Interestingly, Valve's hiring policy is that you should aim to hire someone who is better than you. But what if you hire someone who is better than you, then Valve has a round of redundancies and you're the one who gets the axe?
Here's what Valve's employee handbook [1] says on the matter: '"With the bar this high, would I be hired today?" That’s
a good question. The answer might be no, but that’s actually awesome for us, and we should all celebrate if it’s true
because it means we’re growing correctly. As long as you’re
continuing to be valuable and having fun, it’s a moot
point, really.'
> Interestingly, Valve's hiring policy is that you should aim to hire someone who is better than you. But what if you hire someone who is better than you, then Valve has a round of redundancies and you're the one who gets the axe?
Given what we know about human nature, it probably doesn't work this way. The person who ostensibly worked to hire people better than themselves probably mutters "well, the manual doesn't say anything about firings" and swings the axe, being careful not to cut themselves.
Yeah Valve's company policies are kind of like Santa Claus I really want to believe in them but find it hard.
The whole hiring someone better than yourself idea is great and sounds great. However has been pointed out it may not be the best policy for people especially now that Valve has shown a chink in their armor.
Hiring everyone better than yourself is great for you if you have a nice chunk of equity in the company but when you're just an employee the benefits are much smaller.
In a company with a well known lack of traditional corporate hierarchy, who does the firing? I'm genuinely curious. None of the reports I've read have mentioned it, which leads you to assume Gabe. This would imply an informal hierarchy.
I recall reading a while back that they have a procedure for this. I think they actually decide it through group consensus, so I would imagine the reasons for being fired would have to be pretty real.
That kind of sucks actually... If I was getting fired from a job, knowing that more than half (or whatever percent) of the people felt I was incompetent and should be fired would really rock the friendships I made.
Having a manager fire you gives you a lightning rod to point the feelings at, rather than an entire group of people.
If i was designing it, you'd need to make it focused on the products. Maybe 50% of people feel these products are not 'viable'. If thats the feeling, I think that those fired can walk away feeling ok about the situation (obviously not great).
That's a horrible criteria. Maybe you have a really talented employee who just happens to be working on a doomed product. The right thing to do is to move that employee to a more viable project, not to fire them just because they're working on a dead project.
Firing by group consensus is, I agree, not tenable. This is one point in time where the CEO needs to step up and take a "buck stops here" attitude.
There are generally 3 types of job loss:
1. People who are unethical or obviously toxic. Here, management can step in and fire the person without much blowback. If someone steals, or is a direct threat to your culture (e.g. he starts playing politics and trying to set himself up as the executive empire-builder you don't want) then it's not very controversial.
2. Good employees who run afoul of political nonsense or parochial managerial politics. This is what companies like Valve are trying to avoid, because it poisons the culture.
3. Layoffs for economic reasons.
I assume this is a case of #3. The CEO needs to explain what happened, why it happened, take responsibility for the decision, and assure people that the company will be better after the change than before.
You are spamming the thread (and every other thread that is even tangentially related to open allocation). You've literally made hundreds of posts about open allocation in the last few months. Give other people a chance to talk.
So what? He's contributing reasonably decent comments, who cares if he posts a lot?
Comment sections are essentially infinite, and if the comments are good, I don't see any reason to tell him to stop. However, if the comments are not insightful/interesting, that's another story.
Some of his comments are great, but many just repeat the same talking points. You can't be saying that as long as someone says something insightful they can copy/paste it into as many HN threads as they like, so long as it is on topic.
IIRC, the employee handbook pretty much said that you can work on whatever you want, and choose your own departments. So when doing layoffs, are peer evaluation scores compared companywide (as opposed to just saying, we're letting you go because we're shutting down your department/project).
In the present case, it's a massive layoff and the decision seems to have staid secret internally (https://twitter.com/tom_forsyth/status/301586785899782144), so I'd guess it's a hierarchical decision, with no internal consultation.
This is one time when, even for an anti-hierarchical company, you can't afford to be a democracy. You need a strong leader to communicate the layoff, the reasoning, and the strategic ramifications going forward.
