With the amount of data we have available on projects, why do teams tasked with workplace improvement struggle to use this data to identify problems and deliver change? When I started in lean business improvement many years ago, data-driven improvement was the method of choice. Driven by Six Sigma thinking, we were trained in data collection, analysis and advanced statistics before being let loose in a mining operation to facilitate improvements. With great enthusiasm, we assembled our teams, got out and observed processes, and then prepared for the data to do its magic. More often than not, our first sign of trouble appeared because the data we needed was often not available, and project sponsors were impatient for results and reluctant to give time or resources to gain the missing data. In the end, we used the data we had as best we could, relied on our relationships and newly-taught facilitation skills, and developed and implemented the best solutions available with the project team. We were often left wondering how much more we may have delivered if only we had better data.

Data-driven improvement is still as relevant today but, since then, there has been an explosion in the amount of data available. Technology now allows us to measure performance in real time and the Internet of Things promises even more data through increased connectivity. Remote operating centres and autonomous vehicles can record gigabytes of data by the second. Could all of this new data be the missing piece to implementing better improvement compared to what we used to be able to achieve? Some of our recent projects with various clients suggest we’re still not at our data-driven business improvement nirvana!

Data rarely delivers lean business improvement on its own

We’ve recently worked for various clients in areas such as mining operations, processing facilities and maintenance workshops. Each project involved using vast quantities of data. Although our clients had access to in-house teams of technical specialists and data scientists who could slice, dice and analyse the data, they were still not achieving the results they were expecting when it came to identifying and implementing improvements.

Through our early project observations, we recognised a similar pattern: after the general objective of ‘improvement’ was identified, gigabytes of data were handed to technical specialists or data scientists to ‘see what they might come up with’. Occasionally, a gem would be uncovered. Most of the time, this wasn’t found. People felt the approach led to analysis paralysis and frustration soon kicked in, followed by team disillusion – a death warrant for any improvement project.

Five considerations for data-driven improvement

Here are five core considerations we’ve identified across numerous projects to help teams effectively use increased data availability for improvement:

  1. A clear improvement method is still needed– most confusion occurs when translating operational language into useful information for data scientists, and translating data analysis back into operational contexts. A disciplined improvement methodology, with a clear objective, helps to avoid confusion and provide shared understanding regarding the improvement being sought.
  2. Analysis should first be informed by process– operations team members tasked with understanding and documenting a process (ie the subject of the improvement) need to be disciplined and thorough, avoiding the temptation to rush through it. Documenting the process right sets up the data analysis team for success and minimises or eliminates rework (and frustration) later in the project.
  3. Field-based insights are still valuable– for data scientists tasked with analysis, it’s essential they take time to understand the documented process. If possible, they should visit the work area and ensure they have a clear understanding of the questions that need to be answered through the data analysis phase. Change and improvement cannot be delivered from an office or a spreadsheet alone.
  4. Team diversity remains critical– the way teams deliver improvement hasn’t changed, even with the increased amount and quality of data now available. Teams still need to comprise diverse skills so they can balance detailed statistical analysis with field-based process familiarity to ensure findings can be practically applied in the workplace.
  5. Cross-team interaction isn’t one-off– during data analysis, interaction between operations and subject matter experts, and data analysis teams, needs to happen regularly. In some contexts, it can help to embed someone with operations experience in data analysis teams so they can practically translate information between project stakeholders in a faster and more iterative fashion.

This list is not exhaustive, but it’s a start for teams that are grappling with improvement objectives in light of increased data availability. In theory, more data should support better, faster improvement. However, it won’t get there on its own.

For the wider mining business improvement community, we believe there’s a critical role to be provided in helping projects balance increased data availability with the fundamentals of team engagement and ownership to implement sustainable change. Get this balance right, and we may start to see more examples of effectively using increased data to help teams implement greater improvements and deliver better performance. For more on how our team can support lean business improvement results, please browse our services.