Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I can tell you as a manager the salary bands are generally set by HR for the whole company. I was told point blank at a previous employer that we payed the 70th percentile. I asked how they knew what that was - apparently there are companies that aggregate this data (legally?) and sell it back to HR orgs for $xx,xxx+ a year. HR for some reason wouldn't deviate from this band for top talent (discrimination maybe?)


I remember many, many years ago my boss at IBM explaining that I did not qualify for a raise because I was just over the industry median pay for my band. Honestly, if they had just kept giving me 3% raises every year, it might never have occurred to me to wonder whether I could be making more elsewhere.

The weird thing about paying at the 50% line is that it means that your policy is, in effect, "we do not want any above average programmers."


Average inflation over the last 20 years is 2.1%

Honestly being given anything under 2.5% is just insulting. Pay freeze or COL adjustment in line with inflation is for a few years in bad recessions imo, like 2009-10 or, probably the upcoming 12-18 months.

Being given 3% is in the neighborhood of "better than nothing", but nowhere near what it should be for an employee trying to get ahead and putting in the required effort.


Yup. Our team is only allowed to have one person with "Exceeds standards." One. A bastard form of stack ranking. So the mgr picks based on his gut feelings, and then justifies it with a vague review.

The lucky team member gets a 5% raise, everyone else is usually ranked as "Meets standards" which merits a 2.5% raise. A few people who need time in the penalty box get a 1% raise.

So what happens is your actual performance doesn't matter unless you make the boss look great, or look terrible. He's a non-tech person, so he can't effectively evaluate you on any other criteria.

This leads to employees doing enough work to avoid being fired, or (as is usual with Stack Ranking) sabotaging other workers.


Even 5% is nothing to write home about. Still far under the amount you would get if you switched jobs.

But then you’d have to run the interview gauntlet again, and frankly that’s just so distasteful that I don’t want to.


I'd take 5% in a heart beat. But I'm not willing to run the interview gauntlet, especially as I get older. There's age discrimination, and everyone wants/expects you to be an expert on ephemeral technologies.


Compensation data from companies like Radford comes with three challenges

1. Your internal leveling scheme for SDEs needs to have been sorted out. External titles like "senior" are pretty meaningless, so one reasonable solution is mapping SDEs to an internal numerical tier so you can differentiate between "senior" and actual "senior".

2. Radford data comes with a trillion job titles. You need to identify the relevant ones and how your internal tiers map to them.

3. In markets where compensation is exploding (Seattle a few years ago), compensation data is a trailing indicator of what market rates you need to pay for new hires. You get fresher, but noisier, signals from candidates who reject your offers and from your personal network.


As others have mentioned, Radford[1] is one company that sells that sort of data. Many providers of HR management and payroll systems also have secondary revenue streams by repackaging that data for uses like that. This[2] is one offered by ADP, one of the largest payroll providers in the US.

As a manager, one way to get yourself some wiggle room in the situation you mentioned is using that strict policy against them. Request as detailed a list of roles and compensation bands as possible[2], then creatively interpret the job duties such that you can justify opening the requisition with whatever role provides the most favorable "70th percentile" salary band to attract the talent you want to bring in.

Success varies between companies, but it's not uncommon for larger companies to have latent roles/titles set up in the system that are rarely used, but exist in the benchmark data and have far more favorable salary bands than the roles you'd typically use for your reqs. Or if you can get the full benchmark data itself, you can look through it all and try to get HR to add a net new one to the system based on what you find.

[1] https://radford.aon.com/surveys

[2] https://apps.adp.com/en-US/apps/79131/benchmarking-powered-b...


I worked at a large insurance company in a previous life, this was their stance as well.

They will deviate if the field is hot enough and it translates to money for them. They were paying FAANG money to AI/Data Science fresh out of college.

The McKinseys/Gartners/etc of the world talk about the coming advantage...exec hears it and they become more open to these "critical sectors"


Yes those salary survey's are just a way for businesses to enhance the information asymmetry in negotiations.

I think if software engineers ever decide to collectively bargain, and they want to include comp as an area to bargain on, the easiest thing to do is buy these same datasets and share with members.


Yep, Radford is one of those companies.


IIRC, that's the one they used. This struck me as kinda crappy and anti-competitive. Perhaps this is too idealistic, but I would prefer a more transparent and open market for talent.


OTOH the sectors where compensation is probably most open and transparent in the US such as public sector tend not to be the best-paying.


This is common practice. There are many salary surveyors that will collect, anonymize / aggregate and sell back the data to companies. At Levels.fyi we're trying solve issues that typically plague these salary surveys (ex. data staleness, leveling / title normalization, etc). We do this by flipping the model and collecting data directly from employees. We've also mapped levels to a fluid range rather than rigid step-wise increases. Crowdsourcing has it's own challenges (validity, normalization, etc). That said we've had several high-profile employers sign up for our new data subscription: https://www.levels.fyi/offerings/


Bands are usually per level - if someone is sufficiently good you bring them in as a Principal/Staff/Magic Unicorn/whatever.

This leads to the same phenomenon where you need to work for 2 years to get promoted but if you jump ship your new manager will set your level in line with your salary expectations. And of course the levels are largely meaningless outside of compensation and office politics.


I've noticed this exactly.

At my original company I was performing Senior-level work (like, literally according to the staff level formalism from HR). But I was told by my manager that to be classified as a Senior-level staff, you have to have been performing Senior-staff level work for at least a year. Therefore being promoted within the organization is a long process. Moreover, with the usual 1-3% raise, you're still in the (low-end) of the pay band for Senior staff, so from management's point of view there's no real problem.

.. and then of course, if you're hired in, you just have a few phone calls, couple of interviews, discussion with the hiring manager and you're at a pay-level and job-level far ahead of existing employees even with identical experience.


This exact scenario happened to me - I was a "senior manager" by HR title but I was actually just promoted to team lead in practice. I meant to illustrate the setting of compensation happened almost entirely without engineering input at a company level (in my experience, where I worked, etc.).


Damn! This explains why my company has so many senior managers. Sometimes with just one report :P




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

Search: