Everyone's budget is fixed to an extent on almost any purchase one desires to make, either by your boss, your spouse, your bank account, your credit score or your own innate opinions of worth and value as it relates to physical goods.
My credit score and debt-to-income ratio told me I could afford a house that was 30% more expensive than what I ultimately purchased. I used a variety of factors to determine that the bank was out of its ever loving mind, despite the numbers...mostly my own common sense and the innate lack of wanting to be held hostage by a high mortgage. As a two income household, the goal was to be able to pay the mortgage on a one income household, which we can. The benefit of the lower purchase price allowed margin for extra principal payments, unexpected unemployment, catastrophic illness or death.
The area I work in does not require the latest and greatest tech to be productive and profitable. However, I have never really skimped on hardware regardless. Almost every iMac or MacBook Pro I have ever owned has been near the top of the BTO chain. Core i7, max DRAM, best GPU, most storage I could reasonably afford. I don't do it to brag about what I have, but merely to make sure my purchase will last a reasonably long time and allow for some headroom as code gets bloated and I experiment with new things.
EDIT: Also, almost every Mac I have owned since 2011 has either been a closeout or a refurb, because I hate buying retail.
My argument is that, depending on your workload and area of industry, you would more likely see a better long term return with upgrading the DRAM or the GPU, if your workloads would benefit from more DRAM or a faster GPU, given the performance gains measured by others over the base configurations Apple sells and the fact that DRAM can speed up certain workflows enormously, while the GPU is being used a general purpose processor for more and more operations given the relative lack of progress in CPU IPC gains since the release of Sandy Bridge.
I would even venture to guess that the GPU upgrade is a bit of a quarter toss as an eGPU might be a better use of $250-$350 given that more powerful GPUs exist that can be utilized in macOS workflows than the Vega 16 or 20 Apple offers as a BTO option.
However, if your particular workload is served better by faster CPU clock speeds and cores (i.e. video compression versus video editing), then by all means, you should opt for the fastest CPU you can at the point of sale, especially since it cannot be changed later on.
In my experience, CPUs are already so fast for general and specialized computing to the point that the need for the absolute fastest CPU is more esoteric exercise than actual problem. Because if it were, you wouldn't be wasting time asking about a 100MHz option for a MacBook Pro, and would have already bought an iCore i9 iMac or just about any flavor of iMac Pro and called it a day.
