In Demand Management Now! we talked about how the intense competition for product companies and retailers and the squeeze on margins challenge companies to get a lot better at planning.
Demand Management Technology — Market Positioning
Now we will look at the market along several dimensions. Today, we have long-term domain leaders who focus on demand management only, some in suites with other supply chain or customer-facing applications, or suite ERP players. In addition, and probably most importantly, their industry focus is really what sets them apart.
Who are the players? What do they do? Who do they serve? This topic will be tackled in a number of ways. Major market position takes a higher level view of the Demand Management players in the market according to major attributes and the markets they serve. Later, we will look at the fine grain capabilities and differentiate between retail and manufacturing players.
What Are We Trying to Do in Demand Management?
If we think about what we are trying to do in Demand Management today, there are two foci that converge. One is understanding the customer. The other is understanding products (Figure 1).
One day it may be less important how and where customers buy, but since the integration of information may take very divergent paths according to retail, wholesale, or manufacturing, it is convenient to break out market position that way, knowing full well that manufacturers often have their own retail outlets. This, of course, challenges how we traditionally position players in the market today.
Another element that is revolutionizing the demand management world is mobile/web combined with location-based or geospatial views of the customer. This new source of customer data gives us an important way to understand the customer, which may be more important than traditional demographics.1 Significantly, the kinds of data we will collect and how we analyze them will require major changes in systems and how they integrate. New spaces will emerge based on this data. Conceptually, this should be easy for demand planners to understand, since demand management, de facto, is all about forecasting events. What is not easy is how these systems will communicate with each other.
Soon we confront in our data analysis that where they are and what they are doing will change how we define customers, how we can predict demand, and how we might influence or incent customers based on that data. In our positioning, we do have a few players who are taking that leap forward. Firstly, they are looking at web data: traffic and clicks and so forth. This was the domain of web analytics and marketing software, but had not made its way over to the demand planning players. But that is changing. In addition, social sentiment — which requires inference, semantic analysis, and evaluation of relevance is starting to emerge.2
Demand vs. Supply
As mentioned, the bulk of the market is split between manufacturing-centric solutions and retail-centric solutions. Mergers and innovation often blur the lines for some firms who offer all types. But specialty — or what we call domain expertise — still has an important place, since the requirements of an industry, micro-verticals, or business processes are so exacting. Solution providers have the technical expertise and knowledge to help with a successful implementation. And their dedication to the body of knowledge — taking an almost academic approach to developing new methods — makes the user’s ongoing investment with these companies important.
In Figure 3, we focus in a bit on manufacturing-centric vs. retail-centric capabilities and what one might expect as functional capabilities.
The variety of industries and the process peculiarities of some industries challenge demand planners. On one hand, we have the consumer packaged goods (CPG) industry that has honed its forecasting skills over decades and has made a lot of progress. The so-called build-to-stock crowd lives by an assumption that, ultimately, the products will be consumed, although we do know there is waste, markdowns and obsolescence. We used to call this world push. CPG’s ability to understand true customer demand often is hampered by their distance from the customer in B2C.
CPG can be either process — food, cosmetics, home cleaning products; or repetitive — hand tools, razors, cigarettes, etc. Many of these products have very long lifecycles with only minor modifications over decades, and replenishment planning is a major activity. Whole solution categories — rapid or short-term demand planning — have grown up to address this very problem. Terra Technologies got their start by tackling this short-term replenishment problem in a new way.
Fast fashion has a short lifecycle, with some attribute planning (style, color, size), but generally, there are no plans for replenishment. Mistakes in fast fashion, whether in attributes or pricing — once made — are not retrievable. Firms like First Insights aim to tackle this type of challenge.
On the other hand, we have the configure/build-to-order modes typical of more customized discrete products from auto, industrial, and high tech. This world is more pull. In the build-to-order world — B2B — we tend to have the specificity of the customer’s order, but competitively, we need to provide a reasonable lead time so that there is still safety stock in these chains. A firm like Steelwedge provides opportunity planning, which pulls notifications from the sales team as they gain knowledge of a potential order. Since these are notoriously difficult to forecast, many engineering-to-order companies just don’t do product furcating in the short-term sense. They do use product family, attributes, attach rates, and such to analyze the value of features and components and how well those sell on their own as well as contribute to the sale of the overall product.
Solving those types of problems has challenged planners and providers alike. But as providers focused on those industry or process problems, they developed methods and3 solutions for many of the above.
Figure 4 aligns some of the solution providers with the types of manufacturing problems they solve.
Being closer to the consumer has some benefits when it comes to creating insights. However, the retailer is dependent on stable and responsive supply chain partners. Some retailers, although they might have a private label, consider their dependency on others a trade-off when compared to the complexity of managing their own chain.4 Thus, they have an interest in supporting the supplier’s replenishment efforts.
The retailer, with few exceptions, is also burdened with the ultimate decision of what the market might buy and building their whole strategy around that — from store locations and formats, to merchandise decisions, to consumer pricing, and liquidating the non-sold items.5 They are most likely to leverage the location-based customer insights data to help with short-term issues such as promotion and pricing.
In the Collaborative Zone
We use the phrase Collaborative Integration to call out activities and functions that require a high level of partnering to achieve. Trade promotions and replenishment are examples. These are most effectively done through collaboration and a high degree of data sharing.
