Seasonal scenario planning for matching seed supply and demand: MS Excel decision support tool

This is an MS Excel tool for seasonal scenario planning. This tool correlates historical seasonal rainfall with seed sales. Using expected rainfall forecasts the tool gives an estimate of expected seed sales in the coming season. The tool should be introduced to decision makers in a workshop, presented jointly with "seasonal scenario planning: workshop format" and "seasonal scenario planning: workshop slides".

Many seed companies offer various varieties of the same crop, targeting different agro-ecologies or market segments. For example, farmers can choose between early-maturing, average, and late-maturing maize varieties. In many years, however, some high-demand varieties sell out, while low-demand varieties are eventually shipped back to seed companies, who need to re-package or destroy the seed.

Because farmers often select which varieties to purchase just before the season, it is hard for seed companies to predict sales. Thus, every year, seed companies supply a similar, long-term average mix of varieties. This can lead to the suboptimal outcomes for both farmers and seed suppliers. Variation in farmers’ seed demand between years can, however, be linked to climatic events: in dry years, seed companies typically sell different seed quantities (for example, more early-maturing varieties) than in years with abundant rainfall.

Today, free online seasonal climate forecasts give an indication of total seasonal rainfall abundance with up to six months lead time. These forecasts provide probability estimates for simple seasonal scenarios (dry / average / rainy season). Using such climate forecasts, seed suppliers can anticipate farmers’ seed demand and adapt seed distribution accordingly.

Access the tool

In what context is this tool useful?

Seasonal scenario planning is useful not only to seed suppliers interested in meeting the demand for their existing products, but the principle of scenario planning could also be integrated into other parts of the seed sector, in particular breeding processes. A good understanding of seed demand under different seasonal scenarios could, for example, enable breeders to develop varieties that optimize overall performance considering the expected shifts in future scenario probabilities. For farmers, these cultivars would be risk-minimizing varieties, and different economic risk strategies could be considered during the breeding process, leading to different varieties.

Results achieved

This was just a proof-of-concept study. No practical implementation has happened yet. However, with the intention of identifying the potential way forward, for a more sophisticated version of the tool, this project involved a participatory design process involving seed sector decision-makers from two countries (Zimbabwe and Ethiopia). Over the course of the process, the ambitions, scope, and specific features of the decision support tool were specified and refined, in part diverting from original expectations. The project has generated valuable insights on decision-making processes in seed supply chains, and practical opportunities for better matching supply and demand throught the use of seasonal climate forecasts.

A main take-away is that, generally speaking, risk in seed supply can indeed be mitigated by using seasonal climate forecasts. As a proof-of-concept study, this project has shown that there is industry interest, among both public and private users. In a case study with maize sales data from Zimbabwe, we found that seasonal rainfall has explanatory power for varietal demand. Probabilistic seasonal rainfall forecasts are freely available online and are understood by decision-makers. This evidence underscores the possibility of using forecasts to anticipate varietal demand in practice.

Variations on this method

No variations of this tool have yet been developed, however, a key recommendation for practical applications in the future consists in moving beyond a generic, globally applicable tool, such as the one we experimentally developed. More regional, case-specific decision-support tools that consider locally relevant climate phenomena – going beyond seasonal rainfall – possibly at longer lead times can be developed together with climate scientists. Our low-tech approach to the decision-support tool, consisting in an MS Excel workbook with embedded weblinks, has generated valuable insights. Yet eventually, offering seasonal scenario planning tools as freely accessible one-stop websites would likely provide a more inclusive user experience.