[HN Gopher] The Shitty Stack System
___________________________________________________________________
The Shitty Stack System
Author : positivesum
Score : 103 points
Date : 2023-06-25 20:35 UTC (2 hours ago)
(HTM) web link (tryingtruly.substack.com)
(TXT) w3m dump (tryingtruly.substack.com)
| giancarlostoro wrote:
| Competition drives innovation when its external, if its internal
| it would need to be on a per team basis, and even that feels
| risky. Pitting members of a team against one another seems
| counterintuitive. I think of every company I join as a ship, so
| if another person from another team or my own is struggling with
| something and I can throw any advise or look and help, I do it,
| if I cannot because it is out of the scope of my area of
| knowledge, there's a small chance I might peek and tell them
| sorry I don't see anything that sticks out. Of course if I'm
| super swamped I might not at all, but if I'm reasonably busy or
| not busy enough, I'll pitch in.
|
| I don't think a company ever really recognizes the value in
| teamwork that I have seen, but it doesn't matter to me, the way I
| see it: we're all on the same ship, if you sink, I might sink
| too. Even if I never get rewarded for my efforts, I would much
| rather take the gratitude of other developers. As a result, when
| I apply for jobs I have several people who can vouch for me
| having worked with me, and I'd gladly do the same for them as
| well.
| totallywrong wrote:
| Steve Ballmer doing of course. He's got to be the absolute worst
| CEO of any major tech company ever.
| angarg12 wrote:
| So recently I was discussing with a colleague how misalignment of
| incentives has shafted our team in a certain Big Tech.
|
| I'm part of a high performing team, and for the last few years we
| have been delivering high quality output while hardly increasing
| our team size.
|
| We have a sister team that is, to put it politely, not as high
| performing. They have been struggling to deliver even though the
| have almost twice as many engineers.
|
| Last year we got some headcount to grow our team, but leadership
| decided to shift it to our sister team, since they had problems
| reaching their goals (you know, the squeaky wheel gets the
| grease). This was fine with us since we weren't in a particular
| hurry, and we would pick up the headcount later.
|
| Fast forward and we have a hiring freeze, meaning that our sister
| team was able to grow headcount, while we can't grow ours.
|
| The double whammy? since we don't have as many people as we
| expected, we couldn't pick up projects that were in our roadmap.
| Guess who got those projects instead? our sister team.
|
| Morale of the story? do a shit job and you will be rewarded with
| bigger teams and more scope, which in turn will translate into
| promotions and bonuses.
| tennisflyi wrote:
| Can't be promoted if you're too good/irreplaceable.
| nine_k wrote:
| It could be compensated at least by a pay raise.
| o10449366 wrote:
| I've had the same experience at the big tech company I worked
| at.
|
| I expressed these frustrations to upper management and the
| explanation I was given was that for many of these big
| companies, their goal is to grow their 1,000 sub-businesses to
| 10 million ARR businesses each. Growing these 10 million ARR
| units to 100 million ARR is much more difficult than going from
| 0 to 10 million, and so when they have a successful unit, they
| disproportionally reinvest that success into the poor
| performers (the 0-2 million ARR units).
|
| This felt backwards and unfair to me as an engineer on the
| teams that worked hard to get to the 10 million mark. It really
| sucked seeing my team lose headcount or engineers to less
| successful and poorly designed/managed projects. But from a
| business perspective on the larger scale I see why they do it -
| I just think it creates a poor engineering culture and, in the
| long-term, a sub-par customer experience because velocity in
| product development slows when projects are de-invested or put
| into KTLO.
| wkdneidbwf wrote:
| is it common to roadmap features your team can't deliver at
| their current size? that sounds like a problem.
|
| also, the team that takes on the most work but does it poorly
| getting the promotions and bonuses also sounds like a problem.
|
| both of those thing aren't problems with developers though--
| theyre management problems.
|
| i would interpret the moral as bad management leads to sub-
| optimal outcomes for developers. not advocating correctly for
| headcount, and not showcasing the high quality work your team
| does are literally what your manager should be doing.
| otoburb wrote:
| >> _[...] both of those thing aren't problems with developers
| though--theyre management problems._
|
| Sounds like the sister team had a better manager in some
| aspects of "managing" (e.g. managing the message up the
| chain) compared to the OP's manager who may have been great
| at facilitating (or minimizing meddling with) the high
| performing team but was less skilled in managing up.
| gmarx wrote:
| Decades ago when I was getting started there was a team that
| fell way behind and everyone on the team worked weekends and
| many days until midnight and finally delivered some buggy
| arguable product 6 months late. They were hailed as heroes. No
| one even noticed me and my team delivering on time. I learned a
| valuable and entirely inapplicable lesson.
