Doing more with less: productivity and data analytics
Productivity in the Australian mining industry has fallen significantly in the last decade or so. Much of this has been structural – with the lag between capital investment costs and production revenue causing a drag on productivity, along with an increase in labour costs. But the development of some higher cost mines based on favourable commodity prices has also been a factor.
Now that commodity prices have fallen from their historic highs it is more important than ever before to find new and novel ways to improve returns on capital investment.
This may require complex and multifaceted adjustments for the industry, but the simple answer is that productivity gains will be essential for future profit growth if lower commodity prices become the ‘new normal’.
There has been a similar fall for the energy and water supply sectors and, although the reasons for these are more complex because of regulation and climatic variability, these industries also now need to improve productivity.
The Power of Data Analytics
Using advanced data analytics can identify and help drive improvements in productivity. Here are three examples.
Mining supply chains can present both major constraints and opportunities for productivity improvement in the commodity value chain.
While per tonne costs to transport commodities from source to market may be up to 40% of the total to-market cost, the impact of inefficiencies or disruptions in the supply chain can also feed back to mine production, making supply chain optimisation a topic of intense interest in the current economic climate.
Thankfully, complex commodity supply chains often offer significant opportunities for efficiency gain due to the sheer volume of interactions that must occur to transport commodities across the global market.
Each of these interactions relies on two or more participants to synchronise operations seamlessly on a continuous basis, but any delay or complication can affect all other steps in the supply chain. So any changes to supply chain operations must be assessed on the basis of their aggregated impact.
As an example, a large rail haulage business wanted to better understand the intricacies of commodity freight operations and identify opportunities for organic productivity improvement using existing assets and infrastructure. We combined several databases of operational train running, maintenance, customer order, network operation, HR and financial data to identify commonalities among periods of low productivity (measured in throughput rate).
The analysis indicated several behavioural tendencies by supply chain participants that consistently constrained the system as well as the key event types and geographical locations driving congestion.
Most interestingly, it demonstrated that the multi-user commodity supply chain was not mutually exclusive, and that all participants should cooperate to achieve the best individual outcomes.
A large electricity distributor suspected that some of its zone substation peak capacity forecasts were too high because of poor data quality produced by their SCADA (Supervisory Control And Data Acquisition) systems. They were having difficulty understanding both the trend in load growth and the drivers behind that growth.
We were able to build processes that cleansed poor data from the SCADA record, which in turn enabled a recalculation of the forecast.
We were able to produce a forecast built from a range on input data sources which was more accurate than our client’s existing forecast. This approach meant that we could describe and quantify the range of factors driving growth in peak demand. Our forecast was lower than our client’s original forecast, and this provided them with the evidence they needed to show that a $800 million network upgrade could be postponed for up to three years.
As a mining company moved from exploration to asset construction and a ramp up in production investment, decisions needed to be made in the amount of office space required for back office functions.
Leasing too much back office space would be an unnecessary cost, while leasing too little could lead to costly leasing arrangements in the future and loss of productivity as a result of operations being split across multiple sites.
The main issue here was that proper data records had not been kept on the historical workforce growth.
We utilised novel data sources (such as building swipe card data) and matched these back to HR records to show that current back office headcount numbers had been underestimated. This allowed us to produce a more robust headcount forecast, and our client was able to make better decisions on floor space planning which will ultimately lead to a more productive back office.
The right data in the right condition
The two most common hurdles to using data-driven solutions to productivity problems are data availability and ‘cleanliness’, but there is an increasing array of data sources available for analysis from ‘smart grid’ technology, telemetrics, mobile technology and publicly available sources such as meteorological and census data.
There are also many ways in which data can be cleansed using both rule-based and statistical processes and, increasingly, bespoke data cleansing algorithms.
Big Data Future
Whilst advanced data analytics can appear daunting for organisations just starting their journey towards analytics maturity, the pressure to demonstrate return on capital investment is an excellent opportunity to find quick wins through advanced analytics whilst also preparing your organisation for the era of big data.
Director | Forensic
+61 7 3308 7595