How Valve hires, how it fires, and how much it pays
gamasutra.com
gamasutra.com
> all is good when the money is flowing.
That really struck me. Everything is all good when the money is flowing. It doesn't matter whether you work at a bank, a law firm, a defence contractor or a sales agency - no matter what the structure, or how people are organized, everything works when the money is flowing.
Now, what I'm really interested in is catastrophic failure.
What goes wrong when the money stops flowing? From what I've seen historically - the same things go wrong no matter what the place (see GFC effect at various companies back in '08).
How is Valve going to deal with the "always losing money" proposition?
There's a mixup in correlation and causation here. Similar mixed up conclusions are reached by people looking at other fantastically profitable companies.
Do you like scenario planning? Shell "proves" it works.
Stalinist management? Apple "proves" it works.
Velvet sweatshop? Microsoft "proves" it works.
Data über alles? Google "proves" it works.
Self-directed workplace? Valve "proves" it works.
Microsoft is the only company I've heard of that has (or had) "Morale Budgets", completely within the control of the programmers. In the 90s, according to McConnell, Microsoft would go to any lengths to protect and improve morale.
I can't speak for Shell or Apple.
The point is that people are mixing correlation and causation. There's no reason to think that Valve's self-directed model is the cause of their success, just as there's no reason to think Google are successful because of how they manage their people. In all these cases the massive profitability is pretty much exogenous.
I also stand by what I said about Google and Microsoft. I know people give Michael Church a hard time here, but from my own friends I've heard anecdotes that confirm his sentiments regarding the amount of bullshit that goes down there. Sure, it's a good company and people are happy - as long as they play the politics game. I'm sure a company-wide "index of satisfaction" looks good on the HR handbook, but when your manager blacklists you without your knowledge and prevents your transfer or promotion, that's kind of shitty.
Same thing with Microsoft, which is known for its legendary corporate bureaucracy and political in-fighting. And Apple, which is by all accounts a very high-stress environment to work in.
I don't think communes are as rigorous about admission.
That or the people that have niched themselves into positions of unwritten-but-assumed authority have decided to hire quiet followers.
Just because someone calls an environment a 'flat system' doesn't make it true. There exists just as likely, a very defined system controlled by seniority.
An even bigger difference might be that the financial goals Valve has as a community are more clear than the goal of "build a good anarcho-syndicalist commune". In my experience communities that have a goal to pursue that's external to the existence of the community are the ones that last while those that don't tend to descend into high school-like popularity contests and fail because of that.
BTW, is there anything HN would recommend reading on 1960s-1970s communes and, especially, ways in which they failed?
Like in "The beatings will continue until morale improves"?
Microsoft has, of course, changed. But in the 90s it was probably the best place in the industry to actually work, if the articles and books written about that period are even vaguely accurate.
The interesting thing is that Valve has similar open space areas. You could substituate a photo of one for another. The big difference at Valve is that people have /self selected/ into the other people they are working with, and can move any time.
The ability to choose is worth any likely amount of "morale budget."
It's true that there are a few disgruntled ex-Googlers that may give you this impression.
But I don't believe you'll find many current Google employes who will agree with that.
Multi-billion dollar game franchises have been created in other ways before and since the initial Cabal article was published.
My hobby is Olympic-style weightlifting. One thing that happens a lot in my sport is cargo-culting whichever country happens to be dominant at the moment. Back in the 1990s everyone got very excited about the "Bulgarian" training system, which was quite different from the "Russian" system which had dominated from the 60s.
In the USA in particular, wholesale attempts were made to adopt the Bulgarian system for elite athletes. Results? Very disappointing.
The context is wildly different. Bulgarian coaches had a feedstock of hundreds of thousands of lifters. If a lifter was wrecked by the extremely aggressive Bulgarian method, so what? You just replaced him with another lifter, there were plenty more coming up the pipeline. Given that in a larger sample you can find more outliers, mere numbers predicted a large fraction of the Bulgarian success -- and before that the Russian success.
