How Boeing, Toyota, Caterpillar, and other OEMs can double their current net profit by using smart contracts to become unmanned “virtual companies”, with or without cryptocurrency: Part 6
Going from Sales Forecast to Supplier Orders, aka the “Meat” of the Smart Contract’s Then-Condition
Okay, we’re back to the smart contract! We have our “if” condition from Chain Link; the sales forecast from the AI. What other inputs might the smart contract need? How will the smart contract manage and optimize the supply chain (mostly) without human intervention? What happens now?
Each model in the sales forecast has its corresponding bill of materials (parts list). For instance, the bill of materials for a Yamaha R1M would be: https://www.yamahapartsmonster.com/oemparts/l/yam/5aba776987a86612183a1239/2018-yzfr1m-yzfr1mj-parts (click in any assembly to see the individual lower level parts).
A Kawasaki Ninja H2R would be: https://www.kawasakipartshouse.com/oemparts/l/kaw/5af07c4987a86610e8e11c9d/2018-ninja-h2r-zx1000yjf-parts (click in any assembly to see the individual lower level parts).
In turn, each part is produced by a supplier according to the terms of a production contract. For instance:
Brake holder part number 43034–0108 will be unit price $40 assuming the OEM buys 10,000/yr and the London Metals Exchange price of aluminum is $2,000/ton. For every 2,000/yr less in orders, the unit price goes up $3 (pro-rated). For every 2,000/yr more in-orders, the unit price goes down $2 (pro-rated). The unit price is 15% dependent on the cost of aluminum. This part requires a lead time of 5 weeks to deliver. 2 are required on each Kawasaki Ninja H2R and 1 is required on each Kawasaki Ninja ZX-1000R.
To recap the translation steps:
Incoming sales forecast → Vehicle models → Bills of material →Part numbers →Supplier orders
What should trigger the automated outgoing orders to suppliers, actual customer/dealer orders or just the sales forecast? It depends on the situation.
The smart contract can be programmed to be highly adaptive so that it can make the right call in each different scenario. If the AI sales forecast anticipates a market downturn in the near future, then the smart contract will only allow actual dealer orders to automatically trigger supplier component purchases. The smart contract will also try to get rid of any excess parts inventory it has on-hand and run as lean as possible.
Whether the ordering interface is via the OEM’s self-built web portal (w/ back-end API), a Salesforce interface, or an SAP interface, recall from the graphic in part 5 that Chain Link can handle it all.
If the AI sales forecast is highly confident that a massive boom in sales will happen 5 months from now, then the smart contract will start ordering extra parts from suppliers and stockpiling them in anticipation of the highly likely boom (before any orders have actually come in). It’s better to build up a bank and accept the inventory costs rather than be caught completely unprepared to support those customers in 5 months.
Returning to the brake holder example. Let’s say the AI is 60% confident that 50 customers will order an H2R within the next 5 weeks and 60% confident that 50 customers will order a ZX-1000R in the next 5 weeks, but 90% confident that 50 customers will order a ZX-1000R or an H2R.
The smart contract will go ahead and order 90% x 50 x 1.5 = 67.5 ≈ 70 brake holders, purely based off the AI forecast and before actual orders have come in.
Whether it’s ZX-1000Rs or H2Rs, those parts are probably going to get used up. When it comes to parts that are shared across models, the risk of idle inventory is much lower.
If the OEM is swamped with so many orders that it’s backlogged for the next 13 months, then the smart contract will obviously completely ignore the AI forecast and only go off actual orders.
In a level market with only 3 weeks of order visibility, the smart contract will rely heavily on the AI forecast. After all, the average supplier might need 7 weeks to deliver a part. And given that it’s a level market, the AI’s confidence is even higher than it would be in a downturn or boom (which is already plenty high).
Of course, the smart contract will do much more than order parts in a manner that optimizes sales potential and on-hand inventory. It will also perform risk mitigation and cost reduction. It will constantly balance those two directives to stay on the “cheapest low-risk path” (which is always the goal at the meta-level).
