Cow Alerts: Are They Trustworthy?

Author: Iain Dalton (IfA) - March 5, 2018

Critical to the effective management of data is the level of false positive alerts resulting in wasted time, poor decision making and a permanent loss in the level of confidence in the system.

Staff facing such inaccuracies will become sceptical of the system and adopt methods of working that do not fully utilise the benefits that data can bring to the management of the dairy cow. “Unpolluted” data is core to a successful business – particularly wherefarms rely on analysis of large quantities of data to make crucial decisions andultimately survive.

The solution to high levels of false alerts or indeed too few alerts is to access the system routinely and link all alerts with specific farm based physical cow knowledge, using the lists produced to check specific animals. Adding cow specific behaviour into data enables the system to incorporate such animal behaviour into its alerts.

In addition to the introduction of cow specific behaviours into the data set the alerts thresholds can be altered to ensure that the alerts level is correct, not too many and not too few. A continual process of feedback is required to ensure that the action lists produced are legitimate – data cannot and should not be used in isolation but worked with alongside and reacting with physical data/behaviour which is observed on the farm. This process can then lead to the formulation of farm specific operating procedures (SOPs).