Over the past 15 years, our industry has evolved, grown, and improved by leaps and bounds. Frequently, the rapid pace to the innovations and gains across the industry has been tough to stay on top of. While we’ve advanced considerably over the past few years, there are mammoth opportunities ahead in the near future. I’m convinced there is steep change on the horizon, rooted in learning from dairy farm “big data.”

The term big data refers to commingled data tables from various sources. As we’ve discussed in the past, there are many valuable data tables to wrangle, including, but not limited to herd management records, dairy production and component records, feed management data, financial records, and feed ingredient nutrition data. For some preceding thoughts on the topic, reference these prior “Feeding Fundamentals” articles: “Clean data can power your farm to profit” (in the August 2021 issue) and “Nutrition needs data-driven decisions fueled by economics” (in the December 2022 issue).

Looking for relationships

Turning the page to 2024 puts us closer than ever to unlocking new and novel understandings of factors that contribute to feed conversion efficiency, sustainability, and dairy economic performance. For example, feed management factors strongly relate to dairy performance and efficiency responses. This statement in and of itself isn’t novel or innovative, yet we need to understand the relationships more concretely.

For example, what is the feed conversion efficiency gain associated with another feed push up? The answer is out there, but we haven’t had the data in hand to develop the scientific and economic model. Advanced big data analysis, through human and machine learning and modeling, will uncover this relationship and others. Then performance and profitability gains will follow by investing in the right areas on your dairy.

Delivering the data

Setting advanced data analysis and machine learning aside, let's talk data dashboards to understand the path toward more profitable insights. We readily understand that having your key performance indicators (KPI) right in front of you on a daily or weekly basis is a must. The simplest KPI dashboard takes form with a manually managed worksheet, scraping data from various sources to collate data in a convenient spot. We’ve all likely organized data from different spots into a worksheet. It’s time consuming and cumbersome.

Stepping up from KPI worksheets, one might employ a data dashboard product or service. There are numerous groups that offer customized solutions to organize KPIs in a central location or even an eye-appealing dashboard. I’ve lost track of how many dairy data dashboard tools exist, but a new or alternative dashboard seems to crop up every three to six months in conversations.

The hype and discussion around the next dashboard or model might be warranted, but all these products or services rely upon the same basic foundation as the KPI worksheet described above. We need to source and clean up data from numerous spots.

With the dashboard, the data structure also becomes important. Think of data from different sources like shotgun shells of different gauges and the data dashboard being a 12-gauge shotgun. The data dashboard is powerful, but it requires a specific type of data to empower it. Poorly structured data is problematic for a dashboard, similar to having a box full of 12-, 16-, and 20-gauge shells but just a 12-gauge shotgun. This is a major challenge — and opportunity — for our industry.

This became abundantly clear over the past year as I worked alongside several colleagues in support of a high performing dairy. We came together and sought to determine why dairy feed conversion efficiency broke records in July 2021, under summer conditions that one wouldn’t have expected excellent performance from. We needed to join herd management, nutrition and feed management, production and components, and weather records together prior to embarking on this advanced learning endeavor.

Ultimately, our efforts fell short of expectations due to poor data structure and our inability to put all of the data tables together. We couldn’t get all of the data tables joined and structured, just like we couldn’t take 16- or 20-gauge shotgun shells to go hunting with our 12-gauge shotgun. Rather than be discouraged, I came away from this failure optimistic, as we’d clearly identified the ground we needed to cover prior to deploying advanced learning methods. This is the point we need to take to heart.

Closer to the goal

The mammoth learning and performance opportunities in 2024 and beyond begin with well-structured data. Appropriate data can then be fed into dashboards or used for machine learning, just like a box of 12-gauge shells feeds nicely into a 12-gauge shotgun. Bear with me on the redundant analogy and realize that our industry needs to invest in clean and well-structured data.

For those who approach your dairy with a data dashboard or learning tool, ask how the data are being pulled together and structured. Is the data management approach reliable and scalable? In many cases the answer is “not entirely,” but we’re closer than ever to achieving this goal.

Tying in dairy nutrition, many impactful factors for your herd’s feed conversion efficiency and sustainability are rooted in dairy nutrition and management data. These data take different forms, but through improving data structure and then learning from the data, we’ll uncover the new paths toward profitable dairy farming.