Risk mitigation matters because at the end of the day, you need 100% of your supply chain to deliver. Even if only one critical component is missing, you still can’t sell to your customers. Doesn’t matter if the other 99.9% of the supply chain delivered. In other words, having parts with which to build vehicles is always the highest priority.
Cost engineering matters because the typical OEM needs to implement 3% cost reduction each year just to stay profitable and competitive (Boston Consulting Group).
Smart contracts are capable of automated supplier bidding and selection. To quote Cognizant:
“A supplier could issue a smart contract on a blockchain platform that specifies the product definition, quantity, price, availability date, and shipping and payment terms. A manufacturer could automatically search the blockchain for smart contracts that meet its requirements, verify the reputation of the seller for quality and timeliness based on data on the network, and then complete the transaction, without the need for a manually-generated purchase order. The supplier could then automatically locate and execute a smart contract with a carrier for delivery.”
In other words, the OEM smart contract can automatically hunt for cheaper and/or less risky suppliers 24/7. OEMs may have millions of buying decisions, but the smart contract will optimize every last one.
To understand what that hunt might look like, let’s start by discussing some of the major forms of “delivery risk”.
1. Single-source supplier:
By far the most notorious case-study is the Meridian Lightweight Technologies fire. This supplier was the single-source for instrument panel and radiator components that went on every single F150. Ford is on track to sell a million F150s this year @ $45,000 each. Doing the math, 2–3 months is $7.5-$11.3 billion.
Ford lucked out and only ended up losing $591 million (dealers coincidentally had ~3 months lot inventory). But still greatly damaging Q2 2018 net income ($1.07 billion vs $2.04 billion in Q2 2017).
Sources:
- https://www.detroitnews.com/story/business/autos/ford/2018/07/25/ford-profits-down-nearly-half-due-supplier-fire-china/826985002/
- https://www.mlive.com/auto/index.ssf/2018/05/ford_has_84-day_inventory_of_f.html
- https://www.fool.com/investing/2018/07/22/ford-motor-company-earnings-how-badly-will-a-suppl.aspx
Close calls like this are pretty typical in aerospace & automotive, even if most don’t make the news. There are hundreds of other Meridian Lightweight Technologies out there, just waiting to really trip up an OEM’s supply chain.
The smart contract could be programmed to automatically identify the most dangerous single-source failure points and start reaching out to potential second-source suppliers.
In the event that commodity managers or engineers accidentally change a dual-source to a single-source (as a result of a deal negotiation or design change), then the smart contract would automatically send a stern email notifying them of the errors of their ways.
2. Overseas supplier in volatile economic (or political) climate:
Cheap overseas labor becomes a problem if it comes with volatility. Better to have expensive domestic parts than no parts at all (which equates to sales disruption).
If volatility index high, then email warning to commodity manager, email warning to foreign exchange risk manager, and automatically start reaching out to potential alternative suppliers.
As shown below, pulling the real-time data into the smart contract can be as simple as an XML API call to Neven Valev’s Global Economy Project (presumably with Chain Link). In the future, there will probably also be artificial intelligence services for calculating economic volatility. You would also connect to those with an API (once again, see appendix E at the end of part 20 if you’re confused on what APIs are). Chain Link can aggregate the conclusions of the human research groups and AI services for a well-balanced prediction.
3. Financially distressed supplier:
It’s important to squeeze the suppliers, but not the point where they go bankrupt. Slightly more expensive parts are still better than no parts (which equates to sales disruption).
With the right inputs, the smart contract could be discerning here. If supplier in distress, then hold off on automated cost reduction efforts for one year.
For publicly traded suppliers, the input could be as simple as pulling web API data from Marketwatch.com. For private suppliers, it would be up to them to open up their books. This could either take the form of opening their bank account API to the OEM (more on this in scenario 2 of part 8) or a traditional financial report (submitted through a web portal with standardized format and input fields so that the API data can be passed on to the smart contract in a structured way).