Mere numbers today predict the success of China in the lighter divisions; mere numbers in future will predict that China will steadily improve in the heavier divisions as Chinese youth become taller due to westernised diets with more protein and calcium.
But the cargo culting has begun. Weightlifters already talk about "the Chinese System" as if there was some single, monolithic master plan. There isn't really. There's just a metric shit ton of Chinese weightlifters in the lower leagues and the elite international coaches can pick out the best of the best.
For the same reason, New Zealand is rugby superpower, nobody can beat the USA at gridiron, Australia is barred from entering the international Australian Rules Football contest because even our lowliest semi-pros dominate any such match up ...
Your assumption that the Bulgarian coaches have a feedstock of hundreds of thousands of lifters is very far from the truth - the whole Bulgarian population is considered to be around 7.3M people, which doesn't account very well for all emigrants living abroad. Out of these 7.3M only 16% (~1M) are aged 0-17 (I assume that 14-15 y/o is the usual age that a lifter starts training professionally). That means that at the moment there are around 100K teenagers (aged 14-15), out of which less than 50K are male.
During the past 20 years sports have lost a big chunk of their (state) funding and therefore the attractiveness for young uneducated kids have lowered substantially, further reducing the pool of athletes.
I consider the three most prestigious and popular sports for males in Bulgaria to be football (soccer), wrestling and weight-lifting.
Meanwhile the USA had, until Crossfit, about 3,000 lifters. Not a very big pool to draw from.
You correctly point out that state funding for Olympic sports collapsed some time back and that this removed the quasi-professional pathway.
For a similar reason, few athletes in the anglosphere who is naturally strong and power will go into weightlifting. They're playing gridiron or rugby. Those sports are popular and have lucrative professional leagues.
I don't know much about weight-lifting, but I recall many scandals in the past years about disqualified Bulgarian athletes for doping.
(I am Bulgarian.)
Unless you're running a multi-million dollar fraud, money is the proof it works.
Proof as in: a lot of other smart people do worse.
Of course, Google, Apple, MS, Valve products are all different, and they would probably be bad at doing the other company's job (to a certain extent)
And yes, MS is monopolistic, is inefficient at work, have several divisions that are a money sink. Still, last I heard, they only lose in profits to oil companies.
Yes, some models require more money to work, still, they got that money in the first place and are free to work like that for as long as it works.
Or worse: an article in the Harvard Business Review.
...do many people think or claim that their success is due to their management style? More importantly, is that why we're talking about it? Certainly that isn't the focus of this Gamasutra summary of the EconTalk podcast by Valve's economist.
Valve does talk a lot about their org-style in general, and they do talk a lot about their success. But I'm pretty sure they have a sound, mundane reason for those things: talent sourcing.
Standing out in the tech industry as an employer is tough, and the traditional offerings: "smart people working on interesting problems", "we're growing fast", "we're the market leader" are, excuse the expression, tantamount to banal rape. Valve has a further difficulty though, as expressed by Varoufakis:
"In many occasions people simply don't fit in not because they're not productive or good people, but because they just can't function very well in a boss-less environment."
They need to find talent like everyone else, but beyond that, they need to find talent that won't fail without someone taller telling them what to work on. Valve's sane solution to both problems (lack of talent, lack of talent-preparedness for their org-style) is to get loud about their org-style.
Valve's org-style is so wild and different that it means we could all talk about it until the cows return -- and we do. They routinely make (tech website) headlines just by repeating themselves, which draws crowds. It also causes candidates to self-select, lessening the fit-problem. Don't think you'd like to work at a Place Like Valve? You won't apply. Never thought about this org-style before? You will now.
Why do we love to hear about their org-structure so much? Is it just because it's different, or perhaps because of apparent claims it's more profitable? I think it's because it addresses a real problem. It would appear that silly directives-from-on-high don't exist at Valve, because they internally removed the notion of "on-high". They also claim to have given each employee the autonomy to figure out how and where to do their best work, while being paid enough. It sounds like they struck the creative-work motivators of autonomy, mastery, and purpose[1] on the noggin. Good marketing at the least.
