This isn't going to end well for Matthew, well if he doesn't work the back side of the curve, let me explain;
The challenge is performance is transitory and pay is forever. Specifically, you will have engineers who happen to be on a project at a time when they can have a big impact, and they do and you want to reward that. But then the next few years might be polishing that project's stability and corner cases which will "feel" like very little impact. As a result if you change their pay to reflect their impact when they first deliver, and leave it like that, then the next few years will "feel" like the curve is inverted and your paying and not getting the impact that you would expect from that level of pay.
This is why, invariably, engineer pay tends to evolve into "base pay" + "bonus pay" which can be thought of as the cost of keeping you around and the appreciation for what you have done for us this year. Many people make the "bonus pay" a fraction of your base pay, in that way an increase in your base pay will be reflected in larger bonuses later. Google famously did this and then added a twist which was called a "personal multiplier"(PM). This PM could range from 0 to 3, and the trick was it was kept secret. In that way you could never know whether or not your manager was "playing favorites"[1]. But it gave them a nice talking point about being "open" about how you would be compensated while not letting you convert the how into actual numbers.
The second thing is that everyone is different. And even when they are a lot the same, they are different. Sometimes people will reach a compensation level that meets all of their needs, and lets them save for retirement, and they don't "need" any more than that. Tandem had a great term for that, it was called people who "retired in grade" which was code for someone who was not being motivated by pay any more so they weren't working as hard as it was possible to work them.
The third thing is that sociopaths love a score. They really really do. All you have to do to get the most out of a sociopath is tell them how you are going to score them and let them see how they score relative to other people. (you'll say "pay" in your discussions but the sociopath will hear "score.") These folks will invest all of their time and effort in "winning", and because they are sociopaths they won't necessarily use that tactic of "work harder than those you are competing with." They may in fact spend all of their time developing schemes to sabotage other efforts in order that their effort stands out. Really not a great place to be, really.
The bottom line is that compensation is hard, doing it well requires that you be honest about what you would pay someone to replace the person you are evaluating and verifying that you are paying them at least that much. And when people start going the other way, you need a plan for that too (you can fire them of course, but as the article mentions they have lots of domain specific knowledge that is handy). Generally successful companies move those people to working on projects that are important but not urgent.
[1] To no ones surprise, the managers were playing favorites.