161–170 of 219 posts
In a large organization, NO SINGLE PERSON makes a decision. It's THE PROCESS that makes the decision. And THE PROCESS takes time. One of the greatest skills you can learn at a large organization is identifying, bypassing and avoiding THE PROCESS. It can be done, you just need to figure out how.
Earlier quoted context omitted.
Friction is a continuum. As the company gets smaller, the amount of friction gets lower. Also it depends on what you do for the company, putting a button on the front page of Amazon.com has more friction than putting a button on an internal service support tool or a test tool. That is to say Big companies can be OK if you land in the right department.
Yes, that's probably true, but ... as somebody who actually put a button on the front page of Amazon.com (back in 2010-ish, I wrote the presentation code for an MP3 sample player for album recommendations) it was actually pretty low friction. Write code, test it locally, deploy it as an A/B test, test was not negative - ship it. The only people I had to talk to were translators for internationalization. There was a t…
This might have been true years ago, but my recent experience has sadly been the opposite. e.g. on the projects I worked involving the website, testing locally has been virtually impossible.
I've done work for big insurance companies, "tech companies" like Verizon, or big soda companies, and they're as high friction as you'd expect. Somehow I imagined BigTechCos to use better tools and be nimbler on decision-making, letting employees make their own decisions without as much oversight.
Earlier quoted context omitted.
It took me a long time to figure this out, but the BigCo friction does come with benefits. Particularly benefits of scale and risk reduction. Need your app localized? Surprise, that happens automatically with BigCo processes. Didn’t think about GDPR compliance? Surprise, part of the delay you’re experiencing is that review. Need an awareness campaign? There’s a whole marketing team to help you with that. Thing big an…
Yup, exactly. The complexities of both running a large operation and operating globally are vastly greater than a small operation. Simply the number of nodes that must communicate has a combinatorial explosion. So, there needs to be a standard way of dealing with that, or entropy also explodes. So, bureaucracy happens, with processes that sorta fit everyone but rarely exactly fit, so friction increases. A saying I he…
Love this! Definitely gonna use this sometime.
Earlier quoted context omitted.
I've come to prefer working in companies that are between Series A and Series B funding: After Series A is when the startups really have money for compensation. Series B is where companies start implementing bureaucracy like OKRs, permissions for everything and the org chart starts getting deep.
Interesting. How can one identify what Series a company is? Ask during interview?
At a fruit-themed Bay Area company I can register a domain name behind a load balancer, issue a new cert, kickoff a VM to host it internally all in the same day without any approvals... so long as it is below a certain scale suitable for testing, skunkworks projects, etc. Approvals and meetings only become necessary if I want to deploy a production service, a large-scale service, or both.
Requests for a new repo or Confluence space are fulfilled same day and give me complete admin control over it. I can kickoff custom OS builds and generation of installation images via a self-service portal. I can create new email-enabled LDAP groups via self-service. I decide if the group itself is self-add or not, etc. Anyone else in the company can do the same. Many other things are as simple as direct manager approval.
You don't have to put up with bureaucratic nonsense, it just happens to be common in corporate America where management doesn't trust the peons to make decisions. Or perhaps where everything is driven by division P&L so no one wants to be responsible for spending money outside their fiefdom?
I prefer an environment where individuals have responsibility and yet where we accept mistakes will be made. As long as you learn from a mistake and don't continuously repeat it there is no need for a new process in response to it.
Interesting that this is so complicated in a big Bay Area tech company. At a fruit-themed Bay Area company I can register a domain name behind a load balancer, issue a new cert, kickoff a VM to host it internally all in the same day without any approvals... so long as it is below a certain scale suitable for testing, skunkworks projects, etc. Approvals and meetings only become necessary if I want to deploy a producti…
That being said, to each their own and I know a lot of my coworkers were really happy there.