This is the last article in a three-part series on how to reduce uncertainty in the supply chain to reduce costs. This article deals with reducing uncertainty in the interface with your suppliers and supply network.
There are several ways to define uncertainty. I like two angles on uncertainty (which are not that different):
- Supply chain uncertainty refers to the decision-making process in the supply chain in which the decision maker does not know definitely what to decide due to lack of transparency in the supply chain and the impact of possible actions.
- Supply chain uncertainty refers to the change of the balance and profitability of the supply chain caused by potential and nonpredictable events that require a response to re-establish the balance. An event can be an unexpected order, late delivery from a supplier or a breakdown of critical production equipment.
As you will see in the following three examples, data to do analytics to reduce the upstream uncertainty is available. Data is often found within the company, but is strengthened by bringing in outside data. The second and third examples rely heavily on outside data.
Uncertainty: Will my purchase order be delivered on time?
One problem that hits manufacturers repeatedly is late deliveries from their suppliers. Typically, firefighting will begin to avoid customer service being hurt. It will further reduce the OEE, as equipment will stand idle or experience more set-ups.
Traditionally, the preventive method is to carry additional inventory in order to cope with late deliveries (safety stock). Another commonly used method is to ask the supplier to send an alert if the promised shipment date is not met. This will not avoid the stock-out, but it will enable early contingency planning, enabling better use of internal resources (capacity and materials).
Mentioning all that can happen to one purchase order between the time it is placed and delivered is not relevant to this blog. The constraints, events and human interactions that can happen are numerous.
The first level of analytics is to do predictive analytics based on data we know, for example:
- Supplier performance.
- Order size.
- Order history (what changes have been made to the order).
- Logistics provider performance.
The second layer is to use event stream data to identify if delays are likely, and when the shipment will arrive. Imagine a shipment being delayed one hour from the supplier, delayed one hour on the road to the logistics centre, missing the same-day shipment here, and thus being late for the harbour and missing the ship – with the next ship coming a week later. The data will come from the supplier and the forwarder.
This will enable an earlier response in terms of replanning – potentially leading to choosing a faster delivery method. Further, being better at predicting when an out-of-stock situation is likely to occur and improving the early warning can lead to lower safety stock – and thus lower Net Working Capital.
Recently, the idea of a supply chain control tower that monitors everything has emerged. It is likely that these will be implemented over the coming years – with analytics as a key part to improve the value of data.
Uncertainty: Am I a victim to fraud?
Procurement fraud matters. According to PWC, it is the second-biggest economic crime after theft in terms of losses, and some estimates suggest that businesses lose around 5 percent per year of procurement spending as a result. According to ACFE $3,7 trillion is lost to procurement fraud each year.
Dealing with fraudulent behaviour in procurement is pretty much like searching for the needle in the haystack. Fraudulent behaviour takes many forms besides what we typically regard as fraud, as it also covers errors and exceptions. Fraud can also happen in good faith and without errors – e.g., when the world changes (example: a person married to a supplier comes to work in your procurement department).
Since fraud takes many forms, different techniques are needed to identify the problems. Common sense and simple business rules will go some of the way. However, to catch all fraudulent behaviour, analytics can be a big help in detecting patterns and events. Jeff Dunham has written more on fraud detection in procurement.
Most of the data are available for such analysis – and naturally looking at the data in the invoices is a great place to start. However, data can come from multiple sources, including social media, phone books or company registers.
Uncertainty: Does the supply chain live up to our ethical standards?
Google “supply chain transparency” and you will find links to stories about the fashion industry and blockchain. Blockchain technology can document the chain for each piece of garment. The technology can help supply chain participants record valid data about price, date, location, quality, certification and other relevant information.
This will increase traceability of materials, reduce issues with counterfeit products, and improve visibility and compliance (e.g., with outsourced contract manufacturing).
The blockchain provides the data, and analytics can help identify if there are any ethical issues. This can be relevant in both manufacturing and retail. Blockchain enables a company to build a position as a leader in responsibility by being able to document the origin and route of all the products it sells.
Knowing the details of your supply network and combining it with text analysis on data from, e.g., social media can help you identify issues early and thus react faster – at least before it becomes a scandal in the media. There are lots of examples of companies having to deal with bad press due to their supply network.
Blockchain data can also be used for several other purposes – e.g., as an important input to the supply chain control tower previously mentioned.
Know the details of your supply network and reduce uncertainties experienced when looking at your #supplychain by using #analytics . Click To TweetSumming up
The bottom line is that the uncertainties experienced when looking at your supply chain can be reduced by using analytics. The first example was very operational, while the latter two are more a question of monitoring your supply chain for issues that are unacceptable.
| Uncertainty | Data | Analytics | Value |
| Will my purchase order be delivered on time? |
Supplier performance, order size, order history, logistics provider performance (all from ERP) |
Predictive analytics, event stream analytics | Early planning, lower safety stock |
| Am I a victim to fraud? | Invoices, social media, phone books | Predictive analytics, network analytics, text analytics, machine learning | Reduction of procurement spending, in some cases up to 5% of total |
| Does the supply chain live up to our ethical standards? | Blockchain data – social media | Statistical analysis, predictive analysis, machine learning, etc. | Brand reputation |
Discover more insights on SAS Industrial IoT Information Hub
