agree to an extent, my take on things is that a successful startup needs two founders; a technical founder to do the programming and a business development founder who possess's high amounts of confidence and a vision to push both the development and end goal of the startup."

A Jobs and a Woz.

The biggest problem I see startups having (at least the ones I have insight into, consulting and whatnot,) is their founders are too far from the technical. I agree founding and coding is hard to manage, but without a developer on the same level with the same goals, you're shooting your business in the foot.

Hence the classic combination of a tech cofounder and a business cofounder. Or, the more usual in real life combination of two tech cofounders, one of whom actually does mostly "business stuff".

I have to disagree here. Having just come out of a failed startup, one of the (many) things that hurt us was having the primary product direction and technical expertise belong to different people.

We were racing against time and lost. A lot of time was spent transferring ideas between people, and implementing something that wasn't quite right. I can't imagine any startup in the same situation not having at least SOME similar overhead. For a small startup without funding, this can't be optimal.

I think it's more than possible to start a company as a developer, or even as the only developer. On the beginning, you probably don't have a lot of skype calls to make, or meetings to participate. On this phase, the thing that probably require most of your time is build your product.

And when you finally launch it, if you're lucky, you'll be more busy, but probably you'll also have more money, which will allow you to hire someone to help you. So you'll won't code so much as before, but if you want, you'll still have some time to do it.

The main argument against the article are the fact there's companies managed by developers who still code doing a good service to their customers: Instapaper (as far I know, Marco Arment is the only developer), 37signals (they have a lot of developers, but Jason Fried still codes), RubyMotion, among others.

Maybe this is not the easiest way, but it's definitely possible.

I disagree since a few years ago. Before i was a developer, or better said, a hacker. I wanted to solve problems. Right now, I more or less know hot to solve problems, so develop a product is kind of boring. From time to time something new comes, but is not usual.

Problem => Analyze => Solve (if makes sense to solve the problem).

After years, that is kind of repetitive.

So move up one level is thinking on products, and hacking in a more social level. Learn about usuability, psychology... how the people behave and what the people want. That are things that we the developers should know to be... Better developers.

Hacking is about understanding. Understand how the things in live work. Understanding why a product A success, and a product B fails.

So I will say that founders should be hackers.

What's even worse is when one of the founders can't spend time coding any more due to investor meetings and other concerns about running the business, but still thinks they know what's going on in the codebase, and demands to be able to micromanage it. I know one CTO who fought tooth and nail with multiple meetings to keep AAC compressed audio out of a codebase after agreeing to a plan to improve overall app size, not understanding that the AAC+ compatibility problems in some versions of Android didn't apply. He just didn't understand audio encoding, didn't have time to research it, but he be damned if he didn't get to decide what's used.

Just change the title to "Why CEOs shouldn't be developers". As soon as your startup has a team of people instead of just you and your buddy you can start applying this rule.

Balance is the key to any startup, the opposite side of the coin are founders who have never been in the developer's chair, which is both common and worse.

the opposite side of the coin are founders who have never been in the developer's chair, which is both common and worse.

I disagree. I've worked at a couple of startups founded or co-founded by a guy with no hands on technical experience and he was great. He came up with the ideas, found the right people (or sometimes the other way around) and then let them do what they did best, with only gentle prodding to keep everything going in the right direction.

What he had was a good high level understanding of the relevant fields, plenty of experience starting and running companies, a vast contact network, an amazing knack for raising money and was very good at getting out there and selling our product.

Same situation here.

Am I correct in assuming the following characteristics? The guy you are talking about:

* Trusts you to do your job, and assumes you are pulling your weight in whatever way you see fit. (contrast: "We need your ass in the chair 08:00-16:00 so we know you're working.")

* Trusts you to do your job, and accepts your technical decisions even if he doesn't fully understand them. (contrast: "I just don't see exactly why GitHub would be better than our existing Team Foundation license, so we won't switch.")

Agreed. You need the perspective if you're running a software startup.

No, you don't. I can't code my ass from my elbow, but I know how the technology works, I know where to find the kind of technology I want in my product, and I can find the people who can put it together.

I guess the distinction I'm trying to put is that there's a distinction between "non-technical" and "non-programmer". I can't program, but I know all the technically pertinent details of my product that any VC or customer is going to care about. To some extent, I think it frankly works better that way than if I was doing the coding, because I know what's really important in the product, and won't miss the forest for the trees.

