Hacker News Reader

Somewhat related to this point: A large conglomerate I worked for managed to kill one of their major products by making the mistake of assuming that programmers are easy to replace with any other skilled programmer. So they moved a product team to a new office, knowing that the highest-paid people on the team would not choose to relocate, and assuming it would be easy to backfill them with inexpensive new hires at the new location.

Except, in this case, the programmers in question had decades' worth of hard-won domain expertise that was absolutely critical to their job function. And, this product targeting a niche industry, it was relatively uncommon domain expertise. So much so that the only people on the job market who had it were these folks whom they were ousting. All of whom had found positions at competitors within a day or two.

So then the company lost their ability to maintain the product in any meaningful way, and, within a year or two, customers started moving on as well.

Sounds similar to what happened to my team. In our case, it was "only" two years of hard-won domain expertise, but we were also "trained" by the original founders (who had left at some point after the acquisition). The new team had zero domain knowledge and while we tried to hand everything over as well as we could, it seemed like a lost cause to me. A couple of months later, we found out our old company ended up buying the competition.

>So much so that the only people on the job market who had it were these folks whom they were ousting. All of whom had found positions at competitors within a day or two.

If they were the only people anywhere who had the critical domain expertise, then how could there be any competitors to snap them up?

This story doesn't sound internally consistent to me.

Not the only people anywhere. The only people on the job market.

So the skills in question, while niche, still enjoyed a supply/demand ratio that ensured that the structural unemployment rate for people who had them was 0%. And most companies know to get who you can and then make sure they stay happy.

Unfortunately, large conglomerates tend to be very credientialist. So leadership positions tend to be filled with people who have diplomas from schools with a lot of name recognition, rather than people who have enough of their own domain expertise to understand the product's business environment.

>on the job market.

That's a very nebulous term, as this whole thread bears out. I feel sometimes like my comments are abducted by aliens.

Perhaps we understand the word differently.

To me, it isn't particularly nebulous, it just means, "People who are currently looking for a new job." Which would roughly be composed of people who are currently unemployed, and people who have jobs but are looking for a new one. I suppose the developers in the story I relate would actually be sort of the latter group looking to avoid becoming the former group, but that's perhaps diving a little bit too deeply into the semantics for a conversation that doesn't involve someone playing bongos or a djembe.

I suppose you could expand that to include, "People who are happy with their job, but could be lured away if you offer them enough money." But it's pretty obvious that that's not a useful subcategory of people for the purposes of the company in the story I'm relating.

Not all companies need $500k engineers or can afford them. An optimization at Netflix that improves performance by 1% may save them millions of dollars. The same optimization at a medium-size company may only save them a few thousand.

Also, it's much more difficult to quantify the impact of a software engineer compared to investment bankers and consultants. The latter two are directly making money for the firm.

> Also, it's much more difficult to quantify the impact of a software engineer compared to investment bankers and consultants. The latter two are directly making money for the firm.

Yes, it is difficult, but possible.

Here's quick and dirty analysis - consider Sears, Walmart, and Amazon:

- Amazon pays its engineers near top of market - $1.5T market cap.

- Walmart has a pretty good engineering culture. They acquired a tech company to start Walmart Labs for data science stuff and they even have popular open source software (Hapi.js) - they're the largest company by revenue and $370B market cap.

- Ever been to Sears.com? Me neither - largest retailer in 1980 to bankrupt in 2018, stock price at $0.23 and $94M market cap.

I deduce that investment in software engineers == higher stock price.

I deduce that investment in software engineers == higher stock price.

Easy to do when you cherry-pick three companies.

Run your analysis on 10,000 companies large and small and you might have something.

I wonder if we looked at stock price / developers.

Would it prove or disprove this statemenet?

> Easy to do when you cherry-pick three companies.

* Netflix vs Blockbuster

* Uber / Lyft vs taxi cabs

* Airbnb vs hotels

* Blogs vs print media

* Youtube /Instagram vs TV

* porn tube sites vs porn studios

I could keep going.

I've also tried the corollary - companies that are successful while cutting cost in tech, but I can't think of any successful companies that are cutting investment in tech.

I meant quantifying an individual's impact. For i-banking that's easy. The more money an i-banker the raise, the better they are. There isn't a simple metric you can use for software engineering. Some terrible developers write a lot of code. Some amazing developers barely write any code.

