Saturday, June 27, 2009

Demand Planning - Essence of Forecasting - Part III

In this session, we will look what is CPFR (Collaborative Planning, Forecasting and Replenishment), Consensus forecast and right forecast method during recession period.

Collaborative Planning, Forecasting and Replenishment (CPFR)

CPFR is a collaborative business practice that enables partners to have visibility into one another’s demand, order forecast and promotional data to anticipate and satisfy future customer demand. This is done through systematic process of information and knowledge sharing. IT plays pivot role while sharing the information among partners.

In the past supply chains were burdened by Forecasting, Planning and Inventory systems that were disintegrated from each other, lacked accurate and timely demand information and blurred visibility beyond immediate trading partners. CPFR simplifies and connects overall channel demand planning by providing a single, real-time plan of forecast and supply. The objective of CPFR is to enable channel retailers, distributors, transporters and manufacturers with the capability to synchronize total supply with total demand from one of the channel to the other.

How does CPFR Work ?

Step 1: Initiative begins with collaborative partnership between two or more members of supply chain (retailer, distributors, transporters, manufacturer) with the intent of creating the technical and operating management architectures necessary to address the existing gaps impeding the synchronization of critical supply chain information.

Step 2 : CPFR partners in the upstream (manufacturer) agree to share critical demand information detailing what products are going to be marketed, how they are going to be promoted and merchandized, pricing and when sales cycles are to begin.

Step 3 : Each partners in the downstream (retailer, distributors, transporters) agrees to implement techniques that provide for the real-time sharing of channel inventory levels, point of sales (POS) transaction and internal supply chain constraints. In addition, each trading partner is responsible for ensuing continuous forecast and inventory accuracy as well as database update.
When these requirements have been fulfilled, the illumination of unnecessary inventory buffers and hidden bottlenecks in the distribution network flow should be revealed and initiatives put in place to eliminate them.

Consensus Forecast

There are so many stakeholders like Sales & Marketing, Operation, Purchase, Logistics and Finance in the demand planning decision today. They all need to be involved and agree in a forum to finalize the forecast number. They all need to contribute and may do so at different levels. The level may be product family versus SKU. For example Operation is interested to look forecast data at SKU level whereas Finance and Marketing at product family level. Generally the consensus forecast discussions happen at product family level.

- Operation to confirm production flexibility and capability (capacity planning) to meet the sales projection as per agreed timelines
- Purchase to ensure availability of raw materials and other components in time, to ensure smooth operation
- Logistics to ensure timely availability of infrastructure like warehouse, transportation
- Finance to ensure that sales projected volume are converted to value and find the gap or shortfall against the targeted revenue. This enable Sales & Marketing to plug the gap through appropriate field and Marketing activities.

The S&OP team develops the consensus forecast using information from the strategic plan, the current system forecast, production capacity planning and current inventory levels.

The steps involved in the consensus forecast (bird view).
- The revenue projection from strategic plan (expected turnover) are converted to item unit demand using standard cost information.
- These estimates of item unit demand are modified using historical demand information as well as the latest sales information, marketing activities and other parameters .
- Members of the S&OP team then agree on the consensus forecast according to their capacity. For example if Marketing team plan to run promotional activity during a period can be turned down by purchase due to non availability of materials.

The owner of the demand plan needs to see all inputs in one view. Scenario analysis is key to this process. ‘What if we increase this forecast by 15% next quarter?’ ‘ What is the risk if this customer demand doesn’t materialize?’ Iterate, analyze, decide. This has to be done fast with accuracy.

Let us understand the consensus forecast concept through example.

Generator manufacturing company strategically plan to achieve Rs. 100 Crores sales from generators (House hold and industry models) for the year 2009. This value is converted to item units based on standard cost. Let us assume the quantity to be sold for the year 2009 is 40 Lac units at Product family level. This 40 Lac units further broken into house hold and industry models (say 25 Lac and 15 Lac) and period (monthly, quarterly) wise. At product family level the generators are expected to grow at 8% over last year.

However the actual quantity sold in a period differ from both strategic plan and the forecasted number. House hold model may grow at higher rate (say 12 %) and industrial models may grow at slower rate (say 5%). As per company practice the finance department work out the excess or shortfall of revenue on monthly or quarterly basis against the projected turnover and share the information. In case of shortfall the Sales & Marketing department devise a plan to improve the sales of slow moving product (Industrial generator) through marketing and sales strategy. Despite the marketing plan, the sales are not picked up for industrial model generator. For short term the marketing and sales department focus to improve the sales of house hold generators to meet the annual turnover and this lead to deviate from the projected sales volume. In the final analysis, some product groups may have a lower consensus forecast (industrial generators say 11 Lac instead of 15 Lac) while others have higher forecast (house hold generators say 29 Lac instead of 25 Lac). This change in sales plan should be ratified by production and purchase department as they need to change the production and procurement plans accordingly. Hence any change in sales volume should have the concurrence from Finance, Production and Purchase department. This leads to emergence of consensus forecast.

