Internal document explains why Google has become slow and bureaucratic

[1] https://komoroske.com/slime-mold/

I've got good news though. How do you fix this? Simple, don't overhire in the first place! Womp Womp!

It happened to Standard Oil, IBM, GE, RCA, Kodak, Sears, all of them.

The same failure happens to government agencies attempting central planned economies. Except that government agencies are never allowed to fail, and we all suffer.

When will we call the antitrust experiment from the 80s era “Chicago school of economics” an abject failure? Google should instead be several companies, each with a duty to consumers to compete with other companies in their respective businesses.

Even Google knows it, not that their leadership would ever admit it. Slow, wasteful monopolies are still appealing to those at the top of hierarchies.

I worked at Google for 10 years (2011-2021) and I don't remember a single time where it didn't feel slow and bureaucratic. Compared to the smaller companies and startups I worked at before. Maybe it's not slow compared to academia or other BigCorps where many other Googlers came from though.

Getting anything done always involved going through piles of "paperwork". Getting in the right MDB/Ganpati groups, finding the right persons to put on code reviews, juggling multiple code reviews at a time in order to make progress because your code reviews are just sitting in someone's review queue, "readability" reviews, applying to get readability, mandatory security reviews, design docs for the smallest features, going through rigorous launch processes, etc.

It all made sense in the context of the kinds of production environments Google has, but it was often applied intensely even for very small pieces of work.

On top of that, going through all that and having the paper trail for it was intimately tied to need to build paperwork for going through the perf & promo process, too. So it was entirely built into the culture there.

Also, if there was any interesting work, if you weren't some sort of political beast or very assertive, somebody would just steal it from you before you could do it anyways.

(Oh, and over half the people you met came directly from academia anyways, and so of course they were used to this kind of crap, because that kind of "defend your work", publish-or-perish, and showboating culture is all over academia.)

Has Google gotten slower over time? Maybe? It was sure bloody slow during COVID WFH.

What did get worse overtime was shitty bureaucratic organizational politics. Or what I'd call "shipping the org chart." Customer impacting decisions clearly made because of the structure of various product areas and organizational empires. Especially in the world of things that were outside of Core/Search/Ads and were consumer facing. Terrible.

Working at Google is a good way to learn about infrastructure at scale and make a lot of money. That's about it. The best career at Google is SRE. I wouldn't recommend a SWE career there.

Google's engine, which it has been specatacularly successful at, is obviously search. More specifically, it's search advertising. Google is a dominant player in display advertising (eg Google Ad Exchange, AdSense) but, at least when I worked there, these amounted to <10% of AdWords' revenue. They were a sideshow.

Chrome, Android, Maps and Youtube are all key parts of this engine. The latter three were acquisitions and that gets to the key point.

At some point a company becomes too large to innovate because no new business can compete, or will be allowed to compete, with the company's core business.

Google missed the social media boat for this very reason. Google+ came way too late. Google spent billions on this failed effort. The real mistake was not buying Twitter. I mean Google ultimately offered a rumored $6 billion for GroupOn at one point. They dodged a bullet on that one when GroupOn rejected the advance. Rumor also had it Google tried to buy Twitter in ~2010 when Twitter was seeking a $3 billion valuation.

Imagine if Google had slapped down $6 billion to buy Twitter. How much better off would they be?

There is a ton of process to do anything at Google. This matches my experience. And that's an issue but it's not the main issue.

Grove talks about something mentioned very early on in Alex's webpage[2] (right under the top link, to a genericized/not-Google-specific Slime Mold presentation), leadership of the gardening style (although specifically Komoroske is talking about gardening platforms). Making sure those with potential & good ideas are enabled, might be how I'd describe it.

On a personal note to me, it's amazing how much we assume - have left individual isolated groups/teams connected upwards & sometimes side-ways as the organizational structure. Leaving only higher up levels of the org structure to mediate & suss out truth, not having workers also put in their bid for how other projects should execute or where priority should be is the practice. Im fascinated by the potential of trying a more connected & aware org, where we try to spend more effort internally communicating & sharing bids for where we think we ought be placing our efforts, where we ought be redirecting or reshaping efforts.