You don't need to know how to code to be able to handle the technical responsibilities of actually running a tech startup. You just need to know the technology itself.

Are you profitable? Can you live off that profit?

Not yet, but that's probably because we just started our market entry process about two weeks ago. That said, our operating costs are tiny (all combined they're currently less than my rent) so we'll hopefully get to that level in the near future.

As a founder and programmer, I found this post really interesting and well written, but rather than lamenting about how it must be impossible to do both roles where eventually you will have to give up programming - why not just accept that it is just much harder? Also, I certainly hope that I will never spend 50% of my time pitching..

How about finding a co-founder to handle those things, while you handle development and architecture?

I can relate to this from working as the CTO of a company that grew too fast, and then burnt out. I like building things, and I found myself less focussed due to management and business concerns. I'm now working on other projects, with fewer management concerns, and much happier (and being paid more).

However, I think the inverse of the post's title is also true. Your founder (at least one of them), if you are doing a tech startup, should be technically competent and understand the aspects of software development that are important... including not changing design specifications every week, and knowing how to interact with clients to prevent them from growing a project's scope beyond what your small team can achieve in a reasonable time.

What I picked from it besides the main subject is this:

> "A while back it was PHP in favour of static HTML, then Python in favour of PHP. Lately it’s been Ruby on Rails or node.js in favour of Python … in the future, who knows. I hear Scala is becoming very popular"

As a Scala fan I could't be happier if it's indeed going to pick up for web dev the way Ruby and Python did (in a myriad of frameworks), but is this indeed the general sentiment nowadays? I'll be more than happy if Play framework gets some of the Rails and Django community to help it grow the way the latter two have in recent years.

I'd like to comment that the title should probably be "Why founders shouldn't be developing/programming". This title makes it seem like being a developer excludes you from being a founder/CEO

This is true for another reason. Programmers make horrible designers, and they should stick to programming when the product is user facing. They can forge perfect tools, and sometimes just the tool is valuable enough to make a profitable product. Many times though, the product needs to work for people, and giving the executive decision powers in the hands of most of the developers is a recipe for a disastrous product.

- Programmers make horrible designers

Please don't generalize, this is not always true. Being a programmer (with a formal CS background) who used to make a living as a professional designer, I have met a lot of people like me. Good design is about solving problems, good programmers usually solve problems. When it comes to aesthetics, it's something that people do learn, and programmers can learn, too.

- giving the executive decision powers in the hands of most of the developers is a recipe for a disastrous product

I respectfully disagree here too, for the sole reason that developers can learn the ability of making decisions in a relatively rational way. I would even wager to say they might be more inclined to trust in data.

I should have started with "Most programmers make horrible designers"; my apologies for that. The point is, even making a not-so-bad design, takes a lot of thought; more importantly, it takes courage to push a design concern that will make implementation harder and take longer to finish. I certainly believe that people are amazing, and they are capable of doing amazing things. But also there is the issue of possible vs probable. And, I don't take it lightly when I see a good programmer and designer in the same person.

You're right, because no successful tech companies have ever been started by programmers, with programmer-designed UIs.

I have given room for that. When the product is a tool, then you can't do much, but accept. The experience is usually similar to the one you have if you ask a blacksmith to make you a pair of scissors.

"Programmers make horrible designers..."

Correction: Most programmers make terrible designers. I've met and worked with some that can handle both well and a few that are exceptional in many realms technical and design.

Now the "designer" that glued the Metro interface on to Windows 8 should stick to smartphones. I'm about to write shim to make the OS boot into a CLI interface, so I can avoid the "Tiles" and get somework done.

I pretty much agree. Although I think it's important to have the tech background and have the ability to code (the same ways CEOs need to understand basic accounting) it's hard not to get immersed in the rat race and lose sight of the big picture. As the OP said, you need to spend a big part of your time pitching. Both to investors and as importantly to potential recuits.

Comment bait title aside... That's why you do it in steps: you build the bulk of your product first before you have customers and need investment and then you shift your time to other things. Nobody is going to hire an engineer to free themselves up to pitch for money to pay for the engineers until that is a sound decision.

That's sort of the obvious thing, but on the other side you have people that you can't do any really meaningful product development without customers, or you're likely to build something nobody wants. Is there a good balance there?

Consider the halfway house of paying freelancers to pick up bits of work for you.

