An Example Application Of 'Descriptive Statistics' And 'Inferential Statistics' Via MS Excel On The Job

Descriptive Statistics and Inferential Statistics are handy tools for management decision making at work. I realised and appreciated more the relevance of these two maths tools during my MBA studies in the unit 'Information And Decision Analysis’ as this is where I am applying many of the mathematics leveraging MS Excel. Descriptive Statistics describes what is happening within the data set while Inferential Statistics is using data to make predictions or forecasting. Indeed, both of these Statistics Maths are very handy 'maths tools' for management, especially when the situation demands that "in God we trust, everyone else must bring data".

The Situation

Changes in OTML Executive Leadership Team (ELT) happened in Q2 of 2023. By Q2 of 2023, production performance forecast for 2023 Budget was in dire strait with Cu production revised from 90+ kt to F2 of 84+ kt (and later F3 of 80+ kt). The Business Improvement function was elevated and reinvigorated by the end of Q2 and incorporated into the new Strategy Business Development & Optimsation Business Unit by mid Q3 2023. Work begun on securing the F2 Cu production target of 84+ kt with the project ‘Roadblock Mitigation Strategies and Outcomes’ being established. By mid-September 2023, there was less hope of even achieving the F2 Cu target and the new lower F3 Cu target of 80+ kt was set. In late September 2023, the new MD & CEO Kedi Ilimbit called all production area leaders to the Whitehouse (i.e. company HQ in Tabubil) to discuss the then production performance to date and the constraints and issues impacting: grades, throughput, maintenance and YTD shortfalls, which were the key outcomes of the ‘Roadblock Mitigation Strategies’ project. Since mid 2023 and that reinforcing master stroke meeting, the business started to see this turn around in our Cu production performance around mid-September 2023. A two weekly ELT update was agreed to be convened until the end of the year, however, after only five (5) meetings with two (2) meetings per week in the first three (3) weeks, this was cancelled after the ELT felt that production performance had improved and that the team appeared to have regained back control.

By 15th December 2023 we met and exceeded the F3 Production target of 80,962.73 dmt of Cu by reaching 81,194.86 dmt of Cu with 16 days to spare. Then on the eve of Christmas, another milestone achievement was reached with our business meeting and exceeding the 2023 F2 Production Target of 84,345 dmt of Cu by reaching 84,782 dmt of Cu with seven (7) days to spare. For such milestone achievements, we must document and know ‘What Good Looks Like’. This is following the ISO9001 (Quality Management Systems) basics: what we do, write it down; what we wrote down, follow it; if not followed, correct and continuously improve it.

Following the ISO9001 basics, I had collated our production data for the corresponding period from 1st September 2023 to 31st December 2023 to document ‘What Good Looks Like’.

Following our three (3) BI themes of: production improvement, process efficiency improvement and cost containment, this endeavour is to correlate our efforts in the BI function to identify opportunities for continuously improving our production process.

Every function that is executed in a business is a PROCESS. Every process has five (5) key elements and these are:

1. A process requires Understanding;

2. That process, when we understand it well enough, we will realise that it has Variables; and

3. These process variables must be properly Controlled.

4. Also from our understanding of the process, we should realise that the process has a Capacity; and

5. That process capacity requires Continuous Improvement.

For continuously improving our production process, we understand and know the key production variables that we must properly control to achieve the desired production capacity targets in the key result areas that we budgeted for.

At the end of 2023, with our business now having met and well exceeded the 2023 F2 Production Target of 84,345 dmt of Cu and ending up with 86,791 dmt of Cu, we had to capture the details of this win and know ‘What Good Looks Like’.

The Scope (Leveraging MS Excel)

A MS Excel Workbook was created to gather the daily 'Mine and Mill Productions' and the daily 'Ore Blend' data from 1/09/2023 to 31/12/2023. The strength of relationships between the Production's KRAs and Production's Key Variables that impact on the Production's KRAs can be established. This is so that these variables can be properly controlled with a high degree of confidence and certainty to influenced and achieved the desired KRAs.

The Actions Taken

A MS Excel Workbook named ‘What Good Looks Like_2023 F3 Production KRAs & Variables Data Analysis’ was created on 10/10/2023 to collate the production data.

Production data was then logged daily starting from 1/09/2023 to 31/12/2023. The data collection table was organised to show ‘Descriptive Statistics’ for each ‘Production’s Key Variable’ data set at the top of each data set column.

The Results

Correlation Coefficients (r) showing the strengths of relationship between the Production’s KRAs and Production’s Key Variables were established. Correlation Coefficients (r) is part of 'Inferential Statistics' to assist with making predictions and forecasts.

Another outcome was a summary visual performance dashboard highlighting at a glance which Production’s KRAs will need more effort invested in to improve.

Part of this visual performance dashboard includes the ability to drill down to see the strength of relationship between the Mill's Key Variables in Throughput and the Sub-Variables (like feed size, power draw and grind size) that impact or influence Mill Throughput. Scatter Graph was mostly leveraged in this instance.

Leveraging the ‘Mean’ and ‘Standard Deviation’ of ‘Descriptive Statistics’, the ‘Lower’ and ‘Upper’ control lines of the ‘Control Charts’ were determined.

During the data analysis it was realised that promising opportunities lay in the area of Ore Blending to improve production. A ‘What If’ scenario ‘Calculator’ for modelling the optimum ‘Ore Blend’ to deliver the Budget Production given the ‘Production Budget’ variables was developed. All the user need to do is type in the budget constraints value of: Mill Throughput, Flotation Concentrate Tonnes, Cu & Au Mill Feed Grades and Cu & Au Recoveries. The 'Ore Blend' to deliver the Budget Production will then be automatically returned to a one-pager summary dashboard.

Conclusion

A practical example of a STAR (situation target action result) story of how ‘Descriptive Statistics' and ‘Inferential Statistics' are applied leveraging MS Excel at work has been presented for a production process. A process requires understanding of its variables and these variables must be properly controlled for the process to achieve its desired inherent capacity. For this practical case, production picked up in mid-September 2023 and the business was able to met and exceeded its critical targets which had seemed impossible at the start of 2023. The business needs to understand what the variables in the production process look like and now we know what 'Good Looks Like' and this was through the application of 'Descriptive Statistics' and 'Inferential Statistics' and leveraging MS Excel. These insights will now help us to focus and control the specific necessary Key Production Variables of our production process to optimise our desired Production KRAs going forward.