[1] https://www.leadingsapiens.com/andy-grove-balancing-top-down... https://news.ycombinator.com/item?id=34381928

[2] https://komoroske.com/

Google owns large chunks of the markets they are in, they are not a startup but at the very top. If you're in that position you pretty much want to maximally exploit what you have and simply make money, there's no comparative advantage to prioritize exploration.

You'd basically want to be more bureaucratic, not less, given that bureaucracy is nothing else but the rational allocation and organisation of resources. If you own everything, making things slightly work better will net you a lot of profit.

I think Facebook with its founder oriented corporate structure now has this exact problem as Zuckerberg went on his Metaverse extravaganza instead of actually making the company print more money which is what they're actually good at.

For google, at least before 2006, I did not know personally anybody does write bad code got in there. Until I start to see someone will always copy and paste code from internet and try to make some code barely works made to Google by just consistently practicing coding questions, the trend is quite obvious that none of stuff coming out there has a distinct quality differences compared to other technology companies.

The two aren't incompatible. You don't need to take away people's independence and agency to have good leaders. A good leader doesn't need to be authoritarian in order to lead. It's bad leaders who have to force people to follow them.

And just because you want to be agile and flexible doesn't mean you can't plan ahead. You need a little of both.

What does a top-down approach look like by comparison for a company as large as Google, and is that even feasible, given that top management are likely to be many layers away from code being written?

Or is it something like when a company gets past a certain size then it has switch management styles?

Was hoping the article would get closer to talking about these ideas but it’s lacking in substance imo.

The "finish line" for Google is not the next Search, or real innovation, it is about longevity, maintaining an ecosystem, and "trading the news."

- Windows applications are risky, and inflexible -- Chrome.

- Laptops and Active Directory confusing ISD administrators -- Chromebook/Workspaces ecosystem.

- TikTok is getting a bad rap on the hill -- Shorts.

These are healthy and understandable business moves, but not everything these companies excrete is, good.

This is where buybacks should be terminated by our legislators. Not to punish these players for their rapid success, but to facilitate natural logistics for these hugely impactful, ubiquitouly embedded service providers.

On the less sane side of things:

- Google Wave beat Slack/Teams/Discord to market by 10 years.

- Google+ was 7-8 years late to the party.

- Stadia, obviously at-odds with AR/VR, and Deep ML.

- The only way to prevent "20% time" from hacking away at your own "terra firma" is to punish it, or make it impossible -- overhire and waste your employees time.

There are hotshots there, of course, but the reputation of their engineering team is that they prioritize a 10 to 3 workday, meetings, and a diffstat of 0 most days.

Or as Blind would call it, "decent WLB, high TC".

I am just trying to get a small thing done, not turn the whole damn ship yet that I am faced with the same bureaucracy and inefficiency. But I have had success overcoming that, at the cost of not being very popular.

But the real solution is managerial. It is remarkably simple: To get anything done, require the least possible number of people to be part of that thing being done.

But for managers, they love to throw people and money at problems.

But really, it is upper management that needs to basically micro-manage everyone under them specifically when it comes to headcount.

I fully expect the CEO of alphabet asking random project managers to justify headcount of not just direct reports but who is being involved in their projects.

If you need 2 devs to code, one to qa/test. That's it. You don't get anyone else. Not for redundancy, not for anything else. You don't involve legal or marketing, change management,etc... because that means rules and procedures and politics and meetings and planning.

The problem is very old: "If all you have is a hammer,everything is a nail". For mid and low level managers, all they have is "get more people" when they want things done. And the arrogance of being a bigcorp compounds this because "we are $bigcorp, we have a team for this and a well tested process for that,such and such team can help with".

You cannot compensate for loss of agility by claiming reduction in risk and liability. You should not let lawyers, marketing and middle management choke your company to death. It's up to top brass to piss off everyone in their immediate surrounding so the bottom line is good and the company is healthy. Ultimately, shareholders need to hold management accountable for wasting their money.

Just like the thesis of the slime mold talk, the talk itself is a result of the slime mold and doesn't particularly represent any company-authoritative take on the situation.

It's turtles all the way down!

In the past we've had pretty big companies/organisations that used to be quite innovative, I'm thinking of Xerox, HP, Bell, even Intel at some point, to say nothing of the various departments attached to the US industrial-military complex (starting with NASA), but then the people innovating back then hadn't joined those companies/organisations mostly for the money and money alone.

