Service Delivery Cost per Farmer vs Scale

Strength of Relationship 5/5

  • Strong relationship between driver and outcome variables
  • Results are consistent across analytical models used
  • Very few limitations regarding sample or indicator

Key Messages

Of all the drivers analyzed, the scale of a smallholder-inclusive business model is most strongly associated with the service delivery cost per farmer; even when controlling for the effect of other variables, the larger the scale, the lower the cost. In other words, FarmFit’s data suggests scale is the most effective way to drive cost efficiency. At the same time, it also appears that the intensity (number of farmers managed per staff member) is often reduced at larger scale – in some cases at the expense of quality.

While these findings are aligned with FarmFit and our peers’ expectations, the data and insights are valuable in confirming this relationship. It highlights that while small-scale models can be helpful for piloting (new) interventions or business models, medium and large-scale models serving at least 1,000 farmers, and ideally over 10,000 farmers, are likely to be much more efficient. In addition, it confirms the importance of continuing to develop innovative approaches that mitigate quality trade-offs at larger scales.

We believe the results are due to three main reasons:

  1. Efficiency and economies of scale
  2. Business model maturity
  3. Intensity and quality of service provision

Keep reading to find out more.

Understanding how scale relates to cost

This analysis might be the most straightforward one in the FarmFit Insights Hub, both in terms of expectations and results. Economies of scale are widely accepted as drivers of efficiency.

It’s no different in this case; our data strongly suggests that the more farmers a business works with, the lower the service delivery cost per farmer. Specifically, models working with fewer than 1,000 farmers have a significantly higher cist per farmer than those working with a higher number of farmers.

The differences are substantial; scale seems to be strongly negatively related to service delivery cost per farmer. These results are validated using more rigorous machine learning methods. (Read more.)

Over 80% of our dataset consists of scattered-plot business models working with many individual smallholder farmers. However, the FarmFit team has also looked at other models such as block farm and nucleus farm models, all of which have, on average, much higher service delivery cost per farmer and are considerably less likely to work with 10,000 or more farmers (large-scale in our categorization). This could distort our results; given that these specific high-cost models tend to be small or medium in scale, they could skew our results. However, even when we control for this variable, the results not only hold, but strengthen further. (Read more.) 

Links to other outcomes

When studying the relationship between scale and the other two outcomes analyzed in the Hub, we find that:

  • For Direct Cost Recovery, our data suggests that on average, larger-scale models recover a larger percentage of their service delivery costs to farmers. Click here for more details 
  • For Farmer Value Created, we see the opposite result: Smaller-scale models are associated with more absolute value created for farmers than medium or large-scale models. This has important implications for this analysis - while scale can help business models operate more efficiently, part of the reason for lower service delivery costs can also be due to reduced intensity and/or quality of service delivery. Click here for more details 

In the following sections we dive deeper into possible explanations and nuances behind these results, as well as their implications.

Diving deeper: what might explain these results?

Our quantitative and qualitative data suggests there are several compelling reasons for why we see these results.

  1. Efficiency and economies of scale – Like businesses in other sectors, we see economies of scale. Working at a larger scale allows companies to work more efficiently
  2. Business model maturity – Business models that are at an earlier stage of maturity, such as the piloting or initial set-up phase, exhibit higher costs (for example, in initial design or testing of features). Larger-scale models working with over 10,000 have frequently already passed these initial stages of maturity
  3. Intensity and quality of service provision – Larger scale models may reduce the intensity of service provision, such as a lower staff-to-farmer ratio – resulting in a lower cost, but at the expense of quality.

Implications – so what does this mean for you?

Based on our findings to date on this topic, we see the following implications for different audiences:

Reflections on data limitations and further research

The Hub is an interactive resource that we are constantly updating with new data, new analysis, and validation by our partners. For the results on this page, we would like to emphasize the following:

Major caveats and limitations of our current approach

  1. Our work relies on averages across multiple years
  2. Our analysis cannot fully attribute the effects of efficiency and reduced quality or intensity of service delivery

Next steps: Updating FarmFit’s findings

  1. Control for measured and projected data
  2. Incorporate scale as an outcome indicator

Suggestions for additional research by our peers and partners

  1. Research innovations that can help scale