[1] RSA Animate's adaptation of Dan Pink's talk http://www.youtube.com/watch?v=u6XAPnuFjJc
And in some places? Yeah. Probably legit.
But I the thing is that it's impossible to know if the Valve model is worth studying for clues on how to better manage non-lottery-fountain companies, because the confounding factor of riding a money mississippi is basically totally impossible to factor out.
In conclusion, I'm jealous that I'll never have to face this problem up close.
The more common corporate arrangement is the extortionist command economy in which one serves one's immediate superior or gets fired from a whole company. That is ridiculous and pathological. It might have worked in 1870, but it's starting to fail badly. Billions of dollars of value are being lost due to this outmoded way of doing things.
Valve is clearly not perfect-- they had a layoff earlier this month-- but it appears that they're still doing a better job, on a cultural front, than almost anyone else.
I do think it's important that Valve operate as they have (in being loud about their structure), to point the water out to us fish. It's commonplace to think that work is just the way work is, with managers and workers and you get a job then stay in it until you are promoted or fired or quit.
Valve is doing something different -- and of course it won't be perfect. But it's different, and they're shouting at the top of their corporate lungs about it. Hopefully it will help lead to the development of a spectrum of organizational styles. Even if it just brings some common willingness to mess about with the water and see where we go, that would be great.
Terrible UI? Craigslist.
We tend to look at outliers and make them the canonical example. 37signals has been particularly bad with dispensing startup advice being completely not typical.
Don't we have examples of financially successful projects with various problems? A growing number of employees may bring in communication inefficiency, for example. Now this probably won't undermine company success, and from the point of people at the top it may be all good regardless.
For other values of ‘all good’, though, I'd argue structure and organization matter.
I would guess, not well. The minute that they start losing money and are forced to seek investment/debt/rescue, is the minute that their current org-structure dies.
I don't see how the organization itself costs money, unlike other companies who go for physical perks (food, massages, fancy offices, toys...).
BTW I just wanted to add that despite all the talk about current Valve's efficiency with the old Valve there was only 5-6 years between HL1 and HL2, while its been 6 years since HL2EP2, and 9 years since HL2.
In comparison since HL2 there were Halo 2, 3, ODST, Reach and 4. Id launched Doom 3, Quake 4 and Rage. And we had Bioshock 1, 2 and 3, and Gearbox made both Borderlands games.
Valve is way too focused in catching up to F2P IMHO, and it should be careful since the social gaming party is already over and I've seen more than a couple of F2P games that get no "whales" at all, if you catch my drift.
Original content: http://www.econtalk.org/archives/2013/02/varoufakis_on_v.htm...
Yanis Varoufakis is the economist valve hired to study the economy of steam and valve itself.
Seems worthwhile to mention the (leaked?) Valve Handbook. It explains a lot of this stuff: http://newcdn.flamehaus.com/Valve_Handbook_LowRes.pdf
Steam is the most critical part of the company for their future.
This means people there don't do unglamorous things like maintenance and stability fixes.
In practice, makers are different people from polishers, and Valve's system of hiring the "best" (people with impressive CVs) and salaries based on stack ranking means there is a huge bias towards makers (Who brought the company more money? the person who helped roll out dota2 or the person who fixed edge cases in the steam client?).
Many of Valve's products are full of bugs that don't get fixed for years. And since they only hire the "best", they don't have any testers. Many times patches are released that cause servers to crash within seconds (tf2 falling damage crash). And sometimes they release them on Friday which means it doesn't get fixed until Monday (Overflow error writing string table baseline crash). These patches were also mandatory.
Edit: though, to contradict myself, this is still as gameable as all hell if you do something silly like connecting it to reward and punishment.
The only thing that works is deliberately cultivating a culture where driving out the old bads is as important as creating the new goods.