In many cases, disclosing their internal financials to the OEM may be a better alternative for the supplier than continued relentless cost reduction that leads to potential bankruptcy.
Such openness isn’t even that unusual in this day and age. After all, it usually is a win-win for both parties. To quote a Deloitte study:
“With the trend for increasing cooperation and integration between upstream and downstream enterprises, the auto ecosystem has evolved such that in recent years OEMs have required components suppliers to proactively offer their financial data & cooperate with OEM inquiries”
The smart contract would aggregate the overall risk posed by all these factors and thereby determine the proper amount of inventory to hold. If the risk is high enough, then the smart contract will deem it worthwhile to incur the overhead costs of additional inventory. If the risk is low, then the smart contract will work with a true Japanese-style “Just in Time” approach.
Now for a discussion of some of the major ways to reduce cost.
1. Seek opportunities to re-source to LCCs (low cost countries) and BCCs (best cost countries):
“By allowing supply chain partners to create trusted relationships without the need for banks or, perhaps, even traditional purchasing processes, manufacturers, suppliers, customers, and machines can find each other and do business much more quickly and inexpensively. Even more importantly, they will be able to form more agile supply chains through smart contracts that automatically find, negotiate with and close deals with partners the world over.”-Cognizant
Of course, existing suppliers would always be given the right of first refusal. Production contracts would only be re-sourced from existing suppliers if they declined to match the alternate overseas supplier’s lower price.
2. Convert out-of-contract buys to on-contract:
Sometimes unexpected scenarios come up, i.e. supplier fires, robots periodically building a defective vehicle from imperfect ingredients, etc. Factories are dynamic and complex environments. The unexpected edge cases are bound to come up in a million different flavors.
In turn, this results in OEMs needing to make last-minute emergency orders from suppliers that are outside of a production contract and therefore much more expensive. In some cases (for one reason or another), out-of-contract buys become a recurring habit that persist for several months before somebody gets around to getting it on contract.
Smart contracts could follow up much more aggressively, ensure the loopholes get closed much faster, and stem a lot of bleeding. Humans blink and miss things, the smart contract doesn’t.
But how would the smart contract even know that out-of-contract buys exist if they’re (by definition) beyond the planned, conditional parameters?
The OEM could request that the suppliers submit invoices for out-of-contract buys through a web portal that shares its API data with the smart contract (or a Salesforce/SAP interface that Chain Link can also pull from). In turn, the smart contract will automatically request for supplier bids and 30-piece sample studies for PPAP on every out-of-contract buy in the system.
The PPAP samples would be reviewed by the 3D scanning robots described above and the contract automatically awarded to the cheapest bid from a supplier with sufficiently low country volatility, sufficiently healthy internal financials, and sufficiently good on-time delivery metrics.
3. Change the manufacturing process accordingly when volumes increase:
If volumes for parts with 3D printing as the designated manufacturing process go above 500 parts/yr, then email human cost reduction engineer about converting to injection molding.
If volumes for parts with thermoforming as the designated manufacturing process go above 30,000 parts/yr, then email human cost reduction engineer about converting to injection molding.
There are too many part numbers for humans to keep track of everything that is happening everywhere. Opportunities to save cost are bound to be missed. But the unblinking smart contract wouldn’t let a single opportunity slip by.
Note that manufacturing process changes are tricky enough where you wouldn’t want the smart contract handling everything by itself. Best to have a human engineer in the loop to manage any subtle alterations to the drawing requirements that result from the manufacturing process change.
4. Choosing a cost-effective material:
Let’s say that a cast-aluminum part, X123-A, also has a ductile iron alternative, T777-B. This information can be built into the smart contract.
If the price of ductile iron is favorable, then order T777-B. If the price of aluminum is favorable, then order X123-A.
Chain Link can easily pull and aggregate this information from a combination of London Metals Exchange and other reputable authorities on the web.
Continue to part 7….