What is the overall conclusion of Do You Need a Blockchain? (Wüst & Gervais, 2017) about permissionless vs permissioned blockchains vs central databases?

None is universally best — each is the right answer for some application, and the choice must be made deliberately from trust assumptions, the parties involved, and performance needs.

The take-away is anti-hype but not anti-blockchain. The authors provide a structured methodology precisely because the decision is not trivial and shouldn't be driven by marketing. Weighing required trust assumptions, application requirements, who the writers are, and technical characteristics like throughput and latency, they conclude there are genuinely valid use cases for each option — permissionless, permissioned, and plain centralized databases. The engineering discipline is to determine, case by case, which trust model the problem actually has — and often that answer is "you don't need a blockchain."

Go deeper:

• Wüst & Gervais, "Do you need a Blockchain?" (IACR ePrint 2017/375) — the full paper, including the three worked use cases and the summary comparison.