Blockchain
Blockchain
What is blockchain?
- A distributed ledger shared among many participants
- Mutually verified, hard to tamper with
- No central administrator needed
What blockchain really changes is not the tech stack but the point where reconciliation work between companies disappears. Until now, business transactions meant each party recording in its own ledger, comparing at month-end, and arguing by email when the numbers didn’t match. When everyone shares one ledger, there is only one ledger to check from the start, so the time spent on settlement reconciliation and dispute checks shrinks. Put the other way, for internal data with no counterparty to reconcile against, there is almost no reason to use it.
Blockchains fall into two broad camps. Public chains, open to anyone, are slow: Bitcoin handles about 7 transactions per second, Ethereum a few dozen, and fees fluctuate. Consortium chains, where participants are approved in advance, can process thousands per second, but because an operator is already defined, the original benefit of having no central administrator is largely diluted. Most enterprise projects are the latter, which is why you will always be asked, “How is this different from a shared database?” It comes down to one question: do two or more organizations with conflicting interests need to see the same record?
The most common misunderstanding in the field is that “once it is on the chain, it becomes true.” Blockchain only prevents recorded data from being changed; it cannot filter out lies told before the data is recorded. If an inspector enters the wrong origin, the wrong information stays forever, and removing it means redesigning the chain. There are also real incidents where documents containing personal data were uploaded as is, and the company, unable to honor correction or deletion requests, was accused of violating the Personal Information Protection Act. That is why the standard design keeps the originals in the existing database and puts only the hash values on the chain.
It also helps to reverse the order of evaluation. PoCs that pick the technology first and then look for a use usually burn through six months of budget and quietly fizzle out. Are there three or more participating companies? Do disputes keep recurring because they can’t trust each other’s records? Is the cost of those disputes tallied annually? These three questions are the criteria. If even one is missing, a well-designed shared database is cheaper and faster.
Ilyoung JungExpert in smart factories, equipment management, and Industry 4.0
John KimInstructor John Kim, author of “B2B Marketing Design” and CEO of Nanum ConsultingThe core concept was presented in Satoshi Nakamoto’s (a pseudonym) 2008 Bitcoin paper, and has since expanded beyond cryptocurrency into a range of distributed-ledger applications.
Consider a hypothetical case. Company A, a seafood importer and distributor with annual revenue of about KRW 90 billion (≈ US$67,335,000) and 140 employees, supplies three major supermarket chains and kept running into disputes over proof of origin. Catch certificates and customs documents moved by fax and Excel, so when a claim came in, it took an average of 9 days just to gather the paperwork. The quality control team took on an 8-month traceability project.
- When the quality control team leader reviewed the past year’s 22 claims, 14 were disputes of the kind “we have the paperwork, but we can’t tell which version is genuine.”
- The IT team lead objected: “If it’s just for us, a shared server is better.” After discussion, they narrowed the consortium chain to only the segment where five outside companies (shipping lines, customs brokers, and processing plants) record together.
- The head of procurement took on persuading partner companies, stating firmly that “the input forms stay the same; only the storage is shared,” and over three months ran two rounds of data-entry training for 12 staff at 5 partner companies.
- When the legal team pointed out that the scanned documents contained crew members’ personal data, so deletion requests could not be honored, the team redesigned the system: keep the originals in the existing document system and put only the hash values on the chain.
- After 8 months, the time to gather claim documentation fell from an average of 9 days to 1.5 days, and origin-related returns dropped from 22 to 6 a year. However, only 40% of items have been put on the chain.