As discussed in earlier session, the “best fit” forecast was done by package using different statistical models at SKU level and hence there will be less variation against the actual resulting better inventory management. In case of the consensus forecast, various departmental head participate in a forum and decide the volume keeping in mind the company expected growth at product family level backed up by Sales and Marketing plan. This volume is percolate down to SKU level through top down approach by using SKU contribution analysis. Hence there will be variation against the actual, resulting considerable inventory holding.

What is the right forecast approach in the current recession time ?

This topic was discussed in the Infosys blog and it is worth reading. The link is given below.

Sunday, June 21, 2009

Demand Planning - Essence of Forecasting - Part II

In my earlier blog, I was explaining about forecast fundamentals and in this article we will explore more about Forecasting techniques on Quantitative and Casual with various methods along with time series components.

Quantitative Technique

Uses historical data. In this technique single variable (eg., Sales / Demand) is used.

Time Series Methods consist of

Moving Average
Exponential Smoothing
Holt's Model - Exponential Smoothing Adjusted for Trend
Winter Model – Exponential Smoothing Adjusted for Trend and Seasonal
Times series decomposition
Times series extrapolation

Casual Techniques

Uses the relationship between demand and some other factors (income, population etc) to develop forecast. This uses several independent variables rather than the historical pattern of the time series. These independent variables contain information useful to explain the sales patterns of the dependent variable. For example increase in disposable income (independent variable) lead to increase in sales (dependent variable). There is a correlation between disposable income and sales. This information may help to build a forecasting model more accurate than dependent on sales history alone.

Methods used in Casual Technique

Correlation Method
Regression Model
Econometric Model

However in some cases more than one technique is used . For example ARIMA uses casual (Auto Regression) and Quantitative (Moving Average) techniques. This method is used in advanced forecasting tool.

Time Series

Time Series is a set of evenly spaced or continuous numerical data which is historic in nature, where the basic demand pattern varies little between years. It assumes the factors influencing past will continue in future. It uses statistical models as forecasting tools.

Components of Time Series

Trend – Long term (several years) tendency of a series to rise or fall (upward or downward trend), due to population, technology etc.
Seasonal – Periodic fluctuation in the time series within a certain time period (within a year). These fluctuations form a pattern that tends to repeat from one seasonal period to another. This is due to weather and customs (Diwali, Christmas)
Cyclical – Generally occur over a larger time interval (2 – 10 years duration), and the length of time between time successive peaks or troughs of a cycle are not necessarily the same. Due to factors influencing the economy.
Random – Eratic, unsystematic, "residual" . Random noise or error in a time series. Due to unforeseen events like strike etc.

We focus our attention to Trend, Seasonal, Random components rather than cyclical as it occur over larger time interval.

Times series data contains systematic and random components. Systematic is a combination of Trend, Seasonal and Level. Level is a time series data without having Trend and Seasonal components.

Before using the sales / demand data in a time series one must understand the components present in the data. It is not necessary that all components like Trend, Seasonal, Cyclical, Random should be present in the data. This understanding is critical in choosing the right forecast method.

Given below the table which shows the forecast method to be followed against the given components in the data.

Forecast Method -----------------Applicability
Moving Average --------------------No Trend or Seasonality
Simple Exponential Smoothing ------No Trend or Seasonality
Holt’s Model ------------------------Trend but no Seasonality
Winter Method----------------------Trend and Seasonality

How to choose right Forecast technique and method ? Given below the diagram which shows how the forecast technique and method to be chosen based on the available data.

Acid test for the forecast personal lies in selecting correct forecasting technique and “best fit” forecast method.

Let us compare car buying process against forecast technique and method selection process. For example if a person would like to buy a car, first he will decide on segment like small size, medium, luxury, SUV and then he will choose the brand based on performance (trouble free) and other criteria. Trouble free performance can be compared against “Least Error” in statistical parlance. Hence Forecast personal will choose the right technique first and then select the “best fit” method among various methods which gives least error.

