Prophecy™

Demand Forecasting knowledge

+44 (0)1494 785574

Contact Us!

The Role of Judgement in Demand Forecasting

The role of judgement in demand forecasting image

Forecasting is a Critical Business Process

Forecasting is a critical activity for the majority of companies. Forecasts drive production plans, financial plans, sales force targets and budgets. Improved forecasting has attractive financial benefits. 

The pay-off from a modest investment in improved forecasting can be significant. 

A well planned business, one which can act with confidence knowing that it has an effective quantified plan to deliver its financial goals, can focus on the important, more strategic issues rather than day-to-day fire fighting to obtain arbitrary targets. 

 Forecasting Involves Many Users

Every company's forecasting, planning and budgeting process is likely to be different. However, one thing that they all have in common is that each process typically involves more than one user. Users can be forecasters or reviewers. Users can be recipients of forecasts from other functions, which they then add to or use in another business process. 

Any opportunity for allowing multiple users to work with Forecasts, plans and budgets in a more effective way will result in faster turnarounds, lower costs and better quality forecasts, plans and budgets. Prophecy enables you to integrate your forecasting, financial planning and budgeting processes - (although it is equally at home specialising in just one of these areas). 

It is technically quite difficult to deliver multi-user, multi-function forecasting, planning and budgeting systems. 

Developing your own solution has many hidden complexities, which have already been solved in a comprehensive commercial product like Prophecy. 

Demand planning spreadsheets are also limited in their ability to respond to the challenge of multi- user, multi- function concurrent access to data . 

 Statistics Can't Produce Accurate Weekly or Monthly Forecasts in Many Markets

The ideal solution to the forecasting challenge is a forecasting system which we can leave to generate accurate forecasts, allowing us to get on with the process of running our business. 

Unfortunately, the reality in the majority of markets is that statistical forecasting systems cannot generate accurate weekly or monthly forecasts. 

Why? Because the peaks and troughs in sales are caused by events which will appear as random events to a statistical algorithm - things like promotional activity, competitor activity, sales force incentives, advertising etc.. These events occur at different times of the year and each event is both unique in its sales impact and in the synergies with other events which occur simultaneously. (For example, ask yourself whether the same promotion, repeated, has the same impact on customers as it did first time around. It can't, because it is customers' perception of the promotion, rather than the ‘physical’ nature of the promotion, which drives sales - a subtle but significant distinction.) 

Therefore, in the real world, each promotion or other sales impacting event typically has a high degree of uniqueness. There is almost never the same situation twice. 

Statistical methods are helpful in identifying the underlying or long term trend in the market, customer or product. However, the size of the short term peaks and troughs versus the trend in many cases is likely to be so significant as to make the impact of small errors in the long term trend forecast largely insignificant.  It's those peaks and troughs you want to forecast more effectively.

Therefore, in most cases, there is no magical 'silver bullet' to the forecasting challenge, at least for the short term weekly or monthly forecasts. 

 So, What is the Solution?

If statistical forecasting is inappropriate to short and medium term forecasting, then the only source of accurate forecasts has to be those developed by the business managers themselves. 

Judgemental forecasing in demand planning

The business knowledge and market judgment of these managers needs to be focused on evaluating and interpreting past history and applying the understanding gained to known future events, such as promotions etc.. 

This type of forecasting is called 'judgmental' forecasting. In practice, it is the only way to generate short or medium term forecasts in real world markets.

So, let's focus on what business forecasters may need to help them apply knowledge and judgment to the forecasts: 

They need a user-friendly, flexible environment which removes as many of the manual processes associated with collecting, formatting, calculating and manipulating forecast and actual sales information. 

They need to be able to table and graph their forecasts and past history in the ways which are most suited to the task at hand and their own preferences for viewing data 

They need to be able to save textual information alongside the forecasts and actuals, so that the reasons why numbers came out as they did or forecasting assumptions get saved for future forecasters and analysis (so that the lessons of the past are learnt from rather than forgotten). 

They need to be able to understand their forecast accuracy quickly, and with no manual processes - the system must be able to generate accuracy analysis at the drop of a hat. 

Above all, they need a system which is easy to use and which minimises the 'hassle' factor associated with business forecasting.

In addition, any forecasting, planning and budgeting system needs to integrate well with your existing systems - so that actuals and hierarchy information can be uploaded easily, and forecast data downloaded for use in Production Planning or other applications across the information supply chain. 

