Google Solution Architect

So now you've got the compensation data, are you ready to negotiate? Not quite.

First off, comp data is constantly changing. 2022 numbers are coming in higher than the 2021 numbers listed above, and it’s critical to know the full range for each component to optimize your negotiation. But even more than that, many people think knowing the compensation data is the most important part of a negotiation, but it’s not. You can go tell your Google recruiter that you read online that top of band is $X and that you want that number, but in 90% of cases that won’t work. So what will?

Companies negotiate when they believe you won’t accept their offer. But, since you just finished a long and arduous interview process with them, they know you are likely to accept the job unless you have higher competing offers from other great companies.

This means you need to create leverage and make it believable that you are considering other options but would prefer Google if the compensation is comparable. This sounds easy but quickly becomes challenging if you don’t have other offers or if your other offers aren’t higher than the top of band comp numbers you are targeting.

If you are lucky enough to have a competing offer with numbers above top of band, Google negotiations are often still tough. Google has tons of situation specific rules. For example, if you are negotiating a Google Bay Area SA offer and have a higher offer from Facebook Seattle, we've seen Google recruiters refuse to match that because of "location". That is of course a ridiculous excuse since the Bay Area is a higher paying market, but you need to be prepared to handle these situations. Google will also refuse to match private company equity (e.g. TikTok) and apply a 25% discount. We've seen this derail negotiation plans. And of course most people have heard of their requirement to see competing offers in writing, which is possible to avoid if you follow the right steps. The Google compensation team is essentially a black box, however, with enough testing you can start to piece together what inputs can reliably be turned into a specific output.