Get rid of the people who have joined Alphabet only to get the 500-600k yearly comp and you could get, at some point, some innovation in return.

When Xerox was first trying to commercialize PARC inventions, it was estimated that IBM's bureaucracy was so bad that it would take them nine months to ship an empty box.

Leaning up to go faster is code for: our shareholders are complaining, so 1. Corners will be cut or 2. Layoffs of the "slow and weak" are imminent.

It's never about problem solving. Unless your problem is that you need more money right now.

https://komoroske.com/slime-mold/

The source document/presentation ("Coordination Headwinds") and idea are many years old, and from what I recall were not aimed to be critical of the company and its organization, rather to give its employee readers a feeling of solace and understanding for why getting seemingly small things done can feel overly hard and slow… a less humorous, and less SWE-specific version of "Broccoliman" if you will.

Refer to the red panda that merely wants to serve 5TB of data.

  - https://www.youtube.com/watch?v=3t6L-FlfeaI
  - https://news.ycombinator.com/item?id=29082014

While the “slime mold” story provides a convenient and charitable explanation of Google's bureaucracy and dysfunctions, another explanation is:

  - As an organism, Google's executive decision-making functions are severely impaired by lack of accurate information, dysfunction of its middle-management “organs” or tiers, and impedance mismatches arising from genetic encodings or mental models associated with the slime-mold idea.

  - Despite being on a mission to “organize the world's information and make it universally accessible and useful,” Google is far less successful at doing this with its own information. These characteristics are not unique to Google, but some of Google’s cultural history makes this a particularly challenging problem.

The vast majority of Google’s product areas or business units, are not self-sustaining entities and all rely and draw upon resources, namely headcount, that unavoidably go through a centralized prioritization and allocation process. Ruth (Google CFO) and Sundar’s (Google CEO) jobs are quite difficult if you consider that: (1) they don’t have optimal or very accurate internal information, (2) they are constrained by what is provided to them and what people will actually do, (3) tributary sources of information generally have little understanding of broader processes, aside from what they expect to receive, and also have many incentives to fashion information in a manner that maximizes their own benefit – based on what they assume or understand.

Avery Pennarun (apenwarr@) touched on some of the above internally in a document called “View from Above” (go/vfa) and post-Google, on similar dynamics in a public blog post: https://apenwarr.ca/log/20190926

From my own experience at Google, where I served in engineering leadership roles and joined as an experienced industry hire, is that over time Google's culture – Namely its performance rating and promotion process, over-weight certain accomplishments and neglect critical characteristics and understandings needed from employees as they move into upper echelons of leveling (middle-management and higher), most importantly in the L6/L7 (manager, sr manager) and L8 (director) levels.

Many Googlers in those ranks continued to operate based on the understandings they held as project / technical leads (“TLs”), typically a Level 5 role, and continued to be promoted based on those understandings and/or increased headcount below them. Meanwhile, these roles, and the corporate organism’s need from these roles, are fundamentally from being a TL or project lead.

Compounding this problem is the fact that for most Googlers in these roles, scarcity of resources was rarely a problem, and getting headcount was easy. Most managers know how to ask for headcount but have no idea how to help support the conditions for its creation.

To be less wishy-washy, many people were promoted for continuing to do their jobs as they understood them, and their expanding team size to some degree, rather than understanding the full scope and understanding of the new level or role they were stepping into. And to some degree, it may not have been their fault as the persons and examples before them did the exact same thing.

Some might say this is the Peter Principle in practice, but at Google-scale.

Anyways… to wrap this up, I regularly encourage people to think about and learn to understand their boss’ boss’ problems, and ensure they are able to evaluate and represent their work in relation to those problems. Without establishing that context, your work may be completely misaligned or difficult to value when viewed from above.

At some point the search/ads people are so good, and they are insanely good, that there isn’t much pressure for anyone else to show PnL.

These men and women are gods and goddesses in their fields, and if their worst fault is funding some iffy shit and giving us Blink and Android I can live with it.

Their sideline vertical video business is highly competitive with TikTok and Reels and all they had to do was shit they knew already.

The Google people are scary good.