Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google’s Internal Data Show Engineers Found It Harder to Code from Home

theinformation.com

Post reply on HN
Print view View original

91–100 of 109 posts

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#91

Post by kyleee »

hn_throwaway_99 wrote: post #55
rightbyte wrote: post #37

"The number of [changelists] is a key metric the company uses to track productivity." Don't take the article seriusly.

Why? I'm sure Google has tons of historical data, and they have a large enough pool of developers where they likely get very consistent productivity metrics across the company. If they see an obvious, statistically significant dip in the number of changelists companywide, this seems like very useful data.

did they control for all the other factors other than work from home/office?

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#93

Post by mandelbrotwurst »

dacracot wrote: post #45

So first I can't read the article because there is no way I'm giving them my email so they can spam me later. That having been said, what is the unit of measure? Coding is notoriously difficult to measure and the unit of measure is elusive. We all heard the problems with KLOC, function point, ticket, story, etc, etc. They are all widely variable and seldom reflect the complexity or lack there of.

The non-paywalled snippet describes what they were measuring.

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#94

Post by meken »

tidepod12 wrote: post #58
xt00 wrote: post #18

I would guess based upon my own experience and talking to others: 1) the first 3-5 months were hard to figure out a good rhythm and is super impacted by how busy your house is (do you have kids, live-in parents, special needs kids etc) 2) many people have figured out a good rhythm by now and are doing fine 3) people who were new at the company this year are definitely having a much harder time since just sitting next…

>more likely a mix is going to be a good setup.. where very likely new people should expect to be onsite for fairly extended periods of time until they have a much better sense of what to work on The model that I have seen with moderate-to-great levels of success at the large consulting companies (which have long had a strong remote work culture) is that the companies schedule new hires so that all new hires start at…

What company does this?

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#95

Post by phendrenad2 »

JSavageOne wrote: post #46
phendrenad2 wrote: post #3

There's going to be a BIG push to go back to offices soon. I expect some companies will start experimenting with micro-offices (4-5 employees in a medium wework room).

Highly disagree. Most people I know don't want to return to the office. Any company that makes a "BIG" push to force employees back to offices will see a ton of attrition since now practically every company is remote.

I think we're both right. FAANG-tier companies may push for a return to the office, and a lot of good engineers will go elsewhere.

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#96

Post by still_grokking »

andrew_ wrote: post #30
tidepod12 wrote: post #24

Earlier quoted context omitted.

>at that skill level What skill level is that? For the most part we are talking about a bunch of people ages 22-30 that function by responding to specific feature requests or JIRA tickets from their team leads. There are of course exceptions, but by and large this population of engineers do not do well with unstructured environments, such as the type created by a sudden shift to WFH. I am wholly unsurprised that prod…

I know quite a few brilliant folks there, and I just can't buy the assertion that the majority are drones/codemonkeys that drown without instruction and structure.

> the majority are drones/codemonkeys that drown without instruction

You can't expect anything else alone for statistical reasons:

Companies as large as G need a lot of employees, but there are by definition only a few percent of top people around (as one has to stand out significantly form the mass to be considered part of the top people). As a result a big company can't hire only the best of the best, as there are simply to few available! (Smaller companies can be much more picky in this regard as they don't have to fill as much positions). As a result you will have in a large workforce a large amount of at best average employees, meaning "the majority are drones/codemonkeys".

Additional there is cost. You can usually choose to hire one top dude, or respectively a small team of codemonkeys… For a lot of (most?) tasks a team of drones will do just fine when given the right instructions. So having only a few top people (giving instructions to the less experienced / talented) will work out mostly just fine.

G is especially know for building their process around that reality. They put some effort into handling the fact that most of their hires will be young inexperienced people. They even created a dedicated programming language in which their large amounts greenhorn hires could be potentially productive. That this language is "dubbed down to the max" is a hard business requirement for exactly that reason.

I think dealing with the fact that the majority of people in software are (and will indefinitely stay) just codemonkeys is actually a huge part of a software business' success story as it develops into a larger enterprise. An important factor to successfully grow is the ability to make good use of the "average majority" your workforce consists of I think.

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#97

Post by tidepod12 »

meken wrote: post #94
tidepod12 wrote: post #58

Earlier quoted context omitted.

>more likely a mix is going to be a good setup.. where very likely new people should expect to be onsite for fairly extended periods of time until they have a much better sense of what to work on The model that I have seen with moderate-to-great levels of success at the large consulting companies (which have long had a strong remote work culture) is that the companies schedule new hires so that all new hires start at…

What company does this?

The large consulting companies that I'm aware of (MBB and Big4) have (or had, pre-covid of course) some variant of it. For example, Deloitte built a resort hotel near Dallas and branded it "Deloitte University" for the purpose. I know PwC did yearly training in Orlando for new consulting hires that are hired from college campuses. I believe Accenture does something similar in Chicago at the "Q Center" (though only for ~1 month).

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#98

Post by kyleee »

tidepod12 wrote: post #86
kyleee wrote: post #84

Earlier quoted context omitted.

Honestly sounds like a failure of management if slackers are able to keep up appearances for long enough to cripple entire team efforts.

It has nothing to do with "keeping up appearances". The slackers are easily and quickly identified (usually just from the indicator that they chose to WFH 100% of the time). The problem is that, as I mentioned, it is ridiculously hard to remotely correct a slacker situation even after it has been identified.

so everyone knows the problem people, but it's hard to sanction/fire them? again sounds like managment/organizational failure. Surely judging performance of WFH employees and making efforts to help them is a surmountable management problem

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#99

Post by tidepod12 »

kyleee wrote: post #98
tidepod12 wrote: post #86

Earlier quoted context omitted.

It has nothing to do with "keeping up appearances". The slackers are easily and quickly identified (usually just from the indicator that they chose to WFH 100% of the time). The problem is that, as I mentioned, it is ridiculously hard to remotely correct a slacker situation even after it has been identified.

so everyone knows the problem people, but it's hard to sanction/fire them? again sounds like managment/organizational failure. Surely judging performance of WFH employees and making efforts to help them is a surmountable management problem

Have you ever worked in a management position at a decent sized company? Firing someone is not an easy task. Unless it's an egregious case (violation of laws/policy, sexual assault, etc), firing someone, even someone that literally produces zero work product, can take months or even a year. That's a slacker that is going to sit on your payroll, disrupting your teams flow, and eating your budget for months/a year while you build your case for firing.

Then, even after you fire them (or let's even say by some miracle you were able to fire them quickly), it costs a lot of money and time to hire someone new. That's yet more time and budget that isn't going into your team, and your team is also a person short while you go through the process of hiring and then training.

It is leagues and leagues better for "the slacker situation" to not be a problem at all, and it is much less likely to be a problem if prospective slackers aren't given the opportunity to slack off at home to begin with.

Re: Google’s Internal Data Show Engineers Found It Harder to Code from Home

#100

Post by toast0 »

DoofusOfDeath wrote: post #66

I get the sense that many employers see WFH as a temporary measure rather than long-term shift. And consequently they're not willing to seriously invest in tools and processes that would make WFH maximally productive. Suppose for example that Google enacted this policy: (1) They reimburse each employee up to $30k to upgrade their home office, including soundproofing. After Google trained local builders on effective s…

> (3) Google updates their team rules and meeting rules s.t. if anyone in a team or meeting is remote, then the entire team or meeting is remote.

(When offices are open) Dialing into a well equipped conference room is mucb better than doing a meeting with 5 or 10 people in little phonebooth rooms, if there are even enough rooms. Taking conference calls at open office desks is a non-starter.

Post reply on HN