Forecast Technique : At the outset it may look easy to select Time Series as one need to forecast on historical sales data. But after careful consideration, other inputs from Marketing (promotion, Pricing) and sales (Field activity, competition) need to be considered with weightage while forecasting. This observation may lead to opt for casual method due to usage of many related variables.

Forecasting method : Selecting right forecast method is cumbersome process due to availability of wide variety of methods. For example in Time Series lot of methods, as explained above are available. User need to plot the data in all methods chosen and then select the "best fit" method based on result.

What is “Best Fit” Forecast method ?

Forecast personal plot the historical data in the various forecast methods chosen based on technique. He compares the result / output of each method and choose the method which gives least error. Basically he used to compare the MSE (Mean Square Error) & MAPE (Mean Absolute Error Percentage) of each method and then select the best fitted forecast method. This opens up a project scope in Six Sigma on Forecast method.

Above discussion clearly indicate that the forecast personal should familiarize with Statistical techniques, able to understand the output and interpret the results clearly. This helps the top management to plan their Sales and Marketing strategy effectively.

In my next session we will explore about CPFR (collaborative Planning Forecasting Replenishment), Consensus Forecast and analyze right forecast method during recession period.

Friday, June 19, 2009

Demand Planning - Essence of Forecasting - Part I

This is my second article on SCM in this blog, addressing the issues related to Forecasting. Forecasting has been considered as one of the top supply chain issues in the globalized world. Organizations are striving to predict customer demand as accurately as possible. Accurate forecasting kick-starts demand and supply chain planning

What is a forecast ?

In layman language, Forecast is a statement or inference about the future usually based on historical data. Decision can be taken based on intuition or guts feelings also. In general, the decisions with supporting evidence are widely accepted.

Why Forecast ?

Can't we do away with Forecast ?

Is Forecast necessary Evil ?

Let us study the selling pattern of two computer manufacturers ie., Dell and Wipro to understand the forecasting pattern.

Dell is selling the computers directly to the customers based on their indent. The customer specifies the machine configuration and make the payment through online. Based on the users inputs the computers are configured and shipped to the customers directly as per agreed delivery schedule say 3 weeks. In this case, Dell is reacting in response to the customer demand (“Pull”) , need to keep meager quantity of components (inventory) at their disposal. Delivery schedule to the customer is decided by Dell based on lead time in getting the components from other manufacturer, Assembling and delivery time to the customers. This is an example of make to order with zero lead time as the customer is willing to wait for finished product. This model requires minimal forecast.

In Wipro, the customers pick up the computers from the retailers and hence they need to keep different configuration machines. Wipro need to know in advance, the model (desk top, laptop etc) and geography (state, area, town) wise estimated volume (quantity) to be sold for a particular month, to keep the finished stocks ready at their end before beginning of the period. Wipro reacts in anticipation of customer demand (“Push”). This model requires lot of forecasting effort as they need to ensure that the plant/factory is having the capacity to produce the forecasted (estimated) volume for the month and also to ensure availability of components to produce the saleable goods.

Benefits of Forecast

¨ Increase customer satisfaction
¨ Reduce stock-outs
¨ Schedule production more efficiently
¨ Lower safety stock requirements
¨ Reduce product obsolescence costs
¨ Manage shipments better
¨ Improve pricing and promotion management
¨ Negotiating superior terms with suppliers

Forecast Techniques

Qualitative :- They are primarily subjective. Rely on Judgment, opinion, intuition, comparative techniques or surveys. It is non-scientific nature and hence difficult to standardize or validate for accuracy. This type is used in case of Marketing intelligence

Quantitative : - Uses Time Series ie., set of evenly spaced or continuous numerical data which is historic in nature, where the basic demand pattern varies little between years. It assumes the factors influencing past will continue in future. It uses statistical models as forecasting tools.

Casuals : - Uses the relationship between demand and some other factor (Sales Promotion, discount, Retailer incentive, income, population etc.) to develop forecast. Correlation method, Regression and econometric models are used in this technique.

Types of Forecast

Short term Forecast – This is for short period say 2 to 3 months basically used for operational or production planning. Since the short term forecast drives business on an ongoing basis and has a direct impact on the financial performance of an organization. Ideally, Sales team is responsible for short term forecast.

Long term forecast – It is for the period 1 to 2 years and it is meant to meet the financial obligations. Marketing team is responsible for long forecast.

Characteristics of Forecast
- Forecasts are always Wrong. This is true because forecast is an estimation and not actual.
-Long term forecasts are less accurate than short term forecast.

-Aggregate forecasts are more accurate than disaggregate forecast. It is easier to forecast more accurately at product family level than at SKU level. For example Maruti company can forecast better for all models put together than selective models.

