Hacker News new | past | comments | ask | show | jobs | submit login
Silicon Valley’s Exclusive Salary Database (wired.com)
190 points by ax00x on July 1, 2018 | hide | past | favorite | 70 comments

Glassdoor dramatically underestimates salaries. My position is listed as about half of what I'm actually paid (and I'm "average"). Paysa and LinkedIn are accurate for my position, but I think both generally tend to underestimate most people's total comp as well, particularly for software engineers in the Bay Area. Honestly, as bizarre as it sounds, comp info from Blind seems to be the most accurate. Levels.fyi is fairly accurate as well, but only if you look at the newest entries on the spreadsheet (and not on the level averages on the main page).

I’m from Levels.fyi and one thing we’ve found (seemingly obvious) is how specific skill sets can dramatically affect comp numbers. Ex. AI / ML is the hot stuff these days. To this end, we’ve added a field to collect ‘focus’ for each salary submitted. P.S. there’s also a quick link for getting straight to the salary spreadsheet: www.comp.fyi

Glassdoor is old enough now their data comes from 10+ years ago, and the bulk of their salary numbers come from the depths of the great recession. It's no wonder their numbers are out of wack.

I also highly doubt they adjust their submitted salaries for inflation.

It wouldn't make any sense to adjust the salaries for inflation would it? Only if salaries were moving up with inflation, and they're not. Haven't been for decades, I thought that was across the board.

Yup. I make 3x what glassdoor lists for my company. So do at least 2 other senior engineer friends of mine at the company.

For top paying companies (goog/fb/netflix/uber...), here are the total compensation ranges i derived from various Blind threads:

3. (entry level) software engineer : $175k-$225k

4. (experienced or phd) software engineer : $225k-$300k

5. senior software engineer: $300k-$400k

6. staff software engineer/engineering manager: $400k-$550k

7. senior staff software engineer/senior engineering manager: $550k-$750k

8. principal engineer/engineering director: $750k - $1m

As a data point, my own comp. is above the corresponding range here due to stock appreciation.

I would agree with the caveat being that most of these ranges are for West-Coast based companies, in particular, the Bay Area.

Looks about right for software engineers. For data scientists and data engineers subtract half the stock.

wow. can you include base + bonus + stock info here?

I am currently hw/sw engineer at AMD austin (rank: Staff Engineer) and I get paid 140K base + 80-100K stocks. Yearly refresh is 10-20K. This data makes me cry.

At least we can buy a home for <400k here in Austin. Still, wow.

Senior Engineer example: $180k salary $30k cash bonus $150k stock

senior engineer is how many years of experience out of school? with/without PhD? area?

wow. that is a lot of money. is that for the bay area only?

Bay Area + Seattle + NYC + LA

My take is very different from the Wired article: I think Option Impact is a great resource. Anytime an early stage founder asks me about salary/equity comps, I tell them to check out OI.

If a hiring manager wants to use this data to gouge employees then that's a problem with their ethics, not OI. The place where OI is incredibly useful is for making fair offers, especially for roles you're unfamiliar with. Maybe your company has 6 engineers and you're about to hire your first PM. What should you offer them? A market salary would be $X. You are not trying to figure out X so that you can make a lowball offer; you're trying to figure out X so that you don't offer 1.5X (which hurts your capacity to make future hires) or 0.5X (which might come off as stupid or insulting to the candidate). OI tells you what X is.

Data transparency is a good thing. It helps people avoid dumb mistakes. To be honest, I think it would be great if employees could have access to this data, too.

I think much of what people are objecting to is that, by virtue of its position as an _invite only_ paid service, it provides insight to only the employer party of a two-player game.

It is apparently purpose-built to allow only the employers to arbitrage that extra insight into hiring patterns.

It would be great if OI could let employees see this data.

It would also kill OI's business model: provide a semi-legal mechanism for hiring-collusion among the giants and their wannabes.

It sounds like collusion. I guess the anonymized profiles helps mitigate that.