Of course there is still meaningful signal you can use to differentiate developers, but it's going to be fuzzy. I-banking and consulting increase wages for their employees by firing everyone who doesn't meet a certain bar. In addition to creating a toxic work environment, this practice isn't possible with only fuzzy signals.

I mean, "treating your employees better" doesn't necessarily always mean "paying them more". For one, giving good employees more autonomy and decision power, treating people like adults, focusing on their career development, etc., goes a long way too.

> it's much more difficult to quantify the impact of a software engineer compared to investment bankers and consultants.

This depends to an incredible degree on what kind of "consultant" you're talking about. For the traditional "strategic consulting" type consultants it seems basically as impossible to quantify as for software developers, or rely most other people working in large companies.

I think as an employer you need to know when you have such an engineer, though, or be prepared to lose them.

I’d personally disagree with the cost effectiveness argument though. Unless your engineering work really cannot benefit from more skill (and as often as I’ve heard management think their work is relatively simple, in practice I’ve rarely seen it) then you’d have to replace a highly skilled engineer with a small team that would likely cost more than $500k. Packing your engineering firepower into as few bodies as possible is a highly cost efficient strategy, as most of the successful techs know.

On the other hand, a small team is less vulnerable to getting hit by a bus or hired out.

Investment Banks are a terrible example of low attrition workplaces, as attrition is quite high. It used to be the case (not sure if it still is) that expected attrition was ~100% after the first 2 years as an analyst as you were expected to go to B-school, the buyside or somewhere else.

Less of comparison of attrition, more of a comparison of:

* competing for talent by paying very well and wiling to do so

* investment in training early on

Two things that tech workers do not benefit from as much.

The whole discussion is anchored on retention. If you're not talking about that what is the point of your example?

You're asking that businesses think about long-term goals, efficiency, and what is best for the business. After having worked at over a couple dozen companies, I've never seen such an attitude from any business. Industry and size are irrelevant. Occasionally things line up by accident and the business does well, but every single business I've ever worked at or seen has a culture of shooting itself in the foot:

* No training (Not just for engineers but for anyone, including for specialty positions that simply don't exist outside the business)

* No focus on employee retention

* No significant raises / bonuses (Lucky to get a cost of living raise these days)

* No consideration for employees' goals

* Stressing employees unnecessarily with shit vacation benefits and no sick time

* Stressing employees with shit equipment

* Stressing employees with shit hours

* Stressing employees with shit office conditions

* Reducing benefits to slightly cut costs

The list goes on and on. Businesses succeed despite their own best attempts at hurting themselves, not because of it. Owners, executives, and other stakeholders are deluding themselves into thinking this isn't the case and often driving the business into actual bankruptcy or just below mediocre performance.

It's no wonder most employees are disengaged from their work. When you treat employees like shit, they will treat your business like shit to the greatest extent possible. They will do the minimum and it's extremely hard to change that course once it's been set. Employees are humans. Most businesses treat them like slaves, or machines at best.

There does seem to be a slight difference between stock-owned and founder/family-owned business, in my experience.

Not in my experience.

Only failing companies in 2020 view their engineers as easily replaceable.

That said, there are a lot of failing companies these days thanks to cheap and readily available debt.

> Only failing companies in 2020 view their engineers as easily replaceable.

Indeed.

You should engineer your systems as if everyone is replaceable (including yourself) but treat your employees as if they aren't (even though they are, just not _easily_).

That's simply not true. There are lots of companies in highly regulated industries that remain profitable and have low regard for engineers.

The optimal solution is for management to not let high-productivity talent from leaving - increase their comp to whatever offers they might get in the open market. That's how labor economic works - it's a market.

Life is not that simple, and cannot be boiled down into "it's a market."

People are messy. They leave companies for all kinds of reasons. They just broke up with someone and want to start a new life. They don't like the weather. The have to take care of an ailing relative. They've picked up a hobby (surfing, hiking, skiing) that is inconvenient where you are, or more important to them than money. Very often they just want a change.

It's common for HN-types to try to reduce everything to a numbers game, but people aren't numbers. They're not strings of attributes that can be quantified by an AI. They're human beings, and human beings will always be unpredictable.

While this is all true, comparing the turnover in the tech industry with other sectors - even high-paying sectors - our industry does seem a bit crazy, so it's worth figuring out why this is happening and what we could do to fix it.

Very few reasons for leaving aren't connected to an implicit "and I'm making more money." Very few people take significant pay cuts for reasons that don't boil down to "obscenely toxic workplace."

