February 5, 2016 Forecasting methods

Research, Information, Studies & Advice

Find a
article

Receive news from the site Forecasting methods

Enter your email address

Sign up
Unsubscribe
Validate Demand forecasting is an approach that involves using
qualitative or quantitative methods to estimate consumption of
Mailing
products in the upcoming periods. According to these estimates, we plan to
list, liste inside the company the production and anticipates the launch of the
manufacturing of products in order to reduce delivery times. The
forecasts in this particular context help to produce quantities
Download our sheets close to the actual demand.
Logistics - supply chain
Maritime transport A- The components of demand and methods of
Air transport forecasts
Road transport
Rail transport A1- The components of demand
Transit - customs clearance
The request can be broken down into four distinctive elements:
Employment - career The trend: it materializes the evolution of demand over the long term.
Logistics profession term
LogisticsMagazine The cycle: it determines its evolution in the medium term, in relation to
to the economic or sectoral activity
Seasonality: it highlights all periodic variations.
influenced by time (seasons, months, days)
Random residuals: this refers to the set of non-variations
Useful links explained by other factors
MRP Method
Technical data
Determination of lot sizes
Calculation of the economic lot A2 - forecasting methods
Model
supply Forecasting methods are divided into two groups: the methods
Planning of the qualitative and quantitative methods.
production
PDP, Program Qualitative methods are essentially based on opinion.
production comparison and judgment. It includes:
Calculation of net needs
The opinion polling method (surveys of sellers,
distributors of products
The method of comparison (or historical analogy. Forecast by
comparison with similar products sold in the past;
The Delphi method (or Delphi technique. Response to a series
questions by a panel of experts;
Market studies (application of a questionnaire to

Unable to access external URLs. 1/5

February 5, 2016 Forecasting methods

potential consumers in order to anticipate changes in the

market

Quantitative methods rely on demand extrapolation.

in the past using data from past consumption. Here-
Below, we present a non-exhaustive list of methods
quantitative

Simple quantitative method (taking demand into account

currently more or less a certain percentage)
Moving averages method (average of actual demand)
from the most recent previous periods;
Exponential smoothing method (weighted average by)
exponential coefficients) ;
Trend method (linear projection, exponential)
logarithmic or polynomial of the past trend. Starting from the
simulations, we choose the one that best fits the appearance of the
request
Decomposition method (decomposition of the result of
trend forecasts, seasonality, random effects
Regression and correlation method (combined use of the
least squares line and correlation with a variable of
dependence

All these forecasting methods are not exclusive. They can

to be combined to decrease uncertainties and strengthen the reliability of
results.

B- The general approach to forecasting

Demand forecasts are initially made at the link level.

located downstream of the supply chain (retail stores):

They are then gradually passed on to the links.

suppliers (supplier warehouses, central warehouse of the factory,
responsible for the industrial and commercial plan, responsible for the plan
production director, head of procurement...)

They have a specific periodicity. Depending on the quantities of

available stocks and thanks to the needs calculation application,
they facilitate the planning of store supplies
on the upcoming periods;

In practice, they are systematically done at least once.

once a year for products intended for sale. For stock
internally, they are rarely executed for all articles, we
prefer to apply them to well-selected categories (the
the most profitable, the most expensive, the most sensitive items,
brief the stocks designated as critical;

Unable to access the linked content for translation. 2/5

February 5, 2016 Forecasting methods

When the curve of usual sales shows a trend towards the

growth or regression, forecasts require revision
of thesupply models, minimum stocks and
maximum of the articles.

C - The quantitative forecasting methods

Above, we briefly present the quantitative methods of

most used forecasts in industrial companies and corporations
of distribution.

C1 - Forecasts using the moving average method or

moving averages

The moving averages method relies on the use of the average

previous consumption for a given number of periods. Its
The advantage is that it sufficiently mitigates demand fluctuations.
while preserving its overall appearance. It only takes into account the
consumptions or actual sales from previous periods.

The choice of the number of periods depends on the pattern of demand (demand curve)
sales or consumptions

C2 - Forecasts by the simple exponential smoothing method

The smoothing method takes into account the forecast for the period.
previous. In this forecast, the sudden gap is increased, weighted by a
coefficient α between 0 and 1.

If we consider:
Pn-1 (forecast of the previous period n-1)
α (smoothing coefficient)
Dn-1 (real demand from the previous period n-1)

Forecasting consumption for period (n): P = P n n-1 + α (Dn-1 - Pn-1

)

The choice of the value of α is made through trial and error. The chosen value
is the one that minimizes the forecasting error.

In practice, the coefficient α is close to (1) when the demand is

very fluctuating. However, for a stable demand that does not show
no significant cyclical variation, this coefficient is closer to (0).

C3- Forecasts by the least squares method

Also known as the decomposition method, forecasts by the method

least squares decompose the value of future consumption
in three factors:

T = trend of consumption or least squares line;

Cn = cyclic coefficient (seasonal coefficient or coefficient of
Unable to access the content of the provided URL. 3/5
2/5/2016 Forecasting Methods

seasonality). It is expressed as a percentage;

Rn = residual value of the period. It is expressed in
percentage.

Consumption forecast for period (n): P = T x C x R n n n

The trend is materialized

by a line in the
graphical representation of
consumptions. It presents
the pace general of the
consumptions throughout the
analysis period. It is the
projection of this line on
the future periods that allow
to estimate the quantities of
gross consumption

The cyclic coefficient is a

numerical and estimated value in
percentage. It corresponds to
one variation cyclic
increasing or decreasing
of a chronological series

The residual factor represents

the influence it could have on
upcoming consumption
the set of events
unusual or even totally
[Link] who of a
generally provoke
an increase or a decrease in the
request regarding the
forecasts

The calculation of the trend, the cyclical coefficient, and the residual value
is explained on the sheet:forecast of consumption (pdf)

C4- Forecasts using the needs synthesis method

forecast

This method is the most commonly used by distributors. Its principle is

very simple. In a distribution network organized in supplier/customers,
It is a matter for each supplier store to sum the forecasts.
which are transmitted to him by all of his clients:

Forecasts are starting in direct sales stores

(retail) according to the methods explained above (methods
qualitative and quantitative);
The obtained data is transmitted to the supplier store;
Suppliers of the same level gather the
forecasts of their clients into one and in turn transmit them
consolidated data to their provider;
The information thus rises across the entire network of
distribution to the production plant. The data obtained from
the factory represents for it the forecasts of the

[Link] 4/5
February 5, 2016 Forecasting methods

request.

[Logistics Method] [Logistics Forum] [Warehouse announcements

logistics] [Glossary-logistics] [Warehouse Images]
(c) TheLogistics Consulting Group

Unable to translate the content of the URL provided. 5/5