Another element to the art of layoffs is that you get about one shot every 5 years. One large cut can be painful, but it's not damaging to morale if it's well-justified by business or economic concerns. If you do a series of small cuts, that's when you get people paranoid and the alliances and intrigues start to form.
So, I think the best way to do a layoff in a Valve-style corporation is this:
(1) Fair severance, including career support (positive reference, right to represent self as employed). People will find out what kinds of packages people are getting.
(2) CEO communicates the decision to the whole company and takes full responsibility. No mystical "calibration scores" or HR indirections. He gets up and says, "This is what I did, and here's why I had to do it." I'm not talking about traditional hierarchy, but "buck stops here" leadership. That's needed when layoffs are happening.
(3) Unless the people actually are being fired, explain that it is a layoff and that the people will be eligible to re-apply when you start hiring. (Most won't come back, but this lets the rest of the team feel better about the whole thing.)
(4) Err on the side of a larger layoff, because a series of small cuts is more damaging.
(5) Impose a hiring freeze externally, but keep internal mobility as it is. Many companies impose internal hiring freezes during layoffs and that's actually more damaging.
ETA
(6) Make it very clear that these aren't performance-based firings. The worst is when a company tries to depict a layoff as performance-based. Layoffs are understandable, but trying to depict one as a performance-based firing is being a dick.
Given that Abrash is (was?) working on something Occulus Rift like, I wonder if they just decided "Hey we can use O.R. and not bother with this home grown thing." ?
Kotaku reports that Jason Holtman, Valve's director of business development; animator Bay Raitt, and programmer Tom Leonard, are no longer listed in the company directory:
1756 paid games in the windows store as of today, ouch. granted, none of them seem to be AAA titles, but I'd assume they hear footsteps.
I assume this is probably a driving factor behind project green light? While a heavily curated store made a lot of sense when valve launched steam, with their major competitors now os vendors with light curation can valve really afford to turn away indies? It's not like their store is still browsable or doesn't have a lot of crap in it already.
They've since come out and said that Greenlight as a system isn't solving the problems that they wanted it to, and they're going to try and reduce the barriers to getting games on Steam.
My assumption there is that with the saturation of the games market (especially the "casual" variety) means that it's impossible to pick which horses to bet on. Steam doesn't want to be left out as the only missing platform when the next Angry Birds rolls around, and the obvious solution is to do minor quality control but let everything into the storefront.
Affected employees were asked not to speak about specifics, but the impression we get is that these cuts were driven more by company challenges than by individual performance issues
This is a bit surprising, I'd have thought Valve was trucking in money by the boatloads.
It doesn't generate any income by itself, though. Dota2 and TF2 seem more like tools to drag people into Steam and then sell them other publishers' games.
Is Valve doing poorly, as the article suggests? I thought Steam was incredibly profitable for them. Could this be more about canceling projects that weren't going anywhere?
They aren't doing poorly at all, whatever the cause of these departures is it's not money related. Being private they don't release hard figures but Newell has said they make more money per employee than Apple or Google. Steam itself is not expensive to run either.
Preparing for an acquisition of something more in line with their traditional delivery but aimed at more casual gamers? Or planning on building something new along those lines themselves? Threatened by other players in the console space with retaliation?
I'm betting on close work with Sony on the PS4. Valve has been hostile to MS for quite a while, and friendly to Sony. This would include further Steam integration with Playstation consoles, which started with Portal 2.
This change would exclude hardware people, as Sony/AMD are doing that portion of the design, explaining that area of layoffs.
No inside information here, just trying to read the tea leaves, as Sony's PS4 announcement will be on the 20th.
Soo... there's still a lot of console fanboyism going around even now about this generation, and it's hard to be objective about the hardware without getting yelled down by people shrieking "Cell! Cell! Cell!", but, based on the games that have come out, how they compare with the XBox 360, the hype, and all the breakdowns I've read about how the internals of the PS3 work, basically Gaben here is all but objectively correct. The PS3 is a very poorly balanced machine; it is full of bits and pieces that have a lot of power, but it's virtually impossible to actually hook them all up to each other and have them all firing at full power at the same time. All the powerful parts are crippled by lack of bandwidth to move data in or out of them.
