The interesting part of the controversy about including compensation ranges in job postings is about its intended effect on salary discussions. The question is, would making job salary ranges public help people who decline to read a single book on basic negotiation? The question should probably make my views on the issue fairly predictable, but noticing how uncharitable it was got me thinking about what other solutions could look like.
The problem itself isn’t well defined, as while there are many criticisms and anecdotes about how salary negotiations disadvantage perenially disadvantaged people, almost none of the popular articles I have read really talk about what a formal negotiation looks like. There are parts, principles, tactics, outcomes, methods, academic journals, conferences - it’s a really sophisticated domain.
If you haven’t read a book on it, you might think bargaining and haggling fall into the same category, where rounds of tit-for-tat bids of up and down ending in an agreement to split the difference is how people discover prices. Auctions and markets do work this way, but if your salary and contract rate discussions go like this, you are not negotiating, you are just auctioning yourself off. It’s difficult to discuss the topic with anyone who hasn’t at least read one of “Getting to Yes,” “Never Split the Difference,” “Getting Past No,” or even the classic, “How to Make Friends and Influence People” because unless you have a conscious sense of what your downside or best alternative to a negotiated agreement is, or even some thought about what a high quality agreement looks like, you don’t quite have an oar in the water.
Specifically in regard to posting salary levels in job ads: My objection is that it lets employers signal to each other and create noise in “market rates” that drives down wages, both for the talented people who create the growth and opportunity in businesses for the less skilled people, and fewer opportunities for people who just need any job they can get to get by. The public sector and academia are probably exceptions in this because their hiring still operates more like that of guilds and confraternaties than normal markets. You may notice there is tons of data on individual contributor salaries but conspicuously less on management compensation. Most salary websites seem to be self-reported data, and in decades of contracting I have never once got through a recruiter conversation without them volunteering something like, “yeah, consulting rates are down this year, real pressure on rates, we have to bid conservatively…” It’s a zero-sum game for them. I have worked with probably the best and most upstanding talent vendors in the market, and I exclude them from this criticism, but usually if you are on a call with a recruitment agency, these are people who have already agreed to supply your work at a fixed rate, and all they are doing is trying to knock down your rate so they can keep the difference as their margin. I’m the least Marxist person I know, but this tension between capital and labour creates a lot of bullshit and noise in the market that is intended to isolate and discourage individuals and keep wages down. The sentiment behind advocating the posting of salary ranges with job ads is one I sympathize with, even if I don’t agree that it is a solution.
However, I do think there may be a better future than proposed-salary disclosure in job ads and the Blind-man’s Bluff game that gets played out in most screening calls and job interviews.
To get a sense of what a role pays without creating market collusion to drive down wages, the solution is to change the denominations of salaries from familiar decimal numbers that create arbitrary impressions about what the zeroes in them mean, and express them as numbers of a size or magnitude expressed in bits. People outside tech won’t have an intuitive sense of this, but I will show how it is a powerful enough tool for reframing how you understand compensation numbers, so that it may be worth adopting.
For example, if I told you I made “six figures,” that’s an arbitrary psychological line that puts my value above most hourly wage jobs that aren’t unionized, but it’s also a range between $100k and a dollar short of $1-million, a life changing amount of money for most people. It’s somewhere between the bottom number, and literally ten times that number. Six figures is no information, and arbitrary decimal thresholds are literally meaningless psychological noise.
If I published a range for a job that paid between 50,000/y and 60,000/y, the subtext in that is “don’t even think about asking for more, we’re not growing or doing anything anyone wants, your value is measured in small increments, and someone else will do the work for less.” It doesn’t matter what the number is because when someone tells you what you are worth, if you believe them you have already given up your bargaining power and belief in your ability to make something valuable. This demoralizing dynamic is covered in most negotiation books and it doens’t benefit from elaboration here.
However, let’s say I advertised a job with a 17bit package. What does that even mean? Nothing unless you can do powers of two in your head or have a calculator handy, but what it does yield is five things:
the lowest possible floor number is $65,537
the biggest possible ceiling number is $131,072
the next level up is an 18bit package, which has a ceiling of $262,144
the path up the career ladder has clear incremental steps, and it is non-linear
a more precise magnitude with a growth trajectory, but with privacy that allows reasoning without enabling salary level collusion
Four and five could probably be combined, and I might add that these levels include a fair diversity of roles, where if you learned someone was in a 17bit role, it doesn’t tell you much else about them, and if they were in an 18bit role, the relative differences in possible compensation are so large as not to be really comparable - and this achieves much of the social good of salary disclosures, where instead of giving people a way to envy and compare themselves to others it just gets blurred out, and without creating collusion and wage mean reversion. The more-vague bands emphasize what people have in common, and avoids the narcissism of small differences.
If I’m interviewing for a 17bit role, I know I’m pretty much in the upper middle of that range, and that gives me alot of room to negotiate with questions like, what do I need to do to level up, are there non-salary or other benefits that make the total package value higher, are there incentives for doing well. I can ask things like what puts my skills at the lower end of that range and how much value does this role add and are there ways to increase that? With narrow salary bands, the relationship is much more paternalistic, and in collective bargaining environments, haggling over small stakes makes it even more adversarial.
If I am advertising a FAANG SRE or staff level developer job in the Bay Area and I say the whole package with health and RSUs is 20bit, it signals to all candidates that the role is already at a competitively high floor (starts at $524,289) while leaving the ceiling for that role so high that it is open to discussion and growth, because opportunity for growth is the biggest incentive of all at that level. It’s a signal that they are growing. Compensation above those levels typically becomes a proprietary company secret or sensitive compensation information, unless it is to executives and company officers in the public filings. In senior management roles, 21bit to 23bit packages that include equity and bonuses are common in leading tech and finance companies and powers of two adds some arbitrary granularity for reasoning about what are really huge numbers.
This is a blog post sounding out a new idea that has been rolling around for a while in one way or another. Obviously I have opinions and biases, but some of them come from unique experiences as a tech industry contractor which provide a rare view into a large number of diverse organizations. Salary disclosure pops up in one way or another across industries and market eras, but changes in tech and education mean that we are no longer limited to low sophistication ideas.
Bitwise comp negotiation may be an opportunity for something new.