| Tade0 wrote:
| > They have been struggling to deliver even though the have
| almost twice as many engineers.
|
| My experience tells me that they struggled _because_ they had
| twice as many engineers.
|
| I don't think I've been to a 20+ person project which could
| honestly be called "successful". Hard to say if it's a symptom
| or a cause. Personally I lean towards the former, because when
| someone tries to throw people at a problem they either don't
| know what they're doing or delivering is not their priority.
| Scramblejams wrote:
| Frustrating dynamic. Something similar played out for me in a
| different industry, where a company doled out a bunch of
| different parts of a large project to different subcontractors.
| The subcontractor I was working at was apparently the only one
| that hit the dates and deliverables. So we got put on pause
| (aka defunded) while the other contractors were paid to catch
| up.
|
| I never was sure why some of the unfinished work wasn't shifted
| to us, but the end result was the same as yours -- getting
| penalized for doing the demonstrably superior work.
| softwaredoug wrote:
| Maybe the team should be _resilient_ first and foremost: when
| high performing members leave voluntarily, it can readily accept
| and incorporate new members, while maintain a certain work output
| and high morale.
|
| Such a team has a reasonably amount of voluntary turnover. Too
| little turnover and too high turnover create different problems.
| Either teams stuck in their ways, overly rigid in roles, and
| overly dependent on certain heroic members, the org is terrified
| to lose. Teams with too much turnover lose valuable experience
| before it can be passed on.
|
| In my mind good teams are teaching and learning machines: members
| adopt new/different ideas, different roles, new members, by being
| given safe space to fail and be curious.
| Klinky wrote:
| I don't think artificial turnover is the solution. Teams need
| good documentation and design or else onboarding/offboarding
| will be painful and a huge distraction. If you've got good
| documentation and design methodology, then there is no need to
| churn the team artificially. Keep documentation up to date,
| design well, and redesign areas that are tech debt or not done
| well. That's easier said than done though. There is no magic
| bullet.
| noizejoy wrote:
| > I don't think artificial turnover is the solution.
|
| I didn't see this being suggested in the parent post.
| softwaredoug wrote:
| I actually think there's optimum voluntary and non
| artificial turnover where people naturally move to other
| opportunities. And conversely a calcification where people
| get stuck in a comfort zone.
|
| I don't think people should be forced to leave. Instead I
| paradoxically think people should be hired more interested
| in growth and this more likely to leave / move around the
| company / try different roles.
| Klinky wrote:
| The entire second paragraph is about finding optimum
| turnover to keep the team fresh and dynamic. I think it's
| impossible to control organic voluntary turnover, as you
| don't control people's lives/desires. The alternative would
| be artificial turnover, moving people in and out of teams
| as needed to meet this goal. I think people joining and
| leaving is disruptive regardless of it being organic or
| artificial.
| sgt101 wrote:
| Company I worked at had this inflicted on them by McKinsey,
| untold woe followed. There were suicides, many nervous
| breakdowns.
|
| A colleague (a great man who is utterly unregarded professionally
| but is one of the most profound thinkers I ever encountered) had
| a brainwave which was to calculate how much the stack ranking was
| costing. We discovered that every year the company was spending
| nearly 15% of its budget for management grade employees on the
| process. This estimate came from two sources, HR's official
| requirements and guidance and surveys of the management group.
|
| It turned out that large management consultancy found that it was
| doing the same at the same time we did; this information and my
| colleagues presentations of his teams research resulted in the
| system getting dropped.
|
| God damn the people at McKinsey that inflicted this so widely.
| "In search of excellence" must be counted as one of the artifacts
| and efforts that has been most harmful to democracy since 1945. I
| am waiting for an apology.
| truculent wrote:
| The London Review of Books had a pretty damning article on
| McKinsey and management consultancy:
| https://news.ycombinator.com/item?id=33869800
|
| "One gets the sense that Bogdanich and Forsythe think the
| consultants they write about are rotten apples, but the barrel
| is sound. Their own material makes clear, however, that all the
| services often spoken of as merely helping businesses and
| government departments run more efficiently - management
| consulting, audit, software development - are in fact focused
| on enabling capitalists to enrich themselves further without
| the inconvenient interference of workers, taxpayers or
| regulation. Thanks to the hegemonic model McKinsey and other
| management consultants invented, these firms not only make and
| remake businesses and government in the image of their laissez-
| faire fantasies, but see homo economicus as the last word in
| modern selfhood."
|
| (HN comments: https://news.ycombinator.com/item?id=33869800)
| sgt101 wrote:
| I should add, on the journey to the evidence that finally
| killed the stack rank we did some interesting empirical work.
| This was in collaboration with an organisation that actually
| became part of the dreaded firm.
|
| Still, it was good stuff. In F1 racing you have a unique
| opportunity to study employee performance empirically. Why?
| Well - every year there's a new car. The components that get on
| the new car get on it because of an empirical process. If your
| team can lighten something, create more drag or reduce drag or
| increase power then your component gets on the car.
|
| This insight allowed individual performance to be tracked. The
| results - well it turns out that for almost the whole workforce
| it's luck; future performance cannot be predicted from past
| performance. You screw up one year, but then the next you knock
| it out of the park. There was an exceptional group though.
| Being in the highest percentiles in successive years was
| strongly predictive of being in the highest percentiles for
| future years.
|
| In the follow up work this was investigated and it was found
| that managers could reliably identify the individuals in their
| departments that would be in the group of "persistent out
| performers".
|
| There was no similar group of low performers.
| pinewurst wrote:
| It wasn't just Microsoft - many companies were equally infected
| by this evil meme coming out of Jack Welch/GE which of course
| imploded.
| rwmj wrote:
| Yes it's strange they wrote this whole article without
| bothering to investigate the origins of the system. Not even a
| cursory look at Wikipedia apparently:
| https://en.wikipedia.org/wiki/Vitality_curve
| pinewurst wrote:
| It could actually be even worse as Welch was also keen on
| something called "Topgrading" which mostly never really
| caught on.
|
| https://en.wikipedia.org/wiki/Topgrading
| zac23or wrote:
| A company I worked for did worse than stack system:
|
| This company has many buses. And it started to measure every
| movement of the bus maintenance team (with the help of a
| consultancy). How long does it take to remove the engine? 1 hour.
| And the wheels? After that, they made a calculation and fired
| "unnecessary employees".
|
| The remaining people couldn't do the job, it wasn't possible. An
| engine with a bad bolt is much harder to remove, easily doubling
| time.
|
| In the end, a steering rod broke, killing the driver and others
| on the bus.
|
| Nobody was arrested, nothing happened.
|
| My experience in several companies... led me to create the theory
| (or have I read it somewhere?) of Entrepreneurial Theater.
|
| All the methodologies I was forced to work on basically didn't
| work, someone with power wanted to do that because they did.
| Those who do well in this theater are not good employees, they
| are the ones who perform best in the current play.
| sourcepluck wrote:
| https://www.alfiekohn.org/punished-rewards/
|
| "In this groundbreaking book, Alfie Kohn shows that while
| manipulating people with incentives seems to work in the short
| run, it is a strategy that ultimately fails and even does lasting
| harm. Our workplaces and classrooms will continue to decline, he
| argues, until we begin to question our reliance on a theory of
| motivation derived from laboratory animals."
|
| That was 1993. The people at Microsoft, as well as the author of
| this substack, would do well to have a think about Kohn's
| arguments.
| spamizbad wrote:
| I'd argue Kohn's follow-up work isn't too far out of line with
| the author.
| onos wrote:
| MSFT stock up 100x since then. Edit: I guess this was your
| point, criticism was either heeded or was off.
| einpoklum wrote:
| Sure, and NFT prices went up times infinity. Doesn't mean
| that's benefited human society any.
| whywhywhydude wrote:
| MSFT stock stagnated for more than a decade and jumped once
| they fired Ballmer and got rid of his stack ranking system. h
| ttps://www.macrotrends.net/stocks/charts/MSFT/microsoft/sto..
| .
| gigel82 wrote:
| I heard it's back now (not sure if for this year only due
| to the reduced budget, or just ongoing).
| Aleklart wrote:
| Can confirm, worked for a while back in 2011, never saw more
| dysfunctional, hypocrite and senseless organisation ever since.
| Now I consider it hilarious, but I remember how terrified I was
| of kafkaesque processes I were involved and what CPE was. I still
| keep 4 pages list of KPIs i have to be "green" with.
| elcritch wrote:
| > "People responsible for features will openly sabotage other
| people's efforts. One of the most valuable things I learned was
| to give the appearance of being courteous while withholding just
| enough information from colleagues to ensure they didn't get
| ahead of me on the rankings."
|
| Ah academics are well versed in this method! You would think
| science would be an endeavor to share knowledge, but in the
| current cut throat scientific world this has become a norm.
| Research must look reproducible but not actually be so without
| significant work to fill in the gaps. It doesn't hurt if it
| covers up actually unreproducible results either as long as it
| can get in a top journal.
|
| It also squashes genuinely novel research since it'd be too
| risky. Better to do small iterative tweaks.
