Hacker News Reader

Comment by mattbrewsbytes | Hacker News Reader

At initial glance I also think it would be much cheaper to give out adjustments like this than have like 10% of your staff leave. Some napkin math indicates otherwise: if you have a staff of 10 that each make 100k on average, a 5% COLA costs you $50k vs. 1 person leaving (10%) and costing you maybe $25k in recruitment. That $50k is year over year too, not just a one-time deal, your avg salary is now $105k.

So in this fictional, theoretical employer scenario it would take 3 people leaving (30%) in order for it to have been less costly to just have given out 5% to everyone.

Even with my napkin math I still think its better to take care of your people reasonably vs. use a spreadsheet to drive decisions and hope people don't leave.

View on HN

It also is in the company's advantage for 1 person to leave and now give the other 9 the work of 10 and push them to maintain productivity. Now they're saving 100K per year.

I had a previous company do that to a department that had 40 employees, and three years later it had 5 employees (a couple of layoffs, but mostly steady attrition with no hiring to replace them).

Saved them a ton of money in the short term. But the department also stopped being able to do anything but be in "keep the lights on" mode, almost entirely maintenance and very little new feature development, and they also lost two major clients in the process as well as they couldn't keep up with SLAs, and thus their revenue ended up dropping by several million dollars a year, so I'm not sure if they really saved money in the end.

This is the issue with letting COLA types of pay increases lapse, then that employer may fall out of favor, be known to not treat people right and it could steamroll to where they can't get talent.

That's probably true, but in this company's case, it was because they had a hiring freeze for most of that three years, and a "la la la la I can't hear you" policy to middle management that begged to let them hire more people after that.

Or the few times they did allow it, about 2x more people left than they approved to be hired during that time period so the employee count still went down.

A more realistic model would see “1 person leaving” happen every year, too.

Also, “maybe 25% in recruitment” isn’t all the costs. You have to get the new hire up to speed in your company. If that takes that employee 3 months and also 3 months of the time of the remaining colleagues, that’s $50k in itself.

Finally, you can’t keep not correcting salaries for inflation for many years. The number of people leaving per year would go up dramatically, and you wouldn’t be able to get new hires.

The new hire is going to be at 105k as well. As people leave you will eventually ratchet up to the rate you don't want to pay anyway. You are right that it would probably be exponential the further wages get from market. One day a large portion of people would leave and their replacements will be less effective and cost a lot more.

On the other hand, if you are a executive you can keep rolling the dice, get a nice bonus and leave yourself in two years because you know the real money comes when you switch companies.[0]

[0]https://rationalwiki.org/wiki/IBGYBG

One issue is that you may pay ten people 1MM annually to write your software, but the productivity distribution is likely not only non-equal but also non-linearly so.

Those at the top of the productivity distribution are the more likely to be able to find new employment, so you may be losing more than a tenth of your development output by going that route, and finding a replacement for that type of employee may be much more costly and difficult than average.

Dangerous to model the cost of someone leaving on a spreadsheet like this :) https://twitter.com/benjiweber/status/1480219318273073157

+1 I think this point is really important.

When discussing pay versus employee retention, the toy model people present often implicitly assumes a high percentage of employees will leave if pay lags. But the economics totally change based on the percentage, and empirical data suggests the percentage is actually pretty low.

Reply on news.ycombinator.com