"Basically, you need to perform a binary search. When you get an offer, you know that you can get at least that. Then you repeat the process a bit higher." https://thehftguy.com/2017/01/23/career-advice-and-salary-ne...

Except that isn't a binary search.

But it sounds cool when you call it a binary search. Poetic licence.

It's not a perfect binary search. Not all companies have the same top number and you're not equally valuable everywhere.

It doesn't seem to be a binary search in any sense of the computer science term. There is no half interval mechanics nor would that be appropriate. It seems like traditional bargaining.

In an otherwise excellent article, the missing negotiation calibration and example is glaring. The rest of the article would seem to suggest just as you will get your next big promotion from another company, if you find yourself considering coming down walk away. The heart of the approach is getting away from you becoming invested in being hired by a specific company.

Companies have a range they are willing to pay. Surprisingly, most companies in an area (and domain) usually have very similar ranges. Finding the top has a lot in common with binary searches.

I am not sure I understand your second paragraph.

Could your provide a specific example of how to use a binary search to find a salary? A binary search is dependent on boundaries. I don't see how it could be used effectively to find the boundary for a specific employer. Also, the flow chart shown in the article has no similarities to a binary search. The article describes the more reasonable approach: using networking to find out what a domain in an area pays.

You can't do it for a single employer because you only get one test per employer. You do the search across employers, in an area/domain. That gives you a rough idea of the prospect.

Yes, networking and insider information is much better than any of this.

right, it's the wife search algorithm (fun fact, it's really useful to find a refueling station while on the go)