Case Study

Machine Learning and analytics provide valuable insight for Beef and Lamb New Zealand

Knowing how to apply the right machine learning algorithms to a specific business challenge takes a lot of skill and experience - and our team at Harmonic have both in spades. In this case study, we see how applying the right mix of machine learning algorithms and analytics can derive some incredible insights that go a long way in helping improve farm profitability in New Zealand.


Beef and Lamb represents New Zealand sheep and beef farmers. They bring farmers innovative tools and services, plus credible information and analysis to help farmers make the best decisions for their business.

The Challenge

A wide variation in profitability exists between farmers across New Zealand. Beef and Lamb wanted to understand what key attributes make the top 10% farmers profitable when others with seemingly the same resources struggle to break even.

The Solution

Harmonic carried out data cleansing and preparation of farm survey sample data. This ensured insights from the analysis could be trusted. A comprehensive machine learning exercise was then completed to extract meaningful insights from the data for further analysis.

Harmonic then designed and developed a model using a hybrid of machine learning methods. The model associates farm attributes with profit and identifies the most important profit drivers behind top-performing farmers (for specific farm classes). This approach not only uncovers numerous valuable insights, but it also identifies 13 key profitability drivers (from 900+ attributes), which accounts for 73% of the observed variation in profitability.

The Result

  • Gained reliable insight into the profit drivers and attributes that characterise top sheep and beef farms
  • Gained insight into how farm data is used and how it informs farm management decisions
  • An advanced analytical model that identified the top 13 profitability drivers of the top 10% profitable sheep and beef farms, across different land classes, from amongst 900 variables.
  • Uncovered unexpected but useful insights - e.g. Variation of profit is not related to the age of the farmer, level of debt or other common understanding.

“Harmonic’s robust analytical approach was very valuable to us. The resulting data is useful to support business planning and strategic development tools. We discovered some unexpected insights, gaining a more in-depth understanding of the key attributes, variables, and profit drivers of top sheep and beef farms.”

— Richard Wakelin, General Manager Farm, Beef & Lamb New Zealand