[muninn_]

I don't buy it. Many companies are able to get the perks, culture, and autonomy right without being a multi-billion dollar company. It does enable Google to do certain things (like moonshot projects) that smaller companies can't do, but a lot of the perks I've seen are replicated in other companies.

In fact, I think your comment "Google doesn't do what it does to be successful, it does it because it is successful" is completely ignoring how Google came to be successful in the first place, which was doing what you're saying it does because it's successful. Reading something like "How Google Works" goes over this exact thing in detail.

[Periodic]

I tried to clarify the wording. I'm specifically trying to call out the structure of Google's management and the general way they approach problems.

For example, even within the profitable Google portion of the business, they often have many competing projects. There will be competing search features, competing chat applications, competing cloud services. Google will make big speculative bets or internally competitive bets because it can afford to just let people build things and see what happens. Google, as a company, doesn't really seem to have any idea of what most individual engineers are doing week to week.

Many companies can't do that. If you are a medium-sized company you probably don't have the market position or revenue to afford to be unfocused.

[user5994461]

> Google, as a company, doesn't really seem to have any idea of what most individual engineers are doing week to week.

Correct... but also correct if you replace Google by any company.

I've worked from zero management company (ala. Valve) to large fixed hierarchical structure (government).

The company never knows what people are actually doing. If it tries to find it, people will either game the system or waste time to justify they're doing instead of doing it.

[hkmurakami]

>For example, even within the profitable Google portion of the business, they often have many competing projects. There will be competing search features, competing chat applications, competing cloud services. Google will make big speculative bets or internally competitive bets because it can afford to just let people build things and see what happens. Google, as a company, doesn't really seem to have any idea of what most individual engineers are doing week to week.

This sounds exactly like how Microsoft works. They (famously) even had 2 versions of Windows 95.

[tajen]

I didn't know – and haven't found – the 2 versions of Windows 95 you're talking about. Source? Is it Windows 93 vs 95? Is it 95 vs "Plus!" ?

[dgant]

Might be referring to NT

[vkou]

Yet, free markets beat out top-down managed economies because they have competing teams seeking to provide the same services.

Tech is, to an extent, a winner-takes-all market. If you build out two competing products internally, in the long run, it will cost you less then building one product, and having an outside competitor fight you for market share, by building that second product.

Case in point: I'm sure that Facebook would have loved it if one of their teams built Whatsapp. For 20 billion dollars, they could have funded a hundred competing internal products.

[nostrademons]

For most companies, having two competing products internally does not also prevent you from having two competing products externally. There's no law preventing an outsider from starting a competitor just because you've already got two alternatives.

Whatsapp is a great example - Facebook already had Facebook Messenger, and Facebook Chat, internally. Google had GChat and Hangouts and whatever else they're working on these days. It didn't prevent the outside company with half a dozen employees from eating their lunch.

The parent's point is that because of their core markets in search & social networking, Google and Facebook have the luxury of no major competitors, and so they can afford the internal competition to advance the state of these markets. The market structure came first, not the corporate structure. And if you try that in a market with few barriers to entry (like mobile chat, before everyone had built their network effects), it's just as likely that a competitor outside the company will eat the market, not one inside the company. Probably more likely, since internal projects are hamstrung by things like executive approvals, PR worries, and potential legal issues while startups can just take their product to market and see where the market takes it.

[dispose13432]

They don't seem to play free market. They seem to play "new guy takes all".

For example, Google Voice was put in maintenance mode and Hangouts took off. Then Allo, then ... ?

They very rarely actively push two products with the same purpose.

[vkou]

No, it's not a free market. They also don't try to poach other teams' employees with higher salary offers, push bad press about the other team, try to be ramen-profitable, etc.

But you can get some of the benefits of a free market by having teams work on multiple solutions to one problem. You can also squander them by making the wrong high-level decisions.

[Animats]

Google Voice, previously Grand Central, was in maintenance mode from when it was first acquired. Google never improved it much.

[jrochkind1]

Pretty much _every_ google product is in maintenance mode from when it is first released, and never improved much.

This may be part of the strategy, those that somehow turn into 'profit' are the ones that actually have full-time engineering assigned to them, and get improved? I couldn't say.

For example that I ran into recently, Google Custom Search/Site Search theoretically _has_ an API for creating/modifying custom searches... but it's completely inaccessible, because the only auth method it supports is one that no longer exists. I suspect it is still running on a server somewhere with zero traffic though. If you Google around, you can see for at least a year, _some_ people (prob those with the paying 'Site Search', not the free 'Custom Search') have managed to get answers from this from google support -- the answer is "There is an internal feature request ticket to add OAuth 2 support to this API, but I can't give you an ETA".

[euyyn]

> Google, as a company, doesn't really seem to have any idea of what most individual engineers are doing week to week.

This isn't true. Even I, as a software engineer, can look up with ease what the majority of other people work on.