None of this looks like bitterness or permanent anger, so if Sony builds a better-engineering PS4, with a focus on real performance over benchmarks and marketing hype, there's nothing in that rant that would preclude working with them. I hope Sony has people internally who understand the PS3 to be less than the sparkling gem of design that the marketing department wanted people to think it was.
My guess is that there have been long festering barriers in the company direction.
Let's lay out the cards:
Character Designer for HL2 and TF2, Hardware Designer, Director of BD, 3xSenior Animators, senior tech dev for HL2 ep1 & 2 and L4D, One of the original Portal devs, Senior QA and an Engine Dev...In a Jack Welch-esque move, Valve fired about 10% of their staff. Welch did this to eliminate the "bottom 10%". How employees are ranked depends on how you want to rank them. There are lots of senior old-timers in this list.
Valve hasn't been exactly putting out lots of their own games recently.
Maybe 1 game per-year? For a company of around 400 with essentially zero effort/free publishing model that's pretty poor.
Portal 2, while an excellent game, was not exactly the kind of AAA bit mover that a Half-Life game (on a new bloody engine) would be. And while it moved quite a few copies (something north of 4 million copies), NPD claims Mortal Kombat overtook it in week two of sales. At $20/copy Portal made about $80m. Not bad, but again, that's about a year's salaries. Let's also not forget that it's an IP that was introduced just a few years prior as a student project! (Halo 3 moved twice that many and Halo 4 has already matched Portal 2)
I may be living in a cave, but CS:GO isn't exactly lighting up the internet, and it's a sequel to a 12 year old game that was also one of their hottest properties. Why so long?
Valve is letting their most valuable IPs, HL and CS languish.
HL2:E2 came out 5 years ago. That's long enough to spin up a new game from scratch for a team of this size. HL2 proper is 8 years old.
Non-windows Steam software hasn't exactly been taking the world by storm.
If the numbers are right, between 25-30 employees, that's ~$5m-$6m/year savings, but against around $80m/yr in salaries is just a haircut.
Source is getting old. There's some engine devs that were cut there. I have a feeling HL3 and a new engine are caught up in sequelitis-ville someplace.
The Steambox is cool, but getting into the hardware console market is HARD. Like Elon Musk balls of molten iron hard. Turns out it's just a PC hooked to a TV with an expensive in-house controller to replace the 360 controllers everybody else is already using and happy with?Plenty of other folks are selling those, why take the business risk?
I feel like there's a pattern emerging, I don't think the decision was because of hard-times, but on frustration with corporate direction. Most of the people let go were pretty senior, killing off the old-timers is a time-honored way to move things forward. Hardware is a low-margin suicide business to get into, better to work with partners that'll take that risk and flood the market with fairly cheap, reasonably powerful consoles built with off-the-shelf parts. Moving to Linux kills off most of their library, and indie games aren't the money makers that Valve needs or wants. The "work on whatever you feel like" system hasn't turned out to provide the productivity or direction they want either.
Predictions:
1) Valve is going to build a new cross-platform engine, or modify an existing one (e.g. id Tech 5 or Unreal Engine). They needed to clean house to focus on this. HL3 will be the showcase for this engine. Anything less than 6-7 million sold across all platforms will be considered a disappointment. It'll quickly be followed by Portal 3 or possibly an all new L4D game built with the new tech. They'll refocus to releasing 2-3 games per year.
2) Valve is going to partner with hardware manufacturers for TV connectable PCs sold with dual stick controllers and perhaps a cross-license with the Oculus Rift as the market differentiator, this will ultimately be disappointing as traditional consoles will be cheaper and have better libraries for console gamers and PC gamers don't want to use a dual-stick controller.
3) Valve's ultra-flat org structure will not survive this.
