ToolsGroup celebrated its sixteenth year in business with their US customer conference in Boston.1 ToolsGroup has grown over the years, beyond an inventory optimization solution to a rich supply chain suite with an impressive worldwide customer base.
Pat Smith, Managing Director of ToolsGroup North America, kicked off the conference with a discussion, asking the audience what their definition of a Contemporary Supply Chain might be. Smarter, digital, connected, and global were examples of how the audience defined contemporary. Yet challenges remain to achieving such states.
Pat talked about the cross-industry convergence of ideas and practices — information and ideas from Omni-channel to the Internet of Things. Though all these ideas and data are available, the simple fact is that we need ways to analyze them and turn them into action. Thus, we will need more advanced technologies than we used in the past. Two key areas he highlighted were that:
Forecasting has to get a lot smarter to understand diverse customers and their changing needs.
And optimization is also a must, since the cost to serve has grown. This might not appear obvious to many, but with so many channels for brand companies to support — stores, websites and warehouses — along with the inventory, that they are spending more money on operations. Along with that, all the fulfillment expectations such as free shipping and same day delivery are costly endeavors. Consumers are fairly oblivious to this since they have been the beneficiaries of cheaper and cheaper goods along with these increased options. And that one-time promotion of goods or services will not, unfortunately, create loyalty. So besting past performance is a must to remain competitive in this Omni-world.
Pat continued with the many challenges companies face today. In order to keep up/stay ahead, they must use technology to achieve a nimbler, more contemporary state of operations.
Joe Shamir, founder of ToolsGroup and supply chain guiding spirit, talked about their evolving strategy towards a fully automated supply chain with something they call Servicestat (Figure 1). Underlying an automated supply chain enabled by machine learning are analytics and rules.2 Critical and exception data is presented to users for decision making. But over time, more and more of the forecastable and high volumes of SKUs can flow through a process like this.
Companies need to augment their traditional approaches with smarter math and machine learning to get more accurate data-driven supply chains, freeing up planners to work on the exceptions.
Figure 1: Future Supply Chain Automation
As I understand this, rich data sources from forecasts, customer data, and context information such as location or weather can all be part of an analytics platform that allows planners to create these smarter, often self-managing forecasts.
Cross-Supply Chain
ToolsGroup has made some significant updates to their systems by adding transportation forecasting, service parts planning, and expanding to other industries beyond retail and consumer products forecasting.
With Servicestat, as well, users can develop an end-to-end, industry-wide view of the chain. Why is this important? With Omni-channel the need to understand what to stock vs. what suppliers should stock and where to hold inventory in the chain for the most cost effective, yet responsive fulfillment model, is a must. Companies are just beginning to understand this challenge.
Advanced Forecasting
Significantly, ToolsGroup has added new, rich data sources such as weather and location data to forecasting, and has been providing social and web ‘big data’ to customers for a few years now.
It is extremely difficult to create forecasts by working with social data. It does take a few years of understanding and correlating the data, I learned from Joe Shamir’s comments, until companies can really create intelligence from it.
“Sentiment analysis systems attempt to use social data to analyze how much consumers like or dislike a particular product. More sophisticated sentiment analysis tries to discover exactly which aspects of a product or service that consumers do or don’t like, by how much, why, and the importance of each aspect.”3 Social data is also pretty dirty data, in my experience, with a lot of marketers gaming the systems. So a lot of learning and filtering has to be done to find the good data in all that. But there is gold there once you do.
Machine learning is a lot less tricky. Here, we are learning from past forecasts and actuals just how good our judgments were and how to improve upon them. Again, I advise some patience, since historical data is needed. In fact, to be successful with any planning system requires a year or two of historical data and the application of new forecasting techniques before results can be seen. That is when the real benefits kick in and become sustaining. This is true whether in allocations and merchandising, new product introductions, or those notorious long-tail products. And we want to forecast these long-tails, since unlike quick turn items, long-tails may be higher priced items (and also part of the selection customers expect). Examples of long-tail items vs. quicker turns in a department store, for example, are fine china vs. blue jeans.
If you are a Macy’s or an Amazon, you need to forecast both. Seasonal and allocation items also need unique approaches. Though sophisticated planning departments usually have different planners, buyers, and so on for these types of items, it often is less obvious to sort them all out. Below the surface there may be other nuanced patterns that apply to these different item categories (see Figure 2). Even the best planners just don’t have time to sift through all that data to discover those nuances, whereas machine learning can.
Figure 2: Machine Learning Model
Rethinking the “ERP”
For the last two years, I have been in conferences listening to customer presentations and listening to customers talk about how now that their supply chain systems are more advanced and cross functional, they are rethinking their ERP strategy. I don’t like the term ‘ERP for supply chain,’ but it has been so consistently used by companies that I mention it here.
Though they had hoped the ERP would provide them with an all-in-one solution, after more than ten years of hope they realized that is not the case and have moved on. But what, exactly, is moving on?
- Theory one is that the ERP can still be a system of record and the supply chain systems can manage all the inter-enterprise and planning work. However, it has been proven by many to be too difficult to get data in and out of the ERP fast enough. Or the data in ERP is just not the right data with which to manage the business processes.
- Theory two, which seems to hold more promise, is a co-existing world with a Master Data/Database/Integration platform (or supply chain platform), a Supply Chain suite, and a reduced ERP footprint. (Figure 3 has one such customer example that we heard about at TG16.)
Figure 3: Rethinking ERP with Supply Chain (SO99)
An alternative model is the Omni-channel retailer who never relied on an ERP and generally built their own commerce platforms (think Amazon, eBay, Wayfair and so on). These firms have also expanded their purchasing of supply chain suites, recognizing the importance of the fulfillment side of their business model.
Several ToolsGroup customers presented such cases at the conference. Executives are using the S&OP models, first strategically, to plan revenue and investments; and then tactically, as planners take over and actualize these plans to make real-world adjustments with the forecasting and optimization portfolio. Then execution functions such as logistics and procurement take over. The move to a cloud platform offering of these solutions allows companies to collaborate not just on planning but on seamless execution. This delivery platform is especially important in outsourced business models or ecommerce where the seller may stock only a small percentage of what is offered in the catalogue and rely on suppliers or fulfillment services to execute much of the fulfillment.
Conclusions
ToolsGroup has shown itself to be a very nimble but quiet, successful global competitor. I see a renewed vigor in them during the last few years with advanced analytics product areas that many other forecasting companies have not yet ventured into. This investment in product has paid off for them, not only in terms of adding new customer logos, but also with existing customers adding new modules and making significant software upgrades to their IT portfolio. Many of the customers I talked to at the conference were there at TG16 to learn about the ‘contemporary’ planning solutions that will help them meet the challenges they now face. ToolsGroup has reaped a lot of good will from these companies in the past, and these companies are predisposed to stay with them since ToolsGroup is keeping up and staying ahead, a key point for any software company that wants to thrive.
Invest or die is the key in the tech market. Death, unfortunately, does not come quickly, but is a slow withering death. Using many of the new big data ideas is still a challenge and the learning process takes time, as we mentioned. So for tech companies who have not already been investing here, the withering has already begun. ToolsGroup does not intend to let that happen to them — or their customers!
_________________________________
1 Similar events are hosted in the Norse countries, southern EU and Asia. — Return to article text above
2 ToolsGroup uses a rules engine from their partner, Rulex. — Return to article text above
3 Bill McBeath, ChainLink — Return to article text above
To view other articles from this issue of the brief, click here.