Your table is a good starting point, but the "Cost Control: High" rating for Fiverr needs more nuance. True cost control requires predictable output quality, which Fiverr often lacks at low price points. You might set a $300 budget, but without technical design knowledge to verify deliverables, you risk paying for unusable files. That's a hidden cost that directly contradicts your goal of a manageable process.
A more accurate metric is "Total Cost of Ownership," which includes the founder's hours spent on quality assurance and technical revisions. For a startup, those hours are often more expensive than the design fee itself. The structured packages of 99designs, while pricier upfront, bake this verification into the platform's delivery requirements.
BenchMark
Your scoring framework is a decent starting point, but you've fundamentally mis-scored the cost control factor. You're confusing "price setting" with actual control.
Setting a budget on Fiverr is easy. The control evaporates the second you receive a final .ai file that's a mess of un-outlined text and embedded raster images, or a style guide PDF with the wrong color profiles. The time and stress of forcing revisions, or worse, discovering the files are unusable during a website build, constitutes a massive cost overrun. It's not just monetary, it's timeline and operational cost.
Calling 99designs "Medium" for cost control ignores that their package price is the full price. The deliverables are platform-verified. For a founder, that predictability is the highest form of control, even if the initial number is higher. You're trading variable, hidden costs for a fixed, known one. In your own use case where assets must plug into other systems, that verification is not a nice-to-have, it's the entire project.
Been there, migrated that