| lolinder wrote:
| This reminds me of Bret Devereaux's timely observations about
| coup-proofing in authoritarian regimes. Fostering internal
| competition absolutely wrecks your organization's efficacy in
| accomplishing goals, but if you're operating from a position of
| dominance relative to external threats it can help keep power
| concentrated in the hands of the people who aren't forced to
| participate.
|
| > For low legitimacy forms of government, like autocracies, the
| concern is squarely centered on internal stability, and here we
| see a wave of armies designed primarily for 'coup proofing.'
| Russia's military is actually a pretty good example of how this
| is done. An authoritarian government is looking to both maximize
| the ability of the army to engage in repression while minimize it
| as a threat to its rule. 'Coup proofing' of this sort follows a
| fairly consistent basic model (which I may elaborate on at a
| later date). First, command needs to be divided so that no one
| general or minister of defense can turn the whole defense
| apparatus against the leader. You can see this with how the
| Russian armed forces were fragmented, with Rosgvardiya and Wagner
| Group not reporting to the ministry of defense, but it also
| extends to the structure of the Russian Ministry of Defense,
| where the Army, the Navy and the Airborne forces (the VDV) all
| maintain infantry forces. Setting things up that way means that,
| in a pinch perhaps elite, well-paid and loyal VDV forces could be
| used to counter-balance grumbling disloyalty in, say, the army.
| Of course such fragmented command is really bad if you need to
| launch a conventional war, as, in the event, it was.
|
| https://acoup.blog/2023/06/09/fireside-friday-june-9-2023/
| jameson wrote:
| "Incentive structures work, so you have to be very careful of
| what you incent people to do, because various incentive
| structures create all sorts of consequences that you can't
| anticipate."
|
| - Steve Jobs
|
| My current employer has similar problem. Company hired executives
| from a certain company about two years ago (hint: nepotism) and
| shifted towards incentivizing teams who met OKRs and OKRs were
| only product focused - features.
|
| Though this makes sense on the surface, what it ended up
| promoting is doing whatever to meet the OKRs. Short-term solution
| was often acceptable leaving huge tech debts and it is hard to
| justify doing anything to maintain system complexity in a
| reasonable bound because they do not directly correlate to an
| OKR.
|
| Personal developments aren't part of OKR so junior engineers who
| should be focusing on learning struggled.
|
| Team collaboration became a hostile because dependent teams will
| call out other teams for blocking a feature releases.
|
| Executives shared during all-hands their intent isn't to neglect
| personal development or team collaboration but problem was it
| doesn't matter to employees because those aren't what's
| incentivized.
|
| Perhaps if the team or personal performance review required
| feedback from them (360 degree) and takes a non-negligible
| portion, then things might have been different but it's still all
| about what I got done
| einpoklum wrote:
| > Though this makes sense on the surface
|
| Stopped reading your comment at this point to not bias my
| reply, but - how does it at all make sense on the surface?
|
| Would you like to get this feature-rich product?
| https://t.ly/6KUd ...
| dilyevsky wrote:
| Depends on who your customer is. In large enterprise markets
| it tends to turn into feature bingo
| scottLobster wrote:
| Seems to me the key issue is few people in corporate leadership
| roles actually want responsibility or care about the mission.
| It's the classic Peter Thiel line that "competition is for
| losers". Far easier to change something arbitrarily without
| considering the consequences, because it makes you look
| "dynamic". If anyone criticizes your plan, talk about "necessary
| trade-offs" and just shrug off the criticism.
|
| I see it at my own large corporation which is going through a
| top-down reorg right now under the new executive administration.
| For those of us actually creating and delivering products, the
| conversation is: "So this plan is somewhere between meaningless
| and stupid and will solve a whole lot of nothing, how can we use
| it to get more funding?" No one outside of fresh college grads
| expects leadership from the executives, they're largely clueless
| assholes engaged in corporate peacocking and backstabbing for
| their own shallow reasons, but the clusterfuck of high-level
| backstabbing spits out money; so like kids playing in the
| sprinkler, we position ourselves to catch the water as best we
| can.
|
| I wish it operated differently, and at smaller companies where
| the executives are closer to the ground or have an actual vision
| it appears that it can. But the norm is mediocre and shallow
| leadership, like those aristocratic 19th century British officers
| who could literally buy their commissions, and did things like
| haul grand pianos into war zones.
| GMoromisato wrote:
| What is Microsoft like now? How are people evaluated and
| compensated?
|
| I assume not everyone gets the same salary/bonus, so what
| criteria do they use to decide who gets what?
|
| [I'm not doubting stack ranking sucks--I was at Microsoft
| 2005-2011--I'm just wondering how they fixed it.]
___________________________________________________________________
(page generated 2023-06-25 23:02 UTC)