-Information get distorted when moving away from the customer. When your distribution network contains intermediate channels like distributors and retailers, the demand information passes through them will get distorted. In Marketing parlance this is termed as “bullwhip effect”. It is like ripple created in pond when you throw a stone.

Steps in Forecasting Process

- Establish objectives for the forecast. Reduce the Forecast Error or Variation.
- Determine what to forecast. Volume forecast for Production Planning and Value forecast for company value growth.
- Specify the Axis and Level. Forecast works on Four Axis ie., Product, Customer, Time and Fact. Fact could be volume or value Forecast. Level for each axis varies as per organization classification. For example Product Axis can contain SKU, Brand, Family Group etc.
- Gather historical data for particular period ( 2 to 6 years) and analyze. For example if Maruti company sales were increased in Nov’08 due to 10% discount scheme for all brands. The sales data for Nov’08 is distorted due to discount scheme and this data should be normalized before using it in forecast.
- Select a forecasting method. Various models like Moving Average, Exponential Smoothing, Weighted Moving Average, Regression, Arima, Winter, Holtz etc are available. Utmost care should be taken while deciding the suitable method.
- Make the forecast.
- Present or Interpreting the forecast results.
- Monitor and control the forecast.

Of all the above steps selecting the forecast method and interpreting the forecast result correctly is the herculean task and it requires specialized effort. Statistical error measures (e.g. MAPE, MSE) are popular yet widely misunderstood and misinterpreted. Quite often, end-users are not equipped to interpret the forecasting accuracy through such error measures.

In the next session we will explore about Forecasting technique and methods along with Time Series. Looking forward to your comments and queries.

Wednesday, June 17, 2009

Supply Chain Blog

This is my first article on SCM in this blog. I’ll try to address the issues related to supply chain in lucid way to make novice to understand the concepts clear and grasp the essence.

What is Supply Chain ?

Here goes the story of Elephant and Five Blind men in comparison with Supply Chain.

Supply Chain consists of all functions involved directly (Sales, Production, Purchase, Inventory, Warehouse etc) or indirectly (HR, IT etc) in fulfilling customer request.

Supply chain Management consists of three major components :-

- Supply chain management is a set of approaches utilized to efficiently integrate suppliers, manufacturers, warehouses, and stores
- merchandise is produced and distributed at the right quantities, to the right locations, and at the right time, in order to minimize system wide costs
- satisfying (customer) service level requirements Let me explain the concept in a simple way.

Consider Domino’s marketing department advertising that they can supply quality pizzas to their customers within 30 minutes against the order receipt. The company has issued guidelines to all departments to reduce the cost to increase the bottom line growth. If Logistics department which is responsible for distribution adhere to this guideline, it will jeopardize the marketing objective of delivering the product within 30 minutes. This is where the supply chain personal plays crucial role to integrate different departments to work towards company objectives by aligning the department objective with company objective. Supply chain process can be classified into three macro processes.

Integration among the three macro processes is crucial for successful supply chain management.
There are certain functions in an organization considered as core and others are supportive from Supply Chain Management perspective. Effective integration of these main and supportive functions lead to successful supply chain management.

Generic Supply Chain view from FMCG Sector

I will explain these concepts in my subsequent articles .

SCM is the buzzword and emerging trend in the industry. Top Management team believes that
effective supply chain is panacea for most of their maladies. It is true to certain extent.
Success of Supply Chain depends upon the effective integration of various functions and to the
extent the support it gets from top and other management personals in its decision making.
Apart from this, there are other factors which are responsible for successful supply chain
management. These will be discussed in latter issue.

From Indian scenario, I feel we have not truly embraced the Supply chain concepts. Most of the
companies viewing SCM as a remedy to their critical functions, rather than integrating and aligning all core and supportive functions.

I would like to invite comments on this article.

In the next issue I would like to share my view on Forecast & Demand Planning.

Sunday, June 14, 2009

SCM Forum

I am krish. I am an XLRI, Jamshedpur (India) Alumni specialized in Logistics and SCM. Also holding six sigma green belt certificate from ISI, Delhi. Having more than 15 years of experience in Logistics and SCM and 2 years in IT Department as Application Developer in FMCG industry. Currently I moved to Telecom Industry.

I would like to share my experience and knowledge related to Logistics and SCM through this forum and also keen to use my skill set to address the users query.  This blog covers SCM best practices followed in FMCG and other related sectors. 

I would like to publish my first article related to Forecasting shortly and expecting viewers to share their opinion.