It seems weird that Wired focuses on the negative edge case here. At Vicarious we use OptionImpact to make sure everyone is paid fairly, and I imagine that's the vast majority of users. Deceiving someone into taking a lowball salary is a very shortsighted way to run a business, and I have trouble believing any well-run company actually does this.


If you are running a start up and you're using this to fuck over your very first employees' salary then please take a long hard look at yourself in the mirror. You are decieving the very people you are hiring to make your business a success. Are you really cut out for this kind of work if you deliberately fucked over your first and likely some of the most important employees in your business? Just to save a tiny fraction of some VC's money?

With respect, I think if this is you then you may as well shut the startup now and save everyone the time and pain and money that will inevitably come later (either due to failure stemming from your questionable ethics and questionable business acumen evident from your hiring practices, or when your employees/engineers #1-10 realised you conned them and they leave for someone else and leave you in the lurch at some critical point). It's not going to be a great experience for anyone.

Even later stage companies should wise up. It costs a fortune to recruit talent into a company. Any small bumps in pay you can offer to increase retention is a good idea.

Underpaying based on an information advantage, as a strategy, seems very self-defeating. The reason is that engineers who figure start to suspect that they are overpaid are unlikely to go confront their managers about it - they will first test their value by applying for other jobs. Once they get a much better offer, they are likely to leave rather than use it to negotiate.

I've wondered why this happens so much, and I think some of it comes down to a personality difference between management types and many engineering types. Managers and other negotiating types aren't uncomfortable with the process, to the point where they are often bewildered that someone would go through all of that (leaving the company) when they could have just asked for more money.

Your points notwithstanding, “some VCs money” could not be more wrong.

"Come here and belong!" is the startup siren song.

In my experience, even YC startups don't actually care for their employees outside of their contributions to the team. The only things that matter are the customers, the investors, and the founders-- and sometimes not even the customers.

The ideal of a small company worrying about its rank-and-file, fighting "evil" giant megacorp is frankly just a useful ruse to buy developers on the cheap. Why pay a wage when you can give kool-aid?

Yeah, this is happening, we had to know it would. Imagine if real estate transactions weren't recorded publicly - a small number of real estate agents would create a private database of all transactions, while everyone else would be limited to snippets of conversation and bits of unreliable info posted here and there on the web. The disadvantage to being an outsider to the information would be overwhelming.

This is one reason why I favor making all salary information public. I know it freaks a lot of people out, but I work for a public university where all salaries are public, and while the earth shattering problems I hear about haven't materialized, the information it brings to the workforce is very positive (you can actually see what the university pays your future colleagues when you weight a job offer). The main disadvantage is that our info is public but private colleges keep it, well, private, which probably does put the publics at yet another disadvantage (Stanford can easily find out how much all the profs and staff at Berkeley and UCSF earn, but not the other way around).

If this happened in the private sector, I think there would be a bit of a well deserved reckoning, but far more good would come of it than harm. If a blanket requirement would be too intrusive or unconstitutional, I think the gov't could make it a condition for H1B visas, which does have a nice consistency to it (if a corporate HR dept is going to be empowered to decide who does and doesn't get to work in the US - a massive power usually reserved for government - then it's very reasonable to ask for a high level of transparency).

By this, I mean that if a company requests an H1B visa, they must make all salary info public the way public institutions are required to, not just for the particular job, and not anonymously.

I've also felt that Glassdoor, across the board, very much underestimates real salaries, and this further supports that. Anyone can think of a reason why this happens? One would be that they have old, historial salaries in their averages.

My guess would be that the best paid employees have the least time and interest to participate in Glassdoor. Especially as you progress in your career, stack ranking against generic SE levels becomes less relevant. For instance I wouldn't write off a company just because their reported "senior software engineer" average salary was below my requirements, because I know there are many reasons they could be willing to pay more.

Isn't Glassdoor across all of US whereas this is probably mostly SV based companies?