And, of course, the system you choose must be cost-effective. That is, easy to implement without huge consultancy bills and easy to maintain once the system is up and running. 

How can an environment like this be delivered? 

 Are Spreadsheets the Solution?

Spreadsheets for demand planning are inappropriate because, however cleverly they are utilised: 

  • They are essentially single-user solutions. Most forecasting, planning and budgeting activities come with the overhead of needing to share or publish data to several users or functions. 
  • They do not cope naturally with the need to aggregate data over hierarchies of products and customers. 
  • They result in a lack of control over data. Data becomes fragmented across numerous different spreadsheets and numerous different spreadsheet owners, creating a reconciliation nightmare. 
  • It is hard to automate the loading of actuals data or new products and customers. Manual intervention and rekeying is the norm. 
  • Spreadsheets provide an essentially two-dimensional view of data, whilst users need much more flexibility. 
  • They do not allow you to allocate user security in a flexible or robust enough way - i.e. by product or account, by forecast or budget measure, by time period, by user type.

As a result of some or all of these limitations, forecasters will spend a substantial proportion of their forecasting time on the mundane activities associated with maintaining their spreadsheets and invest insufficient 'quality time' on refining the quality of their forecasts or analysing their accuracy. 

 How Prophecy Can Help You

You could get one custom-written for your organisation. But: 

How cost-effective would it really be, when Prophecy is here today and was written from the ground up to match the solution proposed on this page? 

How soon could a robust solution be delivered? A forecasting project has substantial complexity - for example, Prophecy contains nearly 200,000 lines of code. This represents a large programming project, but it is available today, and fully tested and proven! 

Data Perceptions and Prophecy are here for the long term - you have the assurance of continued support and development for your investment. Internal developers move on or leave - however thorough their documentation, continued quality of support for the system is likely to be affected. 

Prophecy offers numerous ideas which have been proven in action to provide major help and significant time savings for judgemental business forecasters and planners, including: 

  • Flexible and powerful multi-dimensional OLAP style reporting and drill down.
  • Forecasts at any level in Product and Customer hierarchies.
  • Forecast for any number of measures - for example, from Volume to Profit or Profit to Volume, within the same model!
  • Powerful mathematical and modeling aids to forecasting and business planning.
  • Bottom up or top down forecasting (or middle up and down!), budgeting and planning.
  • Flexible graphs.
  • Graphical forecasting - change forecasts via 'drag and drop' on graphs.
  • Statistical forecasting - of course - to 'sanity check' your judgmental forecasts versus the long term trend, for example!
  • Store, compare and analyse different versions of the forecast.
  • Flexible Accuracy Analysis at the press of a button
  • 'Expert system' style commentaries on the forecast - 'outlier' forecasts identified for you by the system.
  • Many forecasters can concurrently access a single database to forecast.
  • Flexible, powerful and easy to use provisions for allocating security to users - by product, customer and fact. Provide Read/Write, Read Only or No Rights (i.e. hidden) on a per-user basis.
  • Support for remote users too!
  • Ease of use - Prophecy is targeted at business forecasters and planners, not statisticians or computer boffins!

In addition, Prophecy is not expensive compared to anything else out there. It is very attractively priced, easy to implement, and easy to administer. 

 Summary

If you're reading this, you recognise the importance of applying quality improvement to business forecasting, planning and budgeting. 

Ideally, the (related) processes of demand forecasting, financial planning and budgeting should be integrated. However, frequently they are not, introducing much potential latency and unnecessary manual effort. 

There are many excellent statistical forecasting software tools available. However, whilst comprehensive in the range of algorithms they offer, these systems do not recognise the reality of business forecasting and planning - that short term peaks and troughs versus the long term trend are driven and dominated by complex behavioural factors - such as promotions and competitor activity. They are excellent tools, but applied to the wrong problem. 

What is required is a system that supports 'judgmental' forecasting, which takes the drudge out of retrieving and analysing data, and which allows sales and marketing managers, finance managers and general managers to work cooperatively on their components of the forecasting, planning and review process, within a common, shared but secure systems environment. 

And where judgmental forecasting is not required (low volume/value items etc.) you can apply the 80:20 rule and forecast these items less often or using Prophecy’s powerful statistical forecasting features.

Prophecy is a mid-priced / highly functional solution, which was written specifically to address these issues. You can implement it quickly and it is easy to develop and maintain.