Put another way: for self-direction to work, the culture needs to rank improvements as highly as novelty. Possibly higher.
Which it never does. At least, I have yet to experience or hear of a (software) engineering culture in which that's true.
I'm not sure if cultures like Toyota or the Nuclear Navy can be considered under this heading, as they mostly worked at a procedural-improvement level, rather than a product-improvement level.
We're into No True Scotsman territory here.
GOOD programmers know a lot of things. Put them in an environment that saps their initiative and they do 1 of 3 things: Leave, succumb or adapt.
This means people there don't do unglamorous things like maintenance and stability fixes.
How to create an incentive for this kind of work is an interesting problem, and I don't know enough about Valve to know if they've solved it.
I will say that the alternative is more dysfunctional. Closed-allocation, corporate alternative: some number of people are staffed on the ugly, career-damaging maintenance projects. The good ones either find a way to play the politics and move, or they quit, the bad ones stay. The end result is that the maintenance work is done by incompetents who don't care. This is a big part of why most legacy code only gets worse over time: the maintenance work is given to people who don't have the clout to do anything else, not to people who care enough about the health of the project to do it well.
Edit: I forgot the Sales. Those damn Steam Sales get me every time.
As for the DRM... has it ever really been onerous? The only case I think of is if you don't always have access to an internet connection, and they have workarounds for that, though they admittedly are imperfect.
True, the hypothetical/potential failure modes suck - being locked out of your account and losing your games; Valve goes out of business and you lose your games; you're offline and can't switch to offline mode and thus can't play your games until you find an internet connection. That said, I've always been of the impression that most people don't mind Steam's DRM that much because doesn't stop them from doing what they generally expect to be able to do and want to do with their PC games, which is sit down at any PC where they can log in on Steam and play them. This is unlike DRMed music and movie files: sure, DRM prevents you from making unlimited free copies for your friends, which I think most anyone would agree is reasonable, but it also can prevent you from doing things like changing formats or resolutions or storing/viewing from certain devices.
http://autodesk.blogs.com/between_the_lines/2013/01/get-free...
I think the interesting part is that in spite of that past behavior and almost no PR, the gaming community just rallies around Valve and will vilify most other publishers that attempt to follow Valve's footsteps with Steam.
No one I know found the always-online aspects of Steam particularly onerous. Especially because when it was introduced, it was mostly used for games that were only interesting online anyway. HL mods, HL2 and mods. That "no one I know" is just anecdata, I admit it. And it integrates features that are actually features. E.g. Steam friends-chat is actually useful, unlike all other inter-game chat I've seen (on the PC).
And the Steam store is such a boon. It makes me want to give game-devs money. As my disposable income has grown (massively), Steam was in the right place at the right time. Brick and mortar game stores don't exactly inspire loyalty. And I like the fact that digital distribution lowers intermediate costs, and I like the fact that I get a share of that savings with the crazy-deep-discount steam store sales.
On the other hand, I can't remember the last time I interacted with DRM from another publisher without being afraid, and without paying a burden that actually was onerous. It messes with my router, or it turns out to be an exploitable rootkit, or it pops up inexplicable warnings, or it eats RAM, or I have to create an online account for one stupid game that I don't even know if I'll like yet, or...
I agree with nlawalker.... getting us acclimatized as a precondition of playing Must Play games (online HL mods after 2003, HL2, Portal) was brilliant.
Of course, if Valve in their wisdom were to decide my account was violating TOS, I'd be pretty boned.
Steam's DRM is only offensive in theory. In practice, it works perfectly well for the vast majority of the users for the vast majority of the time.
a) it doesn't affect the consumption experience too negatively. b) the platforms pros (e.g. ease of distribution) outweigh the cons (i.e. DRM)
I can't remember the last time I cursed Steam's DRM system, honestly. Most of the time I don't even think about it.