For example: I suck at the whole CSS/HTML/JS merry-go-round. Ironic that I mostly work on web apps and that these days that is basically where the app lives. But still.

So I get help. I know some good people; I pay them to fill in my gaps.

What I get from this article and from the Git-commit list (upvoted!) is that founders shouldn't REMAIN developers. Having a coder-founder seems almost fundamental to a tech start-up, no?

Yeah, sure, say that to Facebook & Twitter founders (to name the most known). People can sometimes have sufficient genius to handle creating then founding if the product is valuable.

this seems like it assumes that there is only one founder. this is one of the magical moments where you realize how amazing a cofounder can be. at my new startup (http://matchist.com) im the "nontechnical" cofounder and handle all of the stuff this post mentions founders should do so my cofounder can code. and code. and code.

having a cofounder (especially one with a complementary skillset) makes it so much less frustrating in the beginning phases of a startup

Well, true.

But hence the notion of not being a founder CEO. Let someone else take the reigns and you do whatever you want to do.

Though there exists a lot of potential of messiness in this.

I stopped reading after this, "A while back it was PHP in favour of static HTML." Actually it was PHP in favor of either Perl or C used in CGI, or VB in ASP.

One of the tangental points you raised was wrong so you are wrong!

As a non-developer solo founder, this article resonates with me more to the point of "as a founder, you can't expect to devote all your energies to any one thing". My responsibilities with my startup have varied pretty much by month.

First it was working with potential customers to figure out what kind of product they wanted. Then it was time to find the technologies that would allow me to create that product. Then it was finding a technical staff that could build our product using those technologies. Then it was preparing our sales strategy, materials, and sales staff to sell the product. Now it's primarily marketing, pitching to writers, blogs, and doing what I can to get the word out to consumers. If I had to forecast the next month or two, I'd say that preparing investor pitches and materials will start to absorb an increasing amount of my time, and I will probably end up finding a marketer to take over my marketing responsibilities so I can focus on the funding side.

Each of those stages basically involved devoting myself overwhelmingly to that particular issue that was most crucial to the business at that time (product development, sales, marketing, etc) then finding someone to delegate that work to once I needed to move on to the next step. Frankly, I'd never want to do it any other way, because it means I know every aspect of my business, my product, and my customers, and I know exactly what I need to delegate to make sure things get done right.

Moreover, by knowing every aspect of my business, no one is indispensable to me. When my original programmer decided he had to take a full time/weekly paycheck job for financial reasons, I knew my product and technologies well enough that I was able to seamlessly bring in a replacement to finish the work because even though it wasn't my code, it was my product so I could specify exactly what had to be finished, and what he'd be working on after it was done.

Last point, because it seems to be a widespread misconception here: Programming is only a part of the technical side. My startup uses multiple pieces of technology that my programmers weren't aware existed until I showed the APIs to them. Now, I couldn't have used those technologies without those programmers, because I lack the ability to actually take the APIs and plug them into our code, but I can find the technology that does what I need done, and then have them integrate it. If on the other hand I'd brought one of those developers on as a "technical co-founder", and relied on them to handle the technical side, we would have been dead in the water or would have ended up with an inferior product.

None of this is intended to be antagonistic, or to diminish the importance of programming to a tech startup. I simply state it to show there is more than one way to bake a pie, and that a non-programmer can still handle all the technical aspects of a company needed to create a successful startup. Food for thought!

That's why I'm so happy I have found a great cofounder so I can concentrate on development.

Awesome. Great description of being a developer and of being a founder.

This seems to imply side projects should never become startups.

Amen. Thank you for putting it in such a concise clear manner. Those who think they can do both will end up insane. I've seen it too much and so has everyone who has been in this industry long enough.

Insane. I'd say this goes for VP of Engineering and CTO's as well, whether founders or not. Really any management.

You chose it, but can you do it? Most can't and thus abuse those under them. Sad but true.

EDIT: I will add, that if you're trying to use the latest, coolest technologies as a founder you should just stop now and choose a different career. If you didn't "get it" (proper development) by now, you never will.

Deleted comment

That wasn't what the article was about - it was that once you become a founder, with other people depending upon you, you don't have time to develop any more, not that you can't become a founder if you're a developer. I've heard Zuck still does code occasionally, but I'd really doubt he spends more than about 1% of his time doing it.

Enjoyed this post :D

Reply on news.ycombinator.com