4) Valve will not enter the mobile game market in any significant way. Their dev cycles are too slow, and he mobile market doesn't generate the kind of revenue they need.
5) Valve will stay independent, not looking to be acquired any time soon.
You're still looking at Valve as a game maker. Are they? Or is their raison d'être to facilitate steam? I don't really see a lot of signals that they are slowing down on the steam-front any time soon, and their game production appears to be as slow as ever. I don't think we can correctly judge what's going on without understanding Valve's self perception, its internal identity.
Exactly. If we're being rational we should look at what's been pulling in the majority of their money and what's projected to be in 5-10 years. I have a feeling steam is at or near the top and will be the top. Enabling others to power consoles on Steam makes sense, producing more video games makes less sense.
Valves games are becoming vehicles that consumers can use to create more games and sell through valve. If/when HL3 sees the light of day, its innovation will be to better emphasize this.
"Maybe 1 game per-year? For a company of around 400 with essentially zero effort/free publishing model that's pretty poor."
One game per year is actually pretty good for studio of Valve's stature, especially if you consider that many of those 400 work mostly, and maybe exclusively, on Steam. It is not uncommon for highly profitable studios to release just a few games each decade, especially if they are independent. So it is important to keep things in perspective. Half-Life 2 was released in 2004. Per Wikipedia, this is the list of games developed by id Software (200+ employees) since 2004:
- Wolfenstein 3D Classic (2009)
- Doom Classic (2009)
- Quake Live (2010)
- Rage HD (2010)
- Rage (2011)
- Doom 3 BFG Edition (2012)
Of those, probably only Rage sold a significant number of copies across all platforms. id Software also published the following games in the same time frame:
- Doom 3: Resurrection of Evil — Nerve Software (2005)
- Quake 4 — Raven Software (2005)
- Doom RPG — Fountainhead Entertainment (2005)
- Orcs & Elves — Fountainhead Entertainment (2006)
Of those, my guess is that only Quake 4, and maybe Wolfenstein, sold in significant amounts.
Since 2004, Valve has shipped HL2:EP1, HL2:EP2, Portal, Team Fortress 2, L4D 1, L4D 2, Portal 2, Alien Swarm and CS:GO. While none was the hit that HL2 was, few studios have achieved that level of success in the same time frame.
Or take a look at Blizzard. They have at least five, maybe ten times as many employees as Valve, but have released "only" five or six games since 2004, most of which were WoW extensions.
My point is that if Valve is not shipping games at a faster pace, it's because they can afford it. Steam is probably printing enough money that they can afford to spend as much time as they want iterating on their engines and games. I would be very surprised if you were right about the motivations for the layoff.
"Valve is going to build a new cross-platform engine [...] HL3 will be the showcase for this engine."
This isn't much of a prediction. HL2 was released on multiple platforms, and it is reasonable to assume that HL3 will follow suit.
It's probably somewhat relevant to mention that id did a layoff of their own a little over a year ago. I don't remember how many, but they let at least a few extremely talented people go.
Why is everyone leaving out Dota 2? Sure it's not technically 'released,' arguably to successfully scale to a point they can open the game up to China, but it's the most popular game on Steam, with a huge following growing every month.
I kind of saw this coming when they announced the Xi3 last month as being a Valve hardware platform. Originally when I saw Jeri Ellsworth hired, and some of the secrecy behind it, I felt they were trialing something out, because I sure didn't expect any sort of hardware to come out of it. If they found a small form-factor device already on the market, what good are the engineers they brought in to develop hardware for them?
Of those people who were in the company directory, none of their specialities seem to point overwhelmingly to this being mostly about hardware. There are some incredibly talented people from a range of fields.
My view: Valve is retreating from building the Steambox in-house.
Gamasutra mentions that the layouts might have affected the Android and Hardware divisions the most. The Valve Employee Handbook states that the ideal employee is t-shaped, which, I assume, might not fit the profile of someone hired to do Steambox engineering and prototyping.
Exciting times. I really didn't expect Valve to do something like this. Will be following developments closely.
