Valimotie 27, 00380 HELSINKI
Author: Johanna Småros
Director, Scandinavian Operations, D.Sc. (Tech.)
Many companies regard sales forecasting as a troublesome and wholly unpleasant chore. The truth of the matter is, however, that forecasting is not rocket science, and a wide range of effective and easy-to-use tools are available to support it. The greatest obstacles for the development of forecasting seem to lie in peoples’ attitudes towards it. As such, it is time for us to bust some myths about sales forecasting!
In the text below, I have compiled real and perceived problems in sales forecasting that I have encountered in many companies, along with various views on how to tackle them.
Not long ago, I visited a company that was manufacturing consumer products. The director in charge of managing the supply chain introduced me to the company’s operations and production logic. I ventured to inquire as to how forecasting was handled within the company, to which the director replied: “We don’t use any systems for forecasting. Forecasts are never accurate so we’ve seen it as more prudent to invest in flexibility.” What flexibility meant in the case of this particular company was an end product inventory whose capacity matched a year of consumption. No one can deny that the word flexibility has a nice ring to it but, like so many things, it comes at a price.
Most companies rely on forecasts to control purchasing, production, finished product inventories, and capacity. The accuracy of forecasting is thus directly linked to efficiency and the quality and level of services. Even largely inaccurate forecasts may be enough to get by, but without sufficient accuracy, the company will be forced to compensate for uncertainties in demand with hefty buffers of
Naturally, it pays to develop flexibility, but perhaps it would be a good idea to put a few hours of the week into forecasting or acquiring the proper tools – which is ultimately the most financially sound choice for your company?
The reality in many companies is that forecasts are made for different purposes or edited from different perspectives all over the organization. Even if there is only one official sales forecast, it may be modified and interpreted into a number of different iterations along the way. The typical state of affairs is that production has learned not to trust the systematically overly optimistic forecasts churned out by marketing. In one company, for example, the primary tool for forecasting was an Excel spreadsheet, which automatically sliced 20% off the assessments of the marketing department. In situations such as this, it is perfectly understandable that the people crunching the numbers begin to have some serious misgivings about the feasibility of the arrangement.
It seems plain as day that production should be decidedly less equipped to predict future sales than – you guessed it – the sales personnel. Therefore, the question is how to work your way out of a situation where the people in acquisitions, production or inventory feel that they should take charge of managing things outside their own area. In my experience, there are two common denominators among companies whose forecasting is in order: clearly assigned forecast responsibilities and constant monitoring of the forecasts.
The monitoring measures enable the companies to learn from their mistakes, but learning also requires a great deal of motivation, which is where the responsibilities come in. It is often thought that the responsibilities are defined upon deciding who will create the forecasts, when and how. However, the questions of who, how and when only serve to describe the actual generation of the forecast – they have very little to do with assigning responsibilities. A forecast responsibility can be considered as defined only when the person making the forecast carries the responsibility for its repercussions or, in other words, possible overproduction, shortage of merchandise or, for example, the failure of a campaign that was incorporated into the forecast too late. As long as the acquisitions department is chewed out for overstocks of packaging materials or the production department is reprimanded for surplus production, a sharp buyer or production manager will be sure to tamper with the forecasts.
Fictitious example: A salesman is requested to provide an assessment of the sales of an upcoming campaign. His answer is: “A lot!” When he keeps being pestered to be more specific, he snaps: “How should I know? Figure it out yourselves – you know as much about this as I do!” Why would our poor anguished salesman be so rude and incompetent to boot? This is because he has agreed upon a significant sale with a customer, and he has neglected to set any concrete goals for the transaction or to even begin to assess the risks involved. Fortunately, salesmen like this are few and far between, or are they?
I have seen more than my fair share of situations where various measures (product launches, campaigns and orders of seasonal products) are taken without specifying any respective goals for them. If no goals are defined, monitoring the effects of the measures is pointless, which leads to nothing being learned from previous measures and money being routinely spent only to repeat the same mistakes year after year.
In the top companies, forecasting is an integral part of meticulous sales planning. When determining what needs to be done to reach the sales targets (what products and measures will be used and which customers will be invested in), the preliminary forecasts are automatically generated in the process. The forecasts will naturally need updates along the way when we see how the various plans play out, but updating is a cinch when the basic groundwork is in place.
And yes, the above model does effectively demolish the main argument that the sales people have against the forecast responsibility: “The time I spend on forecasts will take its toll on customer contacts and sales.” Just to be clear, forecasting does not detract from sales but, when done right, creates the foundation for profitable sales!
The saying “the perfect is the enemy of the good” has rarely been as felicitous as when used in reference to sales forecasting. Upon suggesting calculated forecasts to companies, I have often received the following reply: “It wouldn’t work. For instance, we simply lack the information on what our competitors have been doing at the same time. In addition, no data on product displays and customer notifications is available. Even the weather seems to affect the results.” When considering investments in forecasting tools, it seems as though it is not enough for a tool to improve the efficiency of the forecaster and the accuracy of the forecast; the system needs to be perfect or it will be dismissed as unnecessary. Due to this, key customer managers, who are indeed among the most important resources of companies, spend a good chunk of their time making routine updates to forecasts.
Tried and tested models for forecasting the demand of seasonal products are readily available. They may not yield the perfect result every single time, but they produce a solid assessment, update it weekly (or daily, if necessary) and observe the situation in an objective fashion without becoming too attached to the product being sold. How many of us forecasters can make the same claim?
Surprisingly enough, calculation models can also be extremely useful in situations where experts are at their strongest. Regression models for campaign forecasting (i.e. comparisons with earlier corresponding campaigns) have proven themselves many times over. Moreover, calculations have been extremely beneficial in challenging circumstances, such as during Christmas sales in bookstores where demand explodes in a short time span and the product range is almost completely renewed on an annual basis. (How exactly calculations are useful is another story, which I will touch upon in another text.)
What am I trying to say:
What should you think about at this point: “Can my company afford to neglect the development of forecasting procedures?”
We at RELEX have extensive experience in making forecast management more efficient in partnership with an impressive roster of companies. Together with our customers, we have successfully worked on sales planning and forecasting models as well as the use of quantitative forecasting models.
Take the first step towards better forecasting and contact us: email@example.com / +44 7546 124031. An hour’s meeting is enough to go through your company’s situation and to define the first steps!