You can filter down by city. I do agree even then it's underestimating but if I can be arrogant for a moment, I believe people who read HN are accustomed to slightly higher compensation than the average

The median salary for software developers nationally is $103k per BLS. According to Glassdoor it is $104k in SF. Do you really think ordinary developers in the Bay Area are getting paid national median salaries to live in an exceptionally high cost of living area? I think even the least among us are smarter than that.

If you put in, say, $0 for your stock compensation, even though you receive stock compensation, it will average it in with equal weight as entries which included actual values for their stock compensation. The same applies to cash bonuses.

That plus, as you mentioned, including old, historical data with equal weighting to newer data will drag down the averages for both salary and total compensation.

glassdoor is packed with fake reviews. why wouldn’t it also be packed with fake salaries?

How has this not been leaked yet?

Agreed, anyone who has access to this and hasn't leaked it is a collaborator.

Don't be this guy, employees will figure it out.

> He recalls walking away from negotiations with a newly hired employee thinking, “Man, this guy is underpaid—we got a good deal.”

Yes, calling the start of a relationship a "deal", frames that relationship as transactional. That plus bragging about underpaying him simply means "I'll stab you in the back the moment you look away"

Wait, an employee-employer relationship isn't transactional? I agree with you that bragging is a bad look but I also agree with an above commenter who says that the point of using OI is that you're trying to find $X to avoid outlier mistakes.

But I think it's naive to expect that salary negotiation isn't adversarial. Both sides are trying to get an optimal deal. Employees making data-driven acceptance decisions is better than "employees figuring it out" later.

A transactional relationship is fine if the employee is not integral and will remain not integral throughout their employment at the company. But if a founder is hiring the first few engineers, or executives, or anyone of importance-- he or she should be thinking about their futures: in particular, how they might feel about being low-balled vis-a-vis their peers.

"If I pay this smart, black, female engineer less than every other male engineer, just because she doesn't have a college degree, how would that look if everyone found out?" [1]

"Adversarial" is expected in a negotiation. Short-sightedness is not. "Bragging" about stiffing staff isn't just a a bad look, it's a sign that the dealmaker is an animal in human skin.

I'm sure 10% of OI's users have good intentions. The rest are living examples for why hiring collusion is outlawed and subject to prosecution.

[1] Yes, I worked for a startup that did exactly that, and was surprised when the engineers started quitting weeks after the truth came out.

I don't think any of your concerns are necessarily (important qualifier) driven by use of OI data on the employer side, and the use of the typical salary estimate services on the employee side. Of course there are going to be bad actors and toxic workplaces where the data is being used to lowball. I don't deny that it happens. I don't think those that do are going to be able to survive long, because once the word is on the street that a company is doing that it's gonna be very tough for them to hire talent competitively.

FWIW, when I mentioned "bad look", I meant it in the extended sense that you're talking about here, so I agree that employers should never do that, and even that it's a warning sign. I think it definitely impacts the culture of a startup in a negative way and it's a horrible way to treat people.

if you have access to this, it's your moral responsibility to leak it

The most interesting headline today. The article has very little information. Hope to see a followup.

How is this not considered collusion?

Collusion is illegal cooperation. This is legal.

How is this legal? Let's take a hypothetical example:

Joe is employee #2 FoolzRUs, at a Sequoia-backed startup, in CA. The company is small, having 10 employees total, and maybe 6 engineers. Joe is interviewing to get another job, because he feels that his company is not going to last. The prospective employer happens to be a growing company that was funded by Sequoia. The recruiter looks up FoolzRUs in Option Impact and deduces that the Principal Engineer at FoolzRUs sounds like Joe, and thus, his salary. Now she does not have to ask Joe what his salary is, but is she violating California's new law, where an she is barred from asking Joe what his compensation is?

She is violating the law. California doesn't allow seeking his past salary data in the hiring process. This is true even if they don't seek it from him.