Could it be that the people 'let go' are actually leaving to form their own studio, so more like they jumped rather than being pushed. Or a combination thereof.
Its telling that gamasutra can write a paragraph about each one of them - meaning they weren't exactly code monkeys. Valve on CV and some random website knows enough about you to write a paragraph. They'll be fine.
Also, I count 10 not 25. So someones numbers are wrong.
Might want to reread the article... I believe the paragraphs on each person were pulled from the company directory, and not everyone does up on the directory.
This is a really sad day, because Valve is known for its culture and it's very hard for a company to have a layoff at this scale (almost 10%) and keep its culture.
That said, layoffs are a fact of life in business. It seems to me that this is an obvious case of a layoff for economic reasons. I have no idea what those reasons are.
It would be interesting to know how the culture there evolves after this happens.
I wonder how much they thought about a 10% company wide paycut. The "we're all in this together" can (imho) protect a lot of that culture. I can imagine people leaving anyway. But they would leave because they're shutting down specific platform development, not that there's no room for them to contribute.
Layoffs are crushing regardless. I hope for the best for everyone involved.
It doesn't sound like these layoffs were prompted by money problems, so asking people to take paycuts for 'the good of the company' wouldn't solve the problem of there people who no longer fit into the company vision of what they want to be doing.
If everyone agrees the person was less qualified, probably yes. But layoffs often hit people who at least one other person you want to keep thought was good, which hits morale, which in turn also drags down people. Partly it's because the web of connections is complex, and there are lots of reasons one of the remaining employees might have respected the laid-off employee, any one of them enough to instill a bit of a grudge against your decision-making (or even incredulity that you fired the now-fired person before them).
In a way it's the worst with well-run teams that gel. The ideal situation is that everyone works out what they're good at and develop a bit of a rapport. But then if you fire one of the team members, everyone else on the team has a quite first-hand reason to believe the now-fired person was valuable.
I wondered about that, too. That's what HP used to do: pay cuts offset by time off (with top brass taking the cut but working a full week). It seems like it would be better.
I think that's only viable when the reason for the pay cuts can be blamed on forces outside the company (slowing economy, competition from other segments, etc.).
When it's internal, the people who aren't the problem are, rightly, pissed. With the kind of employees Valve has, I'd be surprised if they didn't lose 10% anyway, except they'd be losing their best, not the ones they actually want to go on without.
It seems like it would be really hard to keep up, but I have no specific first-hand knowledge. Can an open-allocation philosophy really work if you can be fired for choosing an allocation that turns out to be the "wrong" one? If that becomes the expectation, suddenly people are back to feeling like they have bosses and assignments they need to fulfill to keep their jobs, only it's even worse because those aren't transparent...
Perhaps more problematic, there's an incentive to try to get yourself into an "indispensable" position that's core to operations. My read of the existing Valve policy is that they wanted you to ignore those kinds of jockeying-for-position considerations and pick something you were passionate about & good at, as long as it was helpful to Valve somehow.
It seems like it would be really hard to keep up, but I have no specific first-hand knowledge.
That's why it's imperative that there be only 1 round of layoffs every 5 years or so.
One round = shit happens, this sucks, but it's impersonal, so let's get back to work.
2+ rounds = people get paranoid, and typical corporate politics set in.
Can an open-allocation philosophy really work if you can be fired for choosing an allocation that turns out to be the "wrong" one?
Open allocation doesn't mean "people work on whatever they want". It means they're individually responsible for making their work useful to the company, and choosing projects that have this effect. It shuts down the traditional middle-management extortion of "you work for me or you don't work here" and gives everyone the same "freedom of the castle" in their choice of projects, but it's not a free-for-all.
suddenly people are back to feeling like they have bosses and assignments they need to fulfill to keep their jobs, only it's even worse because those aren't transparent...
Actually, that's a false security, in that you can do all your assignments well and please your boss and still get laid off... I know what you mean, though. I agree with you: it's going to be hard for Valve to preserve its culture. These are the kinds of times that test people.
