The value of long term data collection

One issue that comes up for us quite often is the need for data or records to solve an issue where none exists, or exists in a fashion that is so unstructured that it can’t be used.

Paper dockets, inconsistent excel spreadsheets that change structure year to year or vendor records that can not be processed by our clients local systems are common issues that we see, mostly leading to longer response times on breakdowns, higher costs of diagnosis and more often than not overall higher running costs.

We tend to provide the same piece of advice to many of these clients when they tell us that it is too late to get data. We use a Confucian saying that the best time to plant a tree is 25 years ago. The second best time is today.

What we mean by this is that although a problem often provides a catalyst that highlights gaps in systems, it also provides an opportunity to open a conversation to ask what could have been. This is often when the decisions that can avoid the next problems are take.

Data methods such as predictive analytics often step in at this point.

So given the next best time to plant a tree is today, what type of tree should we plant? That one is easy to answer. We plant the tree that provides the tastiest fruit. In out case this is a set of data that is sensitive to change.

In the case of a circuit breaker test, this might be the increased spread of trip times during a test. As the breaker fails, the response rate of the breakers is no longer as consistent as a new breaker. By analysing the spread of trip times, we can determine a pattern for a single piece of equipment and determine whether a breaker is likely to fail, down to a single piece of equipment, and feed that data into wider equipment patterns across an organisation.

As structured data can be analysed automatically, triggers can be used to perform background analysis (usually in the cloud) and provide reports of likely failure points to technical staff.

So, what advice do we give to people who think data collection is something you collect after the fact?

Don’t be the one who didn’t think of the second best time to plant a tree.