Surprising how the BLS numbers seem so far off from what's seen on Levels.fyi. Possible explanations: they only factor in base salary (no RSUs or bonuses), the bay area employs a proportionally small number of SWEs (such that the median is reflective of other, cheaper areas), or non-traditional jobs are getting grouped with SWEs as a single career category.

Levels.fyi is heavily biased toward companies where tech is their primary business. There are a lot of software engineers working in support roles for "old line" businesses that don't value or pay their software engineers nearly as well.

My wife works for a charitable foundation, and continuously complains about the quality of their software. Once she asked her IT director why they can't get someone like me in to just fix it, and her IT director was like "We can't afford someone like your husband."

> Levels.fyi is heavily biased toward companies where tech is their primary business.

They are also (AFAIK) based to a large degree on self reporting which you have to imagine would skew high. If I had an amazing salary I’d want to self report it and tell levels and HN all about it. If I have a normal salary, maybe I keep that to myself.

Anecdotally I’ve worked at two companies featured on levels.FYI and my total comp was nowhere near what they were reporting for similar roles... like 50-66%—they were waaaay off. Either I am consistently grossly underpaid or levels biases high.

Your issue might be the “similar roles.” Software engineer numbers on levels is very accurate, but don’t expect a similar role to pay similar, they’re not comparable.

Hi, I’m one of the creators of Levels.fyi. You’re right that there’s definitely likely a bias in the data we collect. But in terms of validation and verification, I just wanted to provide some info on how we source.

Whenever possible, we collect verified offers / pay stubs / W-2s to anchor the data on the site. While our self reported data is definitely the more popular choice for submission, we are able to determine general bands and find blatant outliers for certain companies and remove them. I think we might be the only consumer facing salary site to do this afaik.

Some more info: https://levels.fyi/verified

We also do provide a community reporting feature much like the flag feature here on HN, which we actively monitor.

Even after all we do, we are after all only as good as the data we collect. The more contributions we receive the more we can color in the whole picture. Feel free to email us any suggestions at hello[at]levels.fyi

Thanks for responding. Generally I find the equity values to be way high. Do people submitting their info take into account vesting? I’ve talked to (otherwise smart) people who make claims like: “My salary is $120k and my equity offer was $100k, therefore my total comp is $220k” where obviously he’s wrong because he will only vest, for example, 25% of that equity the first year. So he should consider $145k his TC.

A lot of people seem to make this mistake, so hopefully your data cleaners have solved that.

Yeah, we do specify that the values we collect our annual on our form. But we might be able to make it more clear.

We had this issue earlier on in our first 2 years since our form was much more rudimentary – could it be that the data points you're visiting are from 2019 or earlier? Also in any case, feel free to report those entries, and we'll make sure to take a look.

Levels.fyi is basically spot-on for my company, despite it having a relatively small # of reports, and is also accurate for friends of mine who work at larger companies (FAANG). But, as another reply said, some places will have vastly different pay scales for "similar roles" to SWE - for example, it's known that Facebook pays Data Engineers less than Software Engineers at the same level (mostly by giving them fewer RSUs).

If anything, I would say Levels.fyi underrepresents average total comp for many of the listed companies, because: 1) they don't take into account refreshers, where given, and 2) they don't take into account stock growth (which, yes, is not guaranteed, but it wouldn't be _unreasonable_ from a perspective of estimating Expected Value to assume whatever RSUs you get will track overall market growth, just with more volatility)

Levels should probably have a way of filtering submissions from employees with 0 YOE and 1+, 3+ YOE, as you mention, RSU value can double. So another identically leveled SWE could be making double you, just based off the stock appreciation. Naturally, one would expect the company to offer a package with half as many RSUs after share price doubles.

Since you always enter a company with 0 YOE at that company (or at least, a clean slate of RSU tranches), the data points of employees who 'got in' early aren't relevant to an offer you might receive.

Agreed that that functionality would be helpful, though I haven't noticed a huge number of offers with details that would indicate "stock growth" rather than "initial offer + raises". They're there, but they're usually easy to pick out because they're way out of band. Might affect things on the margin, I suppose.

Hi, I’m one of the creators of Levels.fyi and I just wanted to mention that we actually do have this feature. You can filter our compensation table by # of years of experience or years spent at the company by using the filters next to the search.

https://levels.fyi/comp.html?showFilters

This is also available on our level specific pages, for example for Google L3: https://www.levels.fyi/company/Google/salaries/Software-Engi...

Though I do agree it would be really nice to have this reflected in our visuals or aggregates where possible.

Hey, big fan of your work :)

Love the new reporting feature, way less friction than sending an email.

Do many SWEs not know that IT department staff are paid way less than them?

levels.fyi is not really a good representation of income for even the Bay area, let alone the industry as a whole.

I don't know why people get confused by this; most people working SWE in the bay area will never see a vaunted Levels.fyi equivalent salary (although most of them will exceed national averages for software).

Basing your understanding of software salaries across the industry on levels.fyi is a bit like basing an understanding of national lawyer salaries on a sampling from top NYC corporate law firms.

(NB there is some sprinkling of other useful salary info on levels.fyi, but people focus on the FAANG and similar companies, and the data is much better there)

People conflate the fact with simply desiring that sort of salary; pointing to Levels provides the justification for demanding top-of-market wages. It's not that we deny the existence of lower-paid SWE's, we just place less emphasis on them when evaluating our financial ambitions in the context of our careers. (If I'll be working 40+ hours regardless of company, might as well maximize wages).