|Notify me of new responses|
I read with interest Managing Software Engineers (http://philip.greenspun.com/ancient- history/managing-software-engineers). The recipe is high standards, competent
coworkers and long hours, which hours are encouraged by providing a nice office with
many perks. What are your thoughts on financial incentives for programmers?
A typical venture funded start-up today offers engineers meaningless equity
grants/options. 0.1% in a post Series A company is not unusual unless you're CTO or
VP of Engineering. They may or may not pay a competitive salary, but even if the salary
is competitive, why would programmers want to work long hours when they don't get
to meaningfully share in the upside? It seems they do it because they're young and
naive, they overestimate the probability of a successful exit, they overestimate what
their share of the upside might be. Often they don't even bother investigating these
Can you share what kind of equity ArsDigita programmers had? Have you seen
meaningful equity grants for programmers? What about profit sharing? Perhaps I'm
overestimating how much of the value of the company programmers create? Should
programmer just be happy with a good salary and not expect meaningful equity
The perks are nice, but it's just not enough to motivate me as a programmer. I'd rather
spend more time on my hobbies or side projects than go above and beyond for
founders and VCs. As a potential future founder, who is VC averse, I ask myself what is
the right way to compensate programmers?
-- Ivan Govrilov, December 22, 2015
I wouldn't want to use ArsDigita as an equity example for anything except "Don't take VC money with its standard 'preferred shareholders get the first $X' if the company is already successful." The problem is also severe in today's heavily funded companies. If they raise $50 million to do something simple and the company ultimately sells for $50 million or less, the common shareholders such as employees (including founders) get nothing. It has been so long that I can't even remember the percentage of common that ArsDigita folks had as options depending on when they joined. It turned out to be moot because the company wasn't ultimately worth more than the preferred investors put in so the options weren't worth exercising.
We did have profit sharing in the form of most of the profit being paid out at the end of the year as bonuses. This avoided us having to pay the world's highest corporate tax rates. It also rewarded people for delivering to customers. Anyway, people could earn 2X their (relatively modest) base salary by getting a bonus. (As the CEO, I made myself ineligible for a bonus, unfortunately!)
I think the idea of stock compensation isn't a bad one. At least it keeps everyone aligned on creating value.
Most programmers these days should probably work for Facebook, Google, et al., where their efforts have the most value. These companies give stock grants that are easy to value and their good programmers can earn over $300,000 per year. (Or maybe become friends with Marissa Mayer and get paid well at Yahoo even if you're not doing a good job...)
If you're compensating programmers how should you do it? It depends a bit on the kind of company. If you have revenue and customers you want everyone to focus on getting more revenue and making the customers happy. So a bonus that is related to customer payments is probably good. If it is a pure conceptual startup and what you want people to do is create a working product maybe a couple of structured bonuses. One for the product passing an acceptance test. One for a customer paying for the product. Then the usual stock options in the company.
As income tax rates rise, I do think that stock options, which can qualify for long-term capital gains tax treatment, will come back into style.
-- Philip Greenspun, January 11, 2016
Ivan, I am also curious what Philip's thoughts on that are.
Note that ArsDigita (particularly the Cambridge office, which was the largest) recruited heavily from MIT undergraduates who'd taken Philip's class, and from its free 3 to 6 week "bootcamps" (genius), so they often had an unusually clear idea of a programmer's practical skills and motivation before hiring. Also, the ACS toolkit was such that new hires could be productive on client projects within only a week or two. I suspect Philip's policies were optimized for those circumstances. With engineers averaging older or much more complicated projects, you might want something different. But at least for a while there, when there was plenty of business coming in, Philip's model seemed to work quite well.
When I was an ArsDigita employee, I was perhaps somewhat young and somewhat naive, but I happened to have plenty of reason to work really hard despite having negligible equity upside. One, I'd just switched careers - that was my first professional programming job (vs. being an engineer who coded on the side), so there was a ton to learn and get better at. And since both the pay and work environment were much better than the engineering job I'd left behind, I knew viscerally what an awesome opportunity it was. Two, I liked working hard and getting stuff done (still do), and I liked it even more when surrounded by other folks gung-ho to do the same. As my friend Paul (also ex-aD) or I said to each other at the time, "It's like a tribe, and I like that about it." I'm not sure if I was representative, but at the time giving me more financial upside would have been largely superfluous, thus a waste from management's point of view.
Btw, when Philip's Managing Software Engineers article was first published, a surprising number of people within ArsDigita seemed to think it was mean and awful and Philip never should have published it. That surprised me. I loved the article's honesty and clarity, and thought it made excellent sense and was a great fit for the company culture. But apparently that culture had already drifted substantially by then. I was lucky to catch the tail end of the meteoric rise, before things flatlined and then started downhill.
That article did make one notable mistake though. The "Physical Comfort" section mentions the need for noise control, but then basically dismisses it. ArsDigita Cambridge had completely open cubicles, not even any partitions - a terrible choice. Despite buying headphones to play classical music, I noticed that I was much more productive early in the morning and late at night, because that was the only time the office was quiet. As far as I know nobody ever collected survey data, but I doubt I was unusual in that respect.
-- Andrew Piskorski, December 23, 2015