I used to pirate games left, right, and center. I would occasionally buy a new one if it had a good rep or I liked the series, but the default mode was 'pirate'. Since the A500 days (yo-ho, yo-ho, a pirate's life for me!). It also helped that the Australian software market is rapaciously overpriced.
Steam changed that. Steam didn't get rid of the DRM, but it added convenience like never before seen. Almost all the problems with DRM affecting the legitimate user are gone - I've never encountered a 'licensing server offline' or an 'invalid code' and only one 'sorry, licensing too busy, try again later'. I've rarely had a 'failed to install', and when I have, a 'try again' pretty much worked. The same can't be said of my experience with boxed software or pirated software. Updates are relatively seamless - Steam is acting like a package manager here, avoiding the need for everything to run its own updater. It's also a lot easier to check on any sales steam might be running - not so easy with a brick-and-mortar store(s); rare is the store website that shows the right level of stock and is regularly updated with specials.
Now, to be fair, I didn't start with steam at the inception, I've only been on for four or five years (it did get installed when HL2 came out, but I never got the sequels so didn't bother with it for a while), and when I started pirating I was a student and when I started steam I was working full-time, but even so, the convenience of being able to buy software from a trusted source and install it without having to leave the room made all the difference.
Steam does have its problems - the client is clunky (and has no tabs, grr - it's built on IE) and sometimes games can surprise you if your internet is out and they are 'online required' (which is usually invisible otherwise).
One argument that I noticed changed over the many years is "But online DRM means I won't own the game if I want to play it in X years". I've played a lot of games over the years. I return to very few - once a game is played out, it's gotta be pretty extraordinary to make it to my replay list. I still think DRM sucks, but it's largely irrelevant once the game is done.
Nope, they switched to Webkit a couple years ago. [1] (Still no tabs though..)
>I still think DRM sucks, but it's largely irrelevant once the game is done.
Wouldn't the success of something like GOG [2] kind of prove that there's a market and a reason for DRM-free games? I all but guarantee 20 years down the road, we'll still be playing games from this generation.
Sure, absolutely. I was talking in terms of my own experience with that line. I'd prefer there were no DRM, and as makomk says above, it causes problems to others that I don't see. I just meant that that one particular argument of "no ownership in X years" didn't hold much water with me personally.
This being said, I have bought System Shock 2 twice and lost it both times (once stolen, I think) and today I have been playing a pirated version (hooray for the new texture mods). That game is one of the exceptions :)
But, selfishly, if I could buy every GOG game on Steam, GOG wouldn't get a penny of my money. I value the Steam features more than I value the DRM-free-ness. I'm befuddled that this is my position, but it is.
Considering Valve Corp as a whole supposedly has a total of only about 400 employees and is acting as a developer house, publisher, and 3rd party distribution system rolled into one, I think they've got a pretty fantastic churn rate.
EDIT: Just want to say I wasn't the one who downvoted you.
I know what you mean, but apart from Portal 2 the rest does not really qualify from development from scratch (even Portal 2 is dubious on that aspect, since it takes all the mecanisms from the previous games and expands on it). While I love what they do as a distributor and very much appreciate what they do for Linux support, as a pure game developer they fall short of expectations, and they have not done anything very original for a while. For a company with a "flat organization" that is somehow disappointing.
Doubt this link is remotely accurate.
In general, if there are only 1 or 2 reported salaries on Glassdoor, you want to take them with a significant pinch of salt.
Edit: found the talk and the moment, and it wasn't quite as I remember http://www.youtube.com/watch?feature=player_embedded&v=t... Rather he was talking generically about employees getting paid something that reflects the value they bring.
This works fine for me. The javascript reads it I suppose.
But they do have chief writers, as I guess should be obvious http://www.gamasutra.com/view/feature/132539/valves_writers_...
http://www.gamasutra.com/view/news/186592/Several_out_of_wor...
http://www.theverge.com/2013/2/13/3983348/valve-fires-jeri-e...