Money and Management are the real reasons people move, everything else is an outlier.

Even if you have the best compensation in the market, sometimes people leave just because they want something different. I agree with your general point that we should be willing to pony up to keep people, but ensuring your system is straightforward for other people to pick up and maintain is also a good practice.

The tech companies do invest a certain amount into new-grads. If you are a new grad in Microsoft working on most products you will get a weeklong training on how to build apps on azure (equivilant to azure level 1 certificate). You will also get invited to these pretty cool speaking events where they bring in people who have had TED-talks to discuss your transition.

> If you are a new grad in Microsoft working on most products you will get a weeklong training on how to build apps on azure (equivilant to azure level 1 certificate)

Do you actually get a certificate? That would be equivalent(and would actually be a good idea).

I think it shows up internally I completed the training. They say at the end of the course, that you should be ready to take the official test. I think you get a discount on the test, but you do have to pay for it.

That's good to hear and MS is doing the right thing, but this is not the case at many other companies.

They could probably do more to make it an integrated experience (manager had no idea these things were happening). But I think Microsoft is trying to make people feel like they belong. If these events weren’t going on only CS graduates from top 20 CS schools would feel like they deserve to be there.

"The optimal solution is for management to not let high-productivity talent from leaving - increase their comp to whatever offers they might get in the open market."

That might result in the engineer being paid more than the value of their contribution at their current employer.

Many things could cause this, e.g.

- the new employer is much more productive, can make better use of the engineer, and so can afford to pay more and still make money

- the new employer over-estimates engineer's skill level and likely contribution

While this is possible, it's relatively unlikely. Top performers can generate enormous amounts of value for a company. As a hiring manager, I absolutely believe in the existence of 10x engineers. The reason the term has been so mired in controversy is because folks think the 10x part refers to raw coding ability, but it doesn't. It refers to soft skills, communication, leadership, domain knowledge, business acumen, customer relationships if relevant, and dedication -- overall impact on the company. It's actually quite easy for one developer to have 10x as much positive impact on a small company compared to another.

As a hiring manager, though, I'm stuck in the middle between developers and executive leadership. Executives tend to form a notion of what different roles should be paid, and they just won't bend, even if it means losing talent and spending $1 million+ replacing them over time. The problem is that replacement cost is abstract and easy to ignore, whereas paying a single developer $500k/year to retain them feels very real. Never mind the fact that losing them means multiple projects will be derailed, you'll spend a year looking for a replacement to hire at twice their salary, that replacement will take 3 additional years to reach the same level of domain expertise, it turns out that replacement never takes on the same degree of leadership or earns the same level of trust from colleagues, and in the midst of all of this the absence of your top performer causes two other talented engineers to leave in the following year.

$500k starts to look reasonable in retrospect, but you'll never be able to sell it, so you'll do the above dance instead and spend more money in the end.

The issue is we're not able to quantify the costs of technical debt and lack of domain knowledge.

We rewrote something in React and adding a feature to it took 2-3 hours but doing the same to the legacy Java code took 3 weeks. This cost was easy to quantify and strengthened proposals to remove the legacy Java code.

I was literally asked "Do you think you are worth $x?" in an exit interview. Obviously I do, since that was the offer. What a strange question.

> Do you think you are worth $x?

Yes, actually $x * 20%.

Right? How much am I worth? How much do you have?

>> The optimal solution is for management to not let high-productivity talent from leaving - increase their comp to whatever offers they might get in the open market.

>> Investment bank and management consulting figured this out a long time ago.

That only works if your business makes piles of money, like investment banks. Or FAANG.

The rest of us have a % of money to spend on talent.

The rest of you waste more money hiring 5-10x more mediocre engineers to try to mitigate the bus factor or whatever. And it's still not as good a product as it will be built by a handful of good, over paid and over worked engineers who really know what they're doing.

Most "over paid" engineers are smart enough not to get in a position where they can become over worked :)

That is a hypothesis that may be true in some cases. In other cases it may be better to have two redundant engineers at 150K than one at 500k.

The idea that there is a single solution is absurd.

> The rest of us have a % of money to spend on talent.

If the % of money allocated to talent is so low that the business cannot survive, then the business has no intrinsic right to survive. One possible exception to this would be a mission-driven non-profit, but that's not what we're talking about here.

If the business can't turn a profit and pay engineers at the same time, of course they shouldn't survive, but I don't think that is what they were saying.

There are plenty of viable business solutions that don't require paying every engineer 500k+.