What does a blockchain offer in this case that submitting a line to some sort of B2B logging API where the logs are made public (like certificate transparency) doesn't?
but at least intuitively, a signed blockchain (in the true sense that each block contains a hash of the previous block) does seem to confer some additional security over a plain database.
A plain database would give you a log of events and who logged them, but security and trust resides in the database and how it's authenticated. Anyone with sufficient permissions could modify the database.
A signed block chain would give you a log of events where each party in the chain signs the new block (which includes a hash of the previous block) using their private key. The block can then be verified using each entities public key. To modify this system you would need to be able to breach the trust of the private keys.
{Lettuce from Farm A on date, <>}[Signed by farm A] => {Received lettuce on date, <Farm A block>}[Signed by shipping agent] => {Received lettuce on data, <Shipping agent block>}[Signed by Walmart store], etc
In theory then, Walmart would have a registry of the public keys of each agent in the chain, and upon receiving produce could verify the events of the whole block?
But now you've also raised the technical requirements and are betting the security on non-technical farmers and farm staff. Even if they have technical chops since farm equipment is more complex these days, they likely don't have security chops.
There's already 'country-of-origin' scams in international shipping, so it's not a far stretch to apply the same concept to farming supply chains. Throw a disposable proxy farm in front of your real farms and change the origin. When it's human-entered information, blockchain doesn't prevent someone being dishonest from the beginning, and it lures people into a false sense of security because of IBM's advertising.
I guess limiting fraud to the beginning of the supply chain is a step in the right direction though. I don't think blockchain solves chain-of-custody well though.
1. Packaging it in such a way that it's easy to use for non-technical users.
2. There still has to be a mechanism for establishing trust in the public keys registered in such a system.
For one I'd imagine that farmers are doing some amount of work to go from raw "lettuce in the ground" to "one unit of shipment". Once it's a unit of shipment it could have an ID/Bar/QRcode generated that is then scanned with a handheld scanner that has its own keypair. The scanners keypair would be registered with the farm and so the farm would either explicitly sign off on the package or implicitly have signed that scanners pair as trusted by the farm. So a farmer and farmstaff wouldn't need any technical know how beyond operating a scanner. Depending on the farm adding such a granular tracking system may or may not be easy to implement.
For two, I can see two options. Walmart could verify every step of the chain themselves and register/receive the public key for each step in the chain. (So a farm would register with Walmart as a verified producer, a shipping company would register, etc). Or each step of the change could make a binary decision to trust the step before it and apply pressure down chain to resolve issues, since theoretically each step could fabricate the entire chain before passing it off. In this scenario, Walmart would go to the shipping company and say "E. Coli was detected in Shipment X. We no longer want to be sourced produce from Producer Y. The shipping company then takes this into account when sourcing produce, or else Walmart discontinues business with them.
I think the big benefit is by having an associated public key, you can verify the claimed identity of any agent in the chain and build trust in that agent over time. You could keep popping up disposable proxy farms, but Walmart/Shipping provider/X doesn't have any reason to trust your proxy farm's shipments; you have to establish trust either directly with a buyer of your produce, or establish trust over time as a commodity provider. As you build trust you can probably get better prices.
I'm by no means a blockchain advocate or a supply chain manager of any kind, just doing a thought experiment of how blockchains might be effectively applied in such a scenario. There very well might be better solutions, but I think at the very least a blockchain presents an interesting or novel solution.
Just thinking through, you could also create a sort of "doubly linked list" and include who custody was given to as part of each block, so a farmer is signing a claim of who they gave their produce to. {Farm A produced unit of Lettuce X, Given to Shipper B, <no previous>}[Farm A signed]
But again, that's why I mentioned a blockchain in the true sense, as a chain of blocks containing a hash of previous blocks, because git is a blockchain by that definition.
And it’s any different on a private blockchain?
If the suppliers themselves enter every piece of information into what is essentially a living contract, any error or omission could be considered a breach of contract and can be leveraged.
If it was only about food safety and increased sales because of increased trust the solution you propose, maybe with some kind of basic digital signatures, would be fine, and probably much cheaper.
From what I've heard, Walmart seems to like moneh very much, hence doing something a more expensive way is likely to be intended to increase their bottom line in more ways than one.