I was talking to a client with regard to blockchain use cases and how it may be relevant (or not) to their company. Their main qualm is that – Given that blockchains, both public and private, have complexities and tradeoffs, why do people want to use them?
My response as below:
Blockchains do have their own use cases when it comes to a particular set of activities, especially when the use cases demand for the advantages of open access, immutability and transparency, while being willing to trade-off on certain aspects like requiring high levels of redundancy (hence requiring more storage capacity) and lower throughput compared to centralised services.
From the perspective of public blockchains, for example: distributed consensus when it comes to decentralised transfer of value (e.g. Bitcoin) which is global and unstoppable, or distributed consensus to decentralised computing (Ethereum). With a public blockchain acting as the backbone infrastructure, these use cases could be developed with no intermediary to do the processing, instead relies on economic incentives for nodes to do the processing for them, which due to its decentralised nature does require some fairly complex design mechanism to ensure the network is safe to use.
With private or consortium blockchains, use cases in reality varies depending on industry/implementation. For example, it may be more sensible to have a blockchain system for supply chain as it involves a huge number of stakeholders. Or maybe national land or asset registries, which requires some level of tamper-proofing and cross government body and/or financial institution validation.
It would make less sense to implement blockchain for, for example, my in-house accounting system which the stakeholders are myself and possibly my auditors (just use an in-house accounting software which allows for data export to your auditors). Or to use blockchain to store KYC data which is relevant to only my company (just use an off-the-shelf or bespoke KYC Management Tool, and when you have partners just get together to agree on a particular synchronisation method which could be as simple as a end-of-business-day CSV export/import).
(PS: In fact, most private blockchain implementations could just run on a SQL server then settle the data on-chain for cryptographic auditability. Like Plasma.)
So yes, the hype you see is warranted for use cases which demand for it, and where it is required it often is a clear cut case of being required; On the other hand there are also many use cases which probably does not need a blockchain, which is perfectly fine if you decide not to go ahead with implementing one.