[nunez]

I hope you're not talking about peoples OKRs or internal profiles on Grow or whatever it's called now because you know that stuffs not accurate

[euyyn]

Wasn't thinking of those, no.

[gonehome]

It may end up that even Google can't afford to be unfocused - if they do end up losing their search engine ad revenue because it and standard websites become less relevant it'll probably be what's pointed at.

You can see their disorganization in their software and products - letting ideas like Google Voice that were ahead of their time languish until iMessage comes and takes over, lots of failed software launches etc.

[bitmapbrother]

>Google, as a company, doesn't really seem to have any idea of what most individual engineers are doing week to week.

Isn't that basically the job of the manager?

[muninn_]

I don't disagree with your follow up. Thanks for the post.

[ravitation]

He's not talking about the perks. He's talking about the design of the organization, beyond the culture.

The structures, the roles, the processes.

Autonomy is a value, but values only influence design.

[muninn_]

The post I responded to mentioned (and naturally I mentioned this as well) perks, autonomy, structure, etc... and then says that the missing piece is billions of dollars in ad revenue.

So I pointed out that many companies are able to get these things right (/imitate Google), without billions of dollars in ad revenue. It's not the ad revenue. That's my contention.

[ravitation]

What I meant is that "perks" are just an outcome of a certain organizational structure designed around a set of values.

There are multiple structures that allow for "perks", or are designed for autonomy. And, as you said, others have produced structures with similar results or based in similar values.

But, he's talking about the specific structures at Google.

[jasode]

>Many companies are able to get the perks, culture, and autonomy right without being a multi-billion dollar company.

You're right but I understood what Periodic intended to explain. Unfortunately, he phrased it as "billions in ad revenue" instead of a less common metric of "profit-per-employee".

If the company generates enough profit-per-employee, it can afford to pay for lavish perks.

Consider 2 different companies' revenue,income,#employees:

Walmart: $482b, $24b, #2300000 -- ~$10k per employee

Google: $74b, $19b, # 69952 -- ~$276k per employee

Take one example of a perk such as free catered food which is estimated to cost $20-per-employee-per-day.[1] For Walmart to offer this perk multiplied by 2.3 million employees, it would cost ~$11 billion which is almost half the $24b profit. Google's calculation for 69k employees is ~$349 million which is less than 2% of the profit. And the free food is just one perk... there are also massages, haircuts, laundry, daycare, etc.

Yes you don't have to have a billion in revenue to offer perks like that. SAS Institute of North Carolina in the 1990s didn't yet reach $1 billion in revenue but they offered lavish Silicon Valley style perks.[2] They could afford it because they sold statistics software with high margins (profits). Although they are a private company, we can safely assume they are generating much more profit per employee than Walmart.

So, if we have a hypothetical medium-sized company that has a revenue of $400 million with income of $100m ... to afford lavish perks, we need to generate that $100m with ~360 employees or fewer (Google benchmark). If it requires hiring 9500 employees to generate $100m (Walmart benchmark), it means we will have to pinch pennies and charge the employees for each pen taken from the supply room.

Direction of cause-effect: create a great high-margin business so you pay for free gourmet food.

Attempting the reverse of subsidizing $20 for each employee doesn't mean you'll end up with a high-margin business.

We can generalize the free food case study to other "Google perks/methodologies/culture" to see if we really have the direction of cause & effect correct.

[1] http://www.businessinsider.com/2008/4/googles-ginormous-food...

[2] 60 Minutes story (2002): https://www.youtube.com/watch?v=lvsIcwHavOs

[muninn_]

So I understand what you're saying here, and it makes sense. At the same time, I see career section after career section offering the same kind of perks with much smaller companies. Maybe it's not 3 meals of gourmet food, but free catered lunches, great 401k matching, unlimited vacation, great health and dental benefits. I'm not saying the perks are matched tit-for-tat, but I still stand by my statement that you don't have to be a multi-billion dollar company to afford those perks.

Likewise, Walmart can offer those types of perks. Just not to people working on the floor or in distribution centers.

Thanks for your original reply :)

[e15ctr0n]

> I see career section after career section offering the same kind of perks with much smaller companies. [...] you don't have to be a multi-billion dollar company to afford those perks.

Many of these smaller companies are VC-funded startups that have zero revenue and are burning VC money to provide these perks.

[ghaff]

These are by no means the norm outside of Silicon Valley VC funded companies where they have become table stakes. Good health and dental in many cases but not the rest. And unlimited vacation, where offered is, even ignoring the more cynical interpretations mostly about eliminating liabilities especially under California law.

[morgante]

I'm guessing you're mostly looking at career sections for technology companies. In general, tech companies have quite high revenue per employee.

You're probably also looking at startups, which don't yet have material revenue but plan to reach margins closer to Google's than Walmart's.

[brazzledazzle]

It's not fair to count retail or warehouse employees. Walmart would only need to offer those perks to its corporate employees.