RA Demand Forecasting Services

We take care about your forecasts…


The importance of good forecasts

The ability of a retail organization to forecast demand accurately is the primary condition for optimizing business decisions. The optimization of order quantities, stock levels, or store shipment allocation depends on the aptitude of a retail operator to forecast accurately demand at store/sku level. The optimization of merchandizing initiatives is only possible through the implementation of industry specific causal demand models, where the impacts of price changes, promotions, displays and ads on to future demand can be measured with sufficient confidence.

Demand forecasting is a business process

Although retail organizations recognise the importance of good forecasts, few implement an effective formal process to support it involving suitable resources, skills, methods and tools to ensure the production of accurate and relevant forecasts. Also, little effort is often spent on the systematic monitoring of the quality the different information sources used to build demand models.

Retailers are not good forecasters

The production and management of quality demand models and the generation of accurate forecasts is a specialists’ field, which requires in order to be run successfully expertise in many domains such as statistics, models building, mathematics, data-mining, management sciences and information technology, together with in depth business knowledge of the different retail sectors. However, our experience shows that most retailers fail to get the science and the tools right, sometimes even after investing heavily in hiring statisticians and acquiring complex and expensive software solutions.

Let the experts work it out for you

RA Demand Forecasting Services provide an effective, no risks and cost efficient solution for the production of quality demand forecasts customized to your need. Using our SAAS build solution you will be sure that:

  • The right demand models will be built to meet the specific needs of your industry and organization

  • Sales, price, promotion, display and ad information will be methodically analysed and cleaned to make certain that demand models are built using sufficiently good and relevant data

  • Forecasts at the right granularity level will be computed and made available to your organization

  • Forecast accuracy will be monitored. Historical forecasts and corresponding actual sales will be stored

  • Lost sales will be estimated to derive achieved service levels

  • Detailed reporting will assist your organization in tracking

    • Data quality and issues

    • Demand models robustness

    • Demand forecast accuracy and relevance to support the supply chain and merchandizing decisions

What you have to do

Initially we shall work with your IS and merchandizing teams to help you identifying, collecting and formatting the information sources such as past sales, product information, past and future price, promotions, display and ad information necessary to build the models and compute the forecasts. We shall also agree on the process for channelling data and reports to and from our portal. That is all!

RA demand forecasts production process can be summarised as follows

  • Evaluate your business requirements and accordingly select suitable demand model forms

  • Clean and analyze data

  • Build and maintain demand models

  • Evaluate models quality

  • Produce demand forecasts

  • Monitor demand forecasts relevance and accuracy

A Complete service

RA Demand Forecasting Services follow a comprehensive set of processes ensuring that:

  • The data supplied is of sufficient quality

  • Suitable retail specific demand models are implemented to meet the specifics of the retailer’s industry sector, distribution channels or products types; taking into account in particular:

    • Retail sectors

      • Fresh grocery

      • Dry grocery

      • Clothing

      • Tools (DIY/Garden centres)

      • Decorations and furniture

      • Electric and electronic appliances

      • Entertainment

      • Petrol retailing

    • Distribution channels / modelling of cross-channel demand transfers

      • Stores

      • e-commerce

      • Mail-order

    • Product types

        • Seasonal/permanent

        • Fast/slow mover

        • Fresh/Dry

    • Dynamic advanced causal models

        • Price elasticity coefficients (regular & promotional)

        • Dynamic effects (e.g. promotions dips, promotions frequencies, carry-overs, TPR’s decays )

        • Cross-effects (cannibalisation effects, products affinities)

        • Promotions types

        • Display types

        • Ad types

    • Customized granularity

        • Time (day/week/month)

        • Product (SKU/demand groups)

        • Stores (single store/price or promotion zones)

    • Weather impact evaluation (optional)

        • Product type

        • Geographical areas

        • Weather alerts

  • Numerous reports are systematically generated to account for:

    • The data quality and relevance

    • Models robustness

    • Forecasts accuracy

    • Lost sales estimation and achieved service level calculation

A scalable and evolutionary service

Our “Grid Analytics” data processing and model building platform has been designed to scale to the largest retail organization thanks to complying to latest parallel processing design principles and allowing the access to private or public clouds computing network. All data supplied is encrypted.


Benefits & ROI

RA Demand Forecasting Services represent a low cost, no risks, and effective solution for the production of accurate forecasts within a comprehensive, secure process. Accessing RA Demand Forecasting Services require

  • No acquisition and installation of software

  • No payment of annual maintenance and update fees

  • No acquisition and maintenance of hardware

  • No training

System Integration requires only the sending and receiving of scripted files using an https based connection. We continuously invest in developing better methods, architectures, models and practices to make RA Demand Forecasting Services the best tool for retail operators. Access to our latest development once tested and validated is available to your organization as part of our service with nothing more to do on your side.

The service is available on a pay as you go basis, where the tariff calculated depends upon

  • The number of products modelled

  • The number of forecast runs computed

  • The number of forecast periods

To quantify the benefits of using RA Demand Forecasting Services, we propose to build an ROI analysis table, which will assess and compare the following:

In the ROI table above, section (3) and (4) estimate the inventory costs savings and lost sales reduction gained for every MAPE’ s percentage point drop (i.e. forecast error) obtained. Such analysis, which is typically based on comparing the demand forecast accuracy between the existing “in-house” system/function and RA Demand Forecasting Services, never fail to show that RA Demand Forecasting Services cost less to deploy and run than a “within the firewall” solution, and helps to significantly reduce the inventory holding costs and lost sales by forecasting demand more accurately.

Easy to try
As an initial step, we suggest to bench mark the existing demand forecasting application against RA Demand Forecasting Services on a sample of few hundred products (typically a product category), using similar measures as the ones described in the ROI table above.  Such evaluation is very simple to run and will provide valuable information to any retailer willing to conduct it.

 Design by AdiumDesgin.com