Do you actually need a blockchain? An honest decision guide

Blockchain attracts proposals that would work better, cheaper and faster on an ordinary database. The useful question is not "can we use blockchain?" but "what problem needs it?"

What a blockchain is good at

A blockchain is a shared, append-only record maintained by multiple parties, where no single party can quietly rewrite history. It trades speed, cost and simplicity for shared trust without a central operator.

That trade only makes sense when the lack of a trusted operator is the actual problem.

A short set of questions

Work through these honestly:

1. Do multiple parties need to write to the same data? If one organisation controls all writes, a database with proper access control and audit logs does the job.
2. Do those parties distrust each other, or each other's operator? If they are happy to trust one of them, or an independent service, you do not need a ledger.
3. Is there no acceptable intermediary? If a trusted third party already exists and is accepted, adding a blockchain removes little and adds a lot.
4. Does the record need to be tamper-evident to outsiders, over a long period? Some value is genuinely about independent verifiability.
5. Do you need programmable, automatically enforced rules between parties, such as escrow or settlement? Smart contracts can do this without a central enforcer.
6. Can you accept the costs: lower throughput, transaction fees on public networks, immutability that makes mistakes hard to fix, and public visibility unless you design around it?

If you answered "no" to the first three, use a database. It will be faster, cheaper, easier to change and easier to hire for.

Where it genuinely fits

• Digital assets and tokens that must be transferable between parties without a central issuer holding the books.
• Settlement and escrow logic enforced by code among parties who do not want to rely on one another.
• Shared registries across organisations where no member wants to be the single owner of the data.
• Verifiable provenance and audit trails where independent verification is the point.
• Decentralised applications where users want to hold their own assets and identity.

Common misconceptions

• "Blockchain makes data accurate." It makes data hard to change after recording. Bad data recorded is permanently bad data.
• "It is automatically secure." The ledger may be robust while the application, keys and smart contracts around it are not. Most losses come from those.
• "It will make us faster." Usually it makes things slower and more expensive.
• "We can put everything on-chain." Storage is costly and public. Put hashes or references on-chain and the data elsewhere, and think carefully about personal data and deletion obligations.

Public, private or consortium

Public networks maximise openness and censorship resistance but have fees and public visibility. Private or consortium chains give control and privacy to a known group, at the cost of weaker decentralisation guarantees, and at that point the case against "just use a database" gets harder to argue away. Choose deliberately.

What to take away

Use a blockchain when multiple parties must share and trust a record without a central operator, and when you accept the cost. Otherwise choose a database. An honest "you do not need a blockchain" is often the most valuable advice available.