Of course, they can use general data that they think matches his experience. But not if they're using that as a way to dig for his individual past salary.

Yeah exactly. That long wink while shaking hands is a breaking of eye contact, so no explicit agreement has been formed.

good headline but only useful for startups and i suppose it comes “free” with VC money. regular companies share salaries with an aggregator (i forget the name) and it’s repackaged for sale to other participating companies. USA-wide at least.

I think the aggregator/sharing company is owned by equifax. there might be more than one of them.

This is all about symmetry of information. Which I want to remind everyone that during the who's hiring posts we need to consider. I exclusively upvote companies that post salary ranges. I know some users even downvote those that don't. Plus posting salaries saves everyone time and effort.

This sort-of reads like an ad...

The databases (Glassdoor, etc) I’ve seen are pretty close on base salary. There are outlier mistakes but generally close. They get some of their data from H1B applications. This misses bonuses and equity, and can be a little lower than non-H1B, but the numbers I’ve had first hand knowledge on are 100% accurate.

How would "steve" as a engineer have access to something that presumably would be jealously guarded by the hr director.

Startup. Logins are routinely shared for third party software.

Even some this sensitive - that's just asking for trouble

Yup. It’s not uncommon in the slightest.

Why do data scientists have such a high salary? Everyone seems to be one at the moment.

They don't at all. At typical companies, their salaries are maybe a bit higher than software engineers. At top companies though, their base and bonus is the same, but stock is about half (unless you're a core DS or have an AI/ML PhD). Also, 4-5 years working on a PhD could have been spent gaining experience as a software engineer, which means that all total a data scientist at FANG will make $100k to $250k per year less than a software engineer at the same age.

do they? in the article 136k w/ 4-6 yrs exp really isn't a lot, especially since it has a higher % of PHDs.

i became a data scientist because I love that type of work. I like solving puzzles and diving deep into data and helping to drive decisions. I imagine others want to be / or have become a data scientist for that reason.

Where do I get this?

> the daunting task of hiring a team, from low-level engineers to a new VP

I don't think "low-level" means what the article's author thinks it means.

How do we shut this place - Option Impact - down?

Ideally, regulators would, but we live in a hyper-capitalistic age, so they probably won't.

I mean, I don't know if I'd want anyone to shut it down. Just somebody leak it and the playing field is evened out again (and from the sounds of it, we all get big raises after switching jobs).

Don't shut it down, make it open so we can all benefit. But I imagine no one would update it.

From a data standpoint, using such a resource to determine salaries would like cause salaries to become predictable, based solely on what other people paid for the combination of experiences used to classify the salaries (Ruby programmers with 5 years of experience in startups of size Z get X). So salaries would tend to not diversify much from the median (perhaps tend down a little).

That also neglects the quality of those 5 years of experience as it relates to the actual position. 5 years of designing financial services systems is more valuable to a financial services company than 5 years of building web applications for a dating company.

That is the problem I have with collective bargaining in higher skilled fields — it assumes people are interchangeable based on worthless “years of experience” metrics. Some sysadmin with 20 year experience might not know the first thing about DevOps, while a fresh college graduate might be an expert. Yet the 20 year guy would get paid more despite knowing less.

The problem of outliers not being able to fully maximize their earnings under collective bargaining is a substantially smaller issue than everyone being underpaid without collective bargaining.

It’s a salary database and it’s anonymized. How is this harming the public? A worker is still free to accept or reject any offer they want. Companies aren’t colluding, they are reporting. Nothing stopping companies from paying more or less than what other people are paying.

Because when one side of the table has access to information (especially such detailed and important data such as this) it shifts the negotiating advantage strongly in favor of the companies.

So they should shut it down because someone didn't negotiate a market rate salary for themselves?

This is laughable. Especially if you're an American.

This is systematic manipulation of the market rate, which is far more insidious and bad than someone merely striking a bad deal for themself.

Reporting isn’t manipulating.

Asymmetric information is

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact