4 ms·

This article is a few years old so I hope it has evened itself out by now. I wonder why an 'algorithmic' approach to salaries is not more common.. To achieve t

This article is a few years old so I hope it has evened itself out by now. I wonder why an 'algorithmic' approach to salaries is not more common.. To achieve this there must be a structured career ladder with standardised job descriptions and (most likely) a way to determine a maturity within the role itself, but then it's a case of having a base salary (often heavily influenced by where the employee is) and some multipliers. Out pops a number (or a tight range) and the employee can choose if they take it based on their valuation of their skills. No risk of gender bias and it takes away a lot of the pain of negotiating salaries felt by both management and subordinates. Regularly reviewing the base numbers to adapt for inflation and market trends will ensure that both new and existing employees are treated fairly. At my previous position (head of engineering at an old school sales house trying to become more engineering friendly) we had great success bringing in a structured approach to growth and a corresponding salary calculator. Unfortunately the salary numbers were reserved for management as it wasn't rolled out across the whole organisation but it inspired other, non-engineering, departments to adopt a similar pattern. My vision would be to have these figures open company-wide however I understand that in most medium-large companies this is a time-consuming process. As someone starting a business today I want to ensure that salaries are open (definitely within the company but likely to the world) from the beginning. Employees should feel comfortable knowing that they are compensated fairly for both the work they do and in comparison to others. If anybody is looking for some open inspiration about how this is done, check out Buffer[0] and PostHog[1] (Gitlab[2] have one too but the numbers are not open). - [0] https://buffer.com/salary-calculator/software-engineer-ii/high - [1] https://posthog.com/handbook/people/compensation - [2] https://about.gitlab.com/handbook/total-rewards/compensation/compensation-calculator/

Leherenn 3y ago
Because you give up a lot of flexibility when hiring people. If you want to hire someone asking even just a tiny bit more than what the algorithm says, you have to adjust the algorithm and increase everyone else, play with the role level to get what you want or just say sorry not possible. My experience is that it's often the second one. "Yeah, he's more like level X than Y, but we need someone quickly and he's the only one in the pipeline, so let's bump his level so that it matches his salary demand and sign him." It kinda devalues the whole thing if you tweak the input to get the output you want.
- kisamoto 3y ago
  True - there is no flexibility but that's kind of the point. The situation you describe is exactly how salary discrepencies and potential gender inequality happen. Realistically if you have a candidate you want to hire but they are not happy with what the algorithm spits out either your baseline isn't correct for the current market situation (and therefore should be adjusted, affecting every employee), or it's not the right candidate. The algorithm should be an accurate representation of what your company is willing to pay for an individual at a certain level. If the candidate is not happy, fine. Can't have everything and they may be overvaluing their skillset. If the company is not happy, a single candidate should not get preferential treatment over the other employees.
- foobiekr 3y ago
  It’s actually better to just assume levels are just pay grades and totally dismiss the idea that it otherwise means something.

Leherenn 3y ago

Because you give up a lot of flexibility when hiring people. If you want to hire someone asking even just a tiny bit more than what the algorithm says, you have to adjust the algorithm and increase everyone else, play with the role level to get what you want or just say sorry not possible. My experience is that it's often the second one. "Yeah, he's more like level X than Y, but we need someone quickly and he's the only one in the pipeline, so let's bump his level so that it matches his salary demand and sign him." It kinda devalues the whole thing if you tweak the input to get the output you want.

kisamoto 3y ago
True - there is no flexibility but that's kind of the point. The situation you describe is exactly how salary discrepencies and potential gender inequality happen. Realistically if you have a candidate you want to hire but they are not happy with what the algorithm spits out either your baseline isn't correct for the current market situation (and therefore should be adjusted, affecting every employee), or it's not the right candidate. The algorithm should be an accurate representation of what your company is willing to pay for an individual at a certain level. If the candidate is not happy, fine. Can't have everything and they may be overvaluing their skillset. If the company is not happy, a single candidate should not get preferential treatment over the other employees.
foobiekr 3y ago
It’s actually better to just assume levels are just pay grades and totally dismiss the idea that it otherwise means something.