True, I agree with you that if they can credibly convince the remaining employees that this is a one-time event due to extraordinary circumstances, and they are sure (as much as anything can be sure) it is not going to be repeated for >=5 years, culture is recoverable.
I've never seen that happen before, but I also mostly have information about more bureaucratic companies (especially my dad working for a top-100 global conglomerate as I was growing up), so Valve might do it right. One reason to think they might is the lack of lots of layers of management: multiple layoff rounds are often due to trying to cut the least possible to fulfill a target, but using the most optimistic assumptions, so finding out a few months later that, surprise, it wasn't enough.
I also agree that one round is pretty much the max you can explain away. Once you've gone to two, nobody believes that this is the last one.
It is being tested. This would be a very interesting (if difficult) time in which to be at Valve.
Ultimately, companies have to do this. Businesses can't expand indefinitely and, at some point, contraction becomes part of the process.
The hard part, assuming the best of Valve, is how it will describe what it did to preserve its culture while being respectful of those who are affected.
Here's what I don't understand about firing. Why not give any warning? Why not give people time to find another job? I mean if they were good enough for you to hire them and sit through the 2-3 week interview/hiring process - it should be fair to give them say a month of warning to tap their network, and hop over to another company. It seem unnecessarily painful and arbitrary to just say - "Your fired. Get out" - especially if they are great people, and all you are doing is shutting down their project.
Acts like these remind me of the old courts held by Kings during the medieval times.
The reason is threefold: morale, liability, and inability to use the worker for any long-term planning.
Having a person who knows they will be fired roaming the halls, poorly motivated is an obvious risk for the morale of the remaining employees. And it just drags out the pain of separation, even if the separation is otherwise amicable.
Second, there is some risk, no matter how generous the severance package, that the employee will become disgruntled and desire to harm the company. Even if it's a 1 in 1000 chance, the damage that a disgruntled insider can wreak is enormous. Huge downside, very little upside.
Finally, if they stay, they are now effectively temps. They can't be used in anything but short-term work because you don't know when they'll get the magic offer and be out the door.
A good severance package (say 2 months of salary, plenty of buffer to find a new job) costs just as much as keeping them on for two months after a warning, but has none of the downside risk. The only thing you lose by doing this is the very marginal short-term work you could allocate to them in the interim, which they would be poorly motivated to execute to the top of their ability.
Because time and time again, disgruntled and often soon-to-be or recently-former employees sabotage or steal from their employers. It's an IT security best practice to disable logins immediately when someone is notified or gives notice of termination. Often during the first HR sit-down. It's cold, but at least remaining employees aren't put at risk.
Well, usually that's part of the severance arrangement: you get to represent yourself as employed for some amount of time.
You do want people out of the office once you deliver the news, because even if you fire people fairly, they're not going to be doing useful work. Otherwise, I agree 100%.
Many companies use PIPs instead of severance offers in order to build a "performance"-based case against an employee. (This is sometimes part of a layoff being dressed up as a performance-based cut, and sometimes not.) That works out horribly. It's actually cheaper, considering morale issues, to cut a 3-month severance than to put someone on a 2-month PIP, because most people turn toxic during the "walking dead" period of the PIP. What PIPs are about is externalized costs: the HR office claims it "saved money" on severance payments, but the team and manager have to deal with a walking dead employee for 2 months.
I have never heard of severance being used to artificially extend the termination date (and I expect health insurance vendors would crack down on it), but I have never heard of termination date mattering in terms of getting the next job anyway.
Interesting random photo of a custom built computer in the antec mini case. Most major valve games won't render on a computer in that case, it can't hold a video card.
Here's what Valve's employee handbook [1] says on the matter: '"With the bar this high, would I be hired today?" That’s a good question. The answer might be no, but that’s actually awesome for us, and we should all celebrate if it’s true because it means we’re growing correctly. As long as you’re continuing to be valuable and having fun, it’s a moot point, really.'
Well, not so moot now.
[1] http://assets.sbnation.com/assets/1074301/Valve_Handbook_Low...