BIG DATA AND S&OP – CAN THE TWO COEXIST?

For some years now, one of the main topics on the agenda for business leaders has been S&OP, or sales and operations planning, in Portuguese. Conceived in the 80s, this process was born with the main objective of breaking down barriers between functional areas, ensuring that all company planning points to a common goal. The recent evolution of this concept, of the supporting technology, and the increased maturity of the process in several companies in Brazil and in the world gave rise to the concept of IBP (Integrated Business Planning), with a slightly broader scope and, notably, with a greater participation of finance.

But despite being around for nearly three decades and already established as a great way to manage a large company, the importance of S&OP is still not fully understood in the Brazilian market, even in large companies. The concept hits the market in waves of interest, and lately we've felt that the subject is back on the upswing. However, today a new name enters the competition for the eyes and ears of executives, appearing repeatedly in social media discussions and specialized articles: the Big Data Heralded as the new wave in business value generation, the expression sums up the countless new ways of capturing extremely granular data and generating previously unattainable information. Like all other trends, this one is now vying for companies' time and resources, and we're starting to see references to Big Data considerably exceed references to S&OP.

As enthusiasts of integrated planning, we have begun to assess the extent to which these two concepts compete, and where they converge. The purpose of this article is to explore the synergies of the two processes, showing how the arrival of Big Data can create new ways to create value in the integrated planning process.

WHAT IS BIG DATA

According to the McKinsey Global Institute[1], Big Data it is a set of information whose size is larger than typical database software can capture, store and analyze. The definition is deliberately subjective and changes over time, as the capacity for processing and storing data evolves. According to Moore's Law, computing power as we know it doubles every two years. Thus, a more precise definition, in terms of bytes, of what fits the concept, would easily be lost in a short space of time. In addition, there are also differences between market sectors, as the commonly used tools and the amount of data that each market is used to dealing with are quite different.

In today's world, where more than 600 million smartphones are connected to the internet, 30 billion pieces of information are shared monthly on the Facebook, and points of sale have more and more embedded technology, the sources of information collection are abundant. This new avalanche of data can be both quantitative and qualitative in nature. As examples, we can list citations of products or establishments on social networks, consumer opinions about the quality of products on blogs and forums, product evaluation sites, cell phone geolocation information, transactional information at points of sale and information on loyalty, which can generate insights valuable insights into consumer behavior and preferences.

for IBM[2], there are four main dimensions that define what is Big Data: Volume, variety, speed and veracity.

• Volume: Even though we don't have a precise threshold, we started talking about Big Data from terabytes (1000 GB) to petabytes (1000 TB) of data to be analyzed at once.
• Variety: Data can be collected in different ways, structured or not, such as blog texts, podcast sounds and images on social networks.
• Speed: The information is generated in real time and should ideally be analyzed at the same speed so that decisions can be taken in fractions of seconds, such as traffic information for GPS systems to recalculate the fastest routes for the user.
• veracity: In most cases, there is no filter of who is generating the information or what is the source and veracity. Data qualification mechanisms must be employed to ensure consistency, completeness, updating and accuracy.

Figure 1 – The dimensions of Big Data
Source: IBM

Simultaneously with the new sources of data, new technologies for storing and processing this information appeared. Several software providers rushed to develop new solutions that catered to this new niche. There are several lines of development, but most use approaches such as non-relational databases and parallel processing. As an example we can cite the Hadoop, a platform open-source originated in Yahoo!, which uses an approach of MapReduce to enable the processing, storage and analysis of large volumes of data in parallel. This is basically done by breaking the data into small batches that can be processed on several “commodity” machines at the same time. This technology goes hand in hand with cloud processing, which allows several simple machines to work together, processing large volumes of data at low cost.

UTILITIES IN S&OP

It is clear the value generation potential of the insights that can be extracted from this unprecedented mass of data. According to the same IBM study mentioned above, approximately 49% of the companies that work with Big Data focus on bringing targeted results to their clients. We can cite, as examples of results achieved, the loyalty programs for shopping in supermarkets, which can suggest specific promotions from customer to customer depending on the time of year. Or an automaker, who have access to the behavior of their customers behind the wheel and develop technologies that bring more safety and improve preventive maintenance of vehicles.

But how can S&OP, a tactical planning process that by definition must look at the company relatively aggregated, benefit from data as detailed as the daily sales of a product at a point of sale, for a specific customer? Do we need a new planning paradigm, better adapted to the oncoming data avalanche? The subject is still very new and there are few reports of companies actually starting to use this information in their integrated planning processes. We venture, then, to suggest some ways in which the Big Data can be used in favor of a more robust S&OP.

Both in strategic definitions and in monitoring the execution of plans, the analyzes bring new directions that would have been impossible before. We see the main benefits in four major groups: understanding demand, launching new products, customer satisfaction and monitoring execution.

UNDERSTANDING THE DEMAND: Every good planning process starts with a good sales forecast. Knowing market behavior and how the company's actions impact demand is essential for a good balance of operations, and detailed consumption data can bring insights valuable. With more detailed customer information, it is possible to develop new channel classifications, with greater geographic and consumption segregation, with the possibility of having different classifications for defining promotions, discounts, service time and focus of the sales team. This allows for more surgical marketing actions, with a more scientific approach to measuring their impact on consumer behavior. The analysis of consumer reaction to one marketing action per day, by SKU and by point of sale, added to qualitative information from social networks, brings valuable information on the impact of each action in each market, with a clear analysis of the return on investment for each type of stock. This facilitates the modeling of new initiatives for the future and supports assumptions of increased demand, mitigating the risks of under and over storage. New consumer trends and opportunities for cross selling can be detected locally and extrapolated to your regional market, indicating opportunities for targeting the sales team to reinforce specific products in the portfolio in the coming months, aligning with distribution and production to ensure availability.