The Steam you use today is to that version of Steam as IE10 is to IE3, approximately. It was lacking major features -- working chat, overlay, store, non-steam game support, offline mode, authentication that didn't fail 70% of the time forcing you to restart Steam then try to play your game again, etc.
It's useful and a value add these days, it was just a gigantic nightmare before.
But I'm still very disappointed that the best (only? I don't count Origin) tool like Steam is Steam. By that I mean that we're locked into what Valve wants to do, which is sell us limited licenses that are tied to Valve maintaining their service, significant amounts of DRM, and other anti-consumer policies like no refunds. Some of these are deal breakers that leave significant room for a competitor to eat all of Valve's lunch, so I suspect if Valve doesn't buck up and play nice with its customers that eventually someone will.
But I simply can't trust anything until I start hearing about the negative aspects too, which nobody ever seems to report.
I mean, not everything can be all roses over there. I'm tremendously interested in the way their workplace actually runs, but you can't really learn any lessons or apply it anywhere else without having a fuller picture of what really goes on, warts and all. I mean, no system is perfect, or even close to it.
"Valve pays people very well compared to industry norms. Our profitability per employee is higher than that of Google or Amazon or Microsoft, and we believe strongly that the right thing to do in that case is to put a maximum amount of money back into each employee’s pocket."
Considering that a ten year veteran at Google can make $200k/year, in salary + bonus + stocks, I'm curious if Valve compensation can reach the same level. The Glassdoor page doesn't suggest it can, but people may not be reporting their bonuses or profit sharing.
What if I keep doing these kinds of projects? Will I eventually lose the popularity contest and be ousted?
The amount of money to be made in personal robotics or space robotics probably dwarfs what Valve makes now, but research is risky.
The combination of employee selection and social pressure means that it's probably not really true that you can work on whatever innovative ideas you want to.
This is what irritates me about companies co-opting hobby time and free time. It's not really free time if it's a popularity contest. You don't have to justify your hobbies to anyone (family, maybe).
I work at a company where I have tremendous freedom to organize my own work, and to work on research projects, but there are still areas that are clearly out of scope. I imagine Valve is no different.
Arbitrary decisions from an executive may or may not be better than popularity contests. I guess it depends on the executives and the culture.
I'm just a bit cynical of all the "you can work on whatever you want (that's good for the company -- that others around you agree is good for the company -- that's meant to capture innovation -- but that can't really capture long-term or unpopular innovation)."
So yeah, the way I work at my current job, and the way that people supposedly work at Valve seems great. That said, what I really want is to not have to work, but to work anyway. Woe is me.
Oh, so you recently realized 25 people didn't function very well in a boss-less environment?
In order to function well in a boss-less environment, it helps to have grown up being educated in one. If your education is authority directed, it's natural to expect your adult work be authority directed. It seems to me that Valve's culture is particularly well matched to people who came from democratic schools. You can probably train yourself to be that kind of self-motivated person as an adult but it isn't easy.
To make something like this work, you can't just hire anybody. It won't work for many organizations simply because they have the wrong people with the wrong ideas of how to run a business or organization.
It's also a lot of work to maintain such a system and the wrong hires can ruin it fast.
The design of a company appeals to a certain kind of people, regardless of the model. Do you put a premium on autonomy? Work at Valve! Like structured environments? Try Microsoft. Prefer strong leadership? What about Apple?
Or whatever.
Success is not the take-away. Employee-model fit is.
http://news.ycombinator.com/item?id=5182145
Is there any other companies like Valve and Github that are bossless?
Sr Software Engineer - 97,000 (2 listed at this rate)
Principal Digital Media Engineer - 62,608
Multi-media artist - 77,000
And for fun, some Google employees:
Research Engineer - 206,900
Software Engineer - 105,000
Software Engineer - 214,800
Site Reliability Engineer - 111,600
Site Reliability Engineer - 170,000
Program Manager - 152,000
Information Security Engineer - 150,000
> The payment mechanism is to a very large extent bonus-based ... Bonuses can end up being 5, 6, 10 times the level of the basic wage.