You will note that build-to-forecast which is repetitive manufacturing, consumer goods, etc. has a rich population of solution providers from which to choose. Social sentiment is an emerging space, but the collection and analysis of customer insight — which looks at customers and consumption, and interest (web clicks, twitter mentions, etc.) is being embraced by solution providers who support manufacturers, not just those who support retailers. For example, Logility, who does not do in-store planning but does do customer insights.
So far, we have positioned solution providers from the vantage point of a higher level. We will telescope in a bit more in the next issue. However, we encourage you to build your requirements and RPIs from the lists and concepts we have discussed, since it is the fine detail of how these solutions work and the product maturity in these major market areas that become most important.
Think Process Support
Features and functions aside, demand planning is a process that frequently involves multiple sites, departments and trading partners. So the platforms and tools — workflows, role-based views, UI, and methods to aggregate information — are very important. After all, if you have good math, but can’t share it effectively with others, you won’t have a meaningful process. Often users do not evaluate this aspect effectively and still wind up with less than stellar implementation. Pretty screens and math are not enough. How well the solution brings the constituents together to solve problems, evaluates data in an understandable way, and facilitates the workflow may be just as important as the math. Remember that some ERP providers’ declarations about integration are based on mythology, since today, they are composites of multiple acquisitions and reside on multiple platforms — database, memory resident servers, web — as well. Thus, buyers should evaluate demand solutions from ERP providers on equal footing with domain leaders.6
Users need to give thought to how the data is organized and shared within the solution. For ‘one-number forecasting,’ the slicing and dicing challenge has been addressed by several providers. But processes like S&OP are at the top of the heap when it comes to multi-multi-multi- problems. Of late, several solution providers have built their S&OP in such a way that it can effectively interface across the landscape of people, as well as roll-up and feed data back into various supply chain planning modules to keep things in sync. The multi-site challenge does lead some companies to buy a package just for S&OP.
Steelwedge has focused on this problem for a decade. Initially just S&OP, they have, of late, expanded into other planning problems. Logility’s S&OP can be bought as a separate module and can interface with other solutions, too.SAP’s S&OP on Hanna is also a separate module and is bought as a standalone capability.
Cloud and What It Means Today in Demand Management
Demand Management seems to be one of the last holdouts in cloud. Several providers, such as NetSuite, NeoGrid, and First Insights have been cloud from day one. During the last year we saw several solution providers such as JustEnough, JDA, Logility, Steelewedge, and SAP provide some cloud offerings. Due to the nature of the data, which mostly is not multi-enterprise, these solutions gain less from cloud, since except for collaboration, they don’t have the network of trading partners who work together on one process as you might see in the transportation market.
Cloud has several variants that are useful to discuss in terms of the demand-planning market. Though ‘cloud’ may be offered by solution providers such as JDA and Logility, they are predominantly providing their customers with the customer’s own licensed version of software and data. They can take advantage of cloud economies in the sharing of some resources, but basically, this is a modern version of hosting.
Multi-tenancy has another variant that seems to be gaining some headway: the multi-tenant, many instance. That is, the solution provider is offering a single, configurable code base, but each customer has their own servers and databases. You see this with Vecco, Orchastro, Terra, and Steelwedge. The nature of the processes they are supporting — multi-site, distributed, yet sensitive demand data — requires a unique server configuration, as well as unique methods of data partitioning.
A network of collaboration platforms, such as offered by NeoGrid, allows several enterprises to share data. So software and data may be shared, though strict partitioning occurs, of course. Usually, these are one to many portal type implementations. That is a retailer or large brand company wants to collaborate with their partners, so much of the data is within the context of the owner’s portal. This is the multi-tenant, single instance variant. Most of the demand planning activities are so enterprise specific that this is not most companies’ choice.
Will License Remain?
Most assuredly. If you look at the bulk of the installations today, the world has forecasting packages with a long history that are inexpensive to support and do basic forecasting blocking and tackling work. In-the-box forecasting packages still sell briskly.
At the high end of the market, huge number crunching applications for merchandise planning, pricing optimization, etc. still tend to be licensed onsite efforts.7
In our next installment we will drill down into the Retail sector.
2 Later in our series we will tackle some of the new math being used in demand management. — Return to article text above
3 It is interesting though how many retailers get this wrong, that is they don’t understand their customers. Witness recent challenges for long-term retailers such as Talbots, JCP, etc. — Return to article text above
4 We do see a growing trend for retailers to take more ownership of supply chain activities from product design through transportation. We will discuss this topic in an upcoming the brief. — Return to article text above
5 Neither the manufacturing nor retail models are absolutes, as manufacturers may have their own stores and retailers may make their own products. They also have wholesale partners who may own the inventory burdens and aftermarket liquidation of last season’s inventory. All these choices drive your solution design. — Return to article text above
6 The stories we hear about providing the software for free — as part of the ERP — are also disappointing. The solution provider, as a partner, should be solving the business challenges you are trying to address. Demand Planning should be an evaluation done by the planners, with IT support, and not solely an IT decision. When millions of dollars’ worth of products and customer relationships are at stake, that should be the focus rather than the cost of the software, although that should be within reasonable boundaries, of course. — Return to article text above
7 No doubt, over time, the cloud services provided by Amazon and Microsoft, HP and IBM can provide a foundation for elasticity and ensured performance.
To view other articles from this issue of the brief, click here.