LAUNCH OF NEW PRODUCTS: Estimating sales at the beginning of a product's lifecycle is a major challenge for the planning team. As much as we have several sample surveys of product acceptance and projections from the marketing area, its real acceptance in the market is always a question mark. In these cases, the best approach is to plan as well as possible with the data in hand, and closely monitor execution to raise the right alerts for the operation when sales deviate from the plan. With real-time information on the consumption of new products at retail checkouts, for example, it is possible to respond more quickly to the uncertainties surrounding these launches and guarantee availability. With a view of what is actually being consumed, the “bullwhip effect” is avoided and it is possible to quickly and effectively redirect the geographical distribution of products (directing supply to markets that have a better acceptance of new SKUs), and trigger emergency actions of production and purchase of inputs.

CUSTOMER SATISFACTION: The service level, together with the sales forecast and the cost of serving the market, make up one of the three main groups of strategic indicators to be monitored in the S&OP. Monitoring social networks allows you to capture various information regarding customer satisfaction regarding a product or service. Exact numbers are difficult to estimate, but it is known that only a small part of dissatisfied consumers make formal complaints to the company's ombudsman. However, it is common for many to express their opinion through social networks, boasting to anyone who wants to hear how dissatisfied they are with a product or a service. The use of tools text mining  to measure the incidence of words like “delay” and “delay” next to the company's brand can trigger specific action plans to improve service time. To illustrate the strength of social networks, according to research by NM Incite [1](joint venture between Nielsen and McKinsey, created to study the impact of social media on consumption), at the end of 2012, almost 60% of Americans between 18 and 24 years old seek customer service through social networks and 71% of those who receive good service through this channel recommend the brand to others. A full understanding of customers' needs and expectations should direct supply chain costs, and the study of this market "sentiment" is one of the most vigorous areas that form around the Big Data

Figure 2 – Summary of Big Data benefits
Source: McKinsey Global Institute analysis

EXECUTION FOLLOW-UP: As we mentioned earlier, more detailed information on end customer consumption allows for increased visibility of the chain and faster and more assertive planning of how much to keep in stock in each location, without depending on the inventory policy or the quality of the retailer's storage systems. . A good example are products that are missing from supermarket shelves and that appear in stock but have registration problems. Analysis of differences between sell in e sell out in these retailers can raise alerts regarding the risk of out-of-stocks or product expiration. When the product supplier has visibility into the end customer's consumption and not just the retailer's orders, it will be possible to diagnose this type of problem in real time, when, for example, an unexpected drop in sales happens. This underpins the risk and opportunity analyzes that are the foundation of the decision-making process in S&OP.

THE RISK OF ANALYSIS PARALYSIS

Even with the increase in analysis capacity, it is important to remember that resources remain limited, such as the capacity to invest in IT, the number of people to translate analysis results, and the short deadlines for decision-making. There is still a huge shortage of professionals qualified in data analysis in the market. Thus, one of the dangers of Big data – and more specifically for the planning processes that start to use the information – is to spend a lot of time generating analysis and information, without being able to reach any conclusion in time to actually take action that adds value.

Within the analysis period of the S&OP process, which usually occurs in monthly cycles, it is necessary to understand what happened in the operation in the past and plan for the future, including assessing the impacts of current planning deviations. If the amount of data starts to grow too much, and the possibility of analysis opens up too much, it is very common for the S&OP team, responsible for supporting the managers' decisions, to not be able to generate all the analyzes in time, or to get lost with the “curiosity” to better understand all the novelties that begin to appear.

Therefore, it is necessary to focus when structuring how the work with the Big data within the company. Investment in IT is not the only solution to leverage your company's results with the use of Big data. Sometimes excessive investment in IT can even hinder the evolution of market understanding processes and operations. The ideal is a combination of team, technology, processes and focus.

CONCLUSION

The use of Big Data can bring significant value to many areas of companies in the most varied industries, from financial services to health services. According to the previously mentioned McKinsey study, the good use of this new information would correspond to an increase of 300 billion dollars for the American medical system, 250 billion euros for the public administration of the European Union or up to 60% of margin increase from American retailers.

Figure 3 – Sectors and types of benefits obtained with Big Data
Source: McKinsey Global Institute analysis

The S&OP analytical and decision-making process is a great way to convert information into action and significantly improve the balance between sales and operations, ensuring that the entire organization is aligned to a common goal. This makes it a major client for the deployment of Big Data technologies and, like any major client, it must be included in the project's requirements survey.

However, caution is needed in order to transform this information into real value for the business and guide decision-making and action plans based on the new insights. This change should be seen as a turning point and not just as an organic process improvement. It is necessary to define the objective to be achieved with the new information, what types of new analyzes should be possible and how to generate them. People must be sought who have the analytical capacity to transform terabytes of data into business information, dimension investments in capture technologies, analyze and store data and rethink the company's processes. It will be necessary to convey the necessary confidence to the management team, who will increasingly have to base their decisions on current data instead of looking only at historical information.

Above all, it is necessary to assess the current level of maturity of the planning process before including Big Data dimensions, and ensure that we are already extracting everything we can from the current scenario before taking the next step. You have to walk before you run.

—

[1] “Big data: The Next for innovation, competition, and productivity”, McKinsey Global Institute, 2011

[2] “Analytics: The Real-World Use of Big Data; how innovative enterprises extract value from uncertain data”, IBM Institute for Business Value, 2012

[3] “2012 NM Incite Social Care Survey”, NM Incite, 2012

Author
Diego de Souza, managing partner of Plannera – Soluções em Planning
Caio Rodrigues, project manager at ILOS – Institute of Logistics and Supply Chain