Logistics is experiencing a time of great excitement and spectacle with AI and supply chain data leading the charge. Every month, we’re introduced to new and innovative ways to run a supply chain:
- Chatbots handling Tier 1, even Tier 2, customer support
- Cross-system end-to-end supply chain visibility mapped by AI (Artificial Intelligence)
- Near real-time predictive analytics, proactive notifications, and prescriptive resolution
- Rapid system configurations and dynamic rules adjustment
- Generative human language that makes insights easily accessible
It can feel like the sky is the limit for new techniques and we’re only limited by how creative we are in leveraging them.
There is a “but.”
It’s easy to focus on the outcomes and neglect the process. I doubt many of these solutions are trumpeting the quality supply chain data needed to feed them. If anything, many of them downplay that. These marvels require incredible amounts of clean, consistent, and complete data sets continuously to produce usable output.
A quality outcome requires quality input. Before diving into any AI or advanced analytics solution, ask yourself: is my data ready?
Reid Bishop explores what it takes to extract and present the right data sets for Executives, Operations, and IT Security leaders.
It’s the data, stupid!
In 1992, Bill Clinton’s political advisor quipped, “The economy, stupid” on a list of campaign talking points for ‘what do voters care about?’. Since then, it has become shorthand that there is a simple core explanation for preference–elections are heavily influenced by the electorate’s feelings about the economy.
For AI and all advanced analytics, it’s the supply chain data, [stupid].
Not the output; the input.
“Garbage in, garbage out” isn’t shorthand to be taken for granted. Good data is the basis of any actionable analytics. Yet, it continues to be the most underappreciated asset to the intelligence consumer.
We’re not passive bystanders
As supply chain professionals, we don’t get the luxury of ignoring the process; we’re part of it. We actively create the data that feeds the beast with every action we make in our systems. We’re both the producers and the consumers.
- Are the trucking records accurate?
- Is the order data accurate between systems?
- Is inventory accurate, with positions and availability up to date?
This could be a list of questions, but I think you understand.
Especially important is ensuring that data makes it into the system. Avoid quick actions outside of the process, and make sure that all those sticky notes return to record.
Step up your supply chain data game! Learn more about Deposco’s Supply Chain Intelligence (SCI) solutions. The depth of shipping and executive insights you need, in a dashboard your team can use and trust.
The data we input and update, day to day, ultimately impacts future supply chain analytics. If we want quality output later, we must ensure we’re accurate today.
Lifecycle of a supply chain data project
Most data projects take an incredible amount of ‘data mining’ and clean-up to even start sifting for gold.
With the advancements in Machine Learning, it’s easy to overlook how much information gathering goes into it–both as a precursor and ongoing maintenance effort.
When doing Supply Network Optimization work, I created a timeline document explaining why it took 12 weeks to develop a network model from scratch.
- 1 week of reporting
- 2 weeks for analysis
- 9 weeks for data mining, cleansing, and harmonizing!
As you can imagine, the customer only appreciated 1 week of work; that’s what they wanted to pay for–intelligence. It’s no different than cooking a fancy dinner with prep work that started at 4am; the customer only cares about the hour they sat at the table.
Data analytics must be results-focused, but getting there is hard
The output of AI is exciting. The code and data collection to get there aren’t—unless you are the data scientist. This goes for even the most mundane, traditional business intelligence.
Sourcing, cleansing, structuring, and exploring supply chain data is tedious work that comprises as much as 80% of a project’s timeline. This is a combination of curating data in various locations and formats, and it is difficult to extract explanations from functional owners to make sense of it.
92% of leaders believe trustworthy data is needed more than ever. Read how we make this easy.
According to Gartner, 85% of data science projects fail to yield actionable results, and an estimated 30% of generative AI projects are abandoned after proof of concept.
The data is difficult to obtain, maintain, and leverage; the projects are difficult and often don’t lead to deployable outcomes.
Deposco makes supply chain data easy again
Before going on this journey, think about all that. Why would you want to go it alone?
The dirty secret: you don’t have to!
There is a bright spot to all of this. Let the management system handle it. The bulk of supply chain data creation can be handled natively within Deposco. The foundational database behind our solutions are built-for-purpose. Everything has been developed in house and there is no technical debt from a parade of acquisitions to sift through.
- Single schema – All Deposco fulfillment tasks are managed and recorded in a single database. No worries about different standards, incomplete records, or inconsistent design strategies.
- Built, not bought – Deposco’s Bright Suite has been developed from the ground up as a SaaS WMS for retail, 3PL, ecommerce, wholesale, CPG, and DTC fulfillment. There is no mix of online and on-premise solutions cobbled together from different vendors. It was all built for supply chain data to work together.
- Solutions are built on shared insights – When data is needed for predictive and prescriptive use cases, we eliminate the upfront lift that you don’t want to wait or pay for. You aren’t going through drawn out ETL processes to make your data work. Get to the part you wanted–valuable supply chain data.
- Bespoke, not patched, insights – Finally, our solutions are made for your business sector. So many AI insights are generic or they require you to fit your business into their use case. Because ours is based on supply chain data captured in the system on your actions, the insights are tailored to your use of Deposco. No bolting on third-party software and no painful translation process that may not even work.