Machine Learning

Sustainable Enterprise Technique with Knowledge Analytics | by Samir Saci | Jan, 2025

January 10, 2025
roosho

In one other article, I introduce the mannequin we’ll use for instance the complexity of this train with two eventualities:

  • Situation 1: your finance director needs to reduce the general prices
  • Situation 2: sustainability groups push to reduce CO2 emissions

Mannequin outputs will embody monetary and operational indicators for instance eventualities’ affect on KPIs adopted by every division.

Diagram illustrating cost and environmental impact distribution along the supply chain. Costs of goods sold link to retail, production costs link to manufacturing, and logistics costs link to freight and delivery markets. Environmental impacts include production and logistics footprints, managed by the sustainability department.
A number of KPIs involving a number of departments — (Picture by Samir Saci)
  • Manufacturing: CO2 emissions, useful resource utilization and price per unit
  • Logistics: freight prices and emissions
  • Retail / Merchandising: Value of Items Offered (COGS)

As we’ll see within the completely different eventualities, every situation might be beneficial for some departments and detrimental for others.

Do you think about a logistic director, pressured to ship on time at a minimal value, accepting the disruption of her distribution chain for a random sustainable initiative?

Knowledge (could) assist us to discover a consensus.

Situation 1: Decrease Prices of Items Offered

I suggest to repair the baseline with a situation that minimizes the Value of Items Offered (COGS).

The mannequin discovered the optimum set of vegetation to reduce this metric by opening 4 factories.

Icons representing manufacturing plants of various sizes and capacities, ranging from small factories to large industrial facilities. Each icon highlights capacity differences and potential production output.
Manufacturing community for Situation 1 — (Picture by Samir Saci)
  • Two factories in India (high and low) will provide 100% of the native demand and use the remaining capability for German, USA and Japanese markets.
  • A single high-capacity plant in Japan devoted to assembly (partially) the native demand.
  • A high-capacity manufacturing facility in Brazil for its market and export to the USA.
Sankey diagram showing supply chain flows from production locations to markets. Japan, India, and Brazil production supply units to markets in Japan, the USA, Germany, Brazil, and India, with flows varying in size to represent volume distribution per market.
Resolution 1 to reduce prices — (Picture by Samir Saci)
  • Native Manufacturing: 10,850 Models/Month
  • Export Manufacturing: 30,900 Models/Month

With this export-oriented footprint, we now have a complete value of 5.68 M€/month, together with manufacturing and transportation.

Stacked bar chart showing the costs of goods sold (COGS) analysis by production location. The chart includes fixed costs (blue) and variable costs (red). The total cost is broken down into Japan (2.07 M€/month), Brazil (1.42 M€/month), and India (1.52 M€/month), with the highest total at 5.68 M€/month
Whole Prices Breakdown — (Picture by Samir Saci)

The excellent news is that the mannequin allocation is perfect; all factories are used at most capability.

What concerning the Prices of Items Offered (COGS)?

Stacked bar chart showing COGS breakdown by market, highlighting transportation (green), production (red), and fixed costs (blue). Japan has the highest COGS at 4.12 €/unit, followed by Germany and the USA, while Brazil and India have the lowest at 80 and 50 €/unit respectively.
COGS Breakdown for Situation 1 — (Picture by Samir Saci)

Apart from the Brazilian market, the prices of products bought are roughly consistent with the native buying energy.

A step additional can be to extend India’s manufacturing capability or scale back Brazil’s manufacturing facility prices.

From a value viewpoint, it appears good. However is it a very good deal for the sustainability group?

The sustainability division is elevating the alert as CO2 emissions are exploding.

We have now 5,882 (Tons CO2eq) of emissions for 48,950 Models produced.

Bar chart displaying CO2 emissions by market location and source. The USA market has the highest total emissions (4,980 tons CO2eq), with transportation contributing 3,870 tons and production 1,110 tons. Emissions for Brazil, Germany, India, and Japan are significantly lower, with Brazil at 55 tons CO2eq
Emissions per Market — (Picture by Samir Saci)

Most of those emissions are as a result of transportation from factories to the US market.

The highest administration is pushing to suggest a community transformation to scale back emissions by 30%.

What can be the affect on manufacturing, logistics and retail operations?

Situation 2: Localization of Manufacturing

We swap the mannequin’s goal operate to reduce CO2 emissions.

Icons illustrating a variety of manufacturing site configurations, representing low-capacity and high-capacity factories. The image compares different plant types based on their environmental and operational characteristics.
Manufacturing community for Situation 2 — (Picture by Samir Saci)

As transportation is the foremost driver of CO2 emissions, the mannequin proposes to open seven factories to maximize native fulfilment.

A Sankey diagram depicting production and market flows for different locations. The USA, Germany, Japan, Brazil, and India are shown as production points linked to their respective or export markets with varying unit volumes represented by flow widths.
Provide Chain Flows for Situation 2 — (Picture by Samir Saci)
  • Two low-capacity factories in India and Brazil fulfil their respective native markets solely.
  • A single high-capacity manufacturing facility in Germany is used for the native market and exports to the USA.
  • We have now two pairs of low and high-capacity vegetation in Japan and the USA devoted to native markets.

From the manufacturing division’s viewpoint, this setup is much from optimum.

We have now 4 low-capacity vegetation in India and Brazil which might be used means beneath their capability.

A bar chart comparing variable and fixed costs by production location (USA, Germany, Japan, Brazil, and India). The total cost is prominently displayed, highlighting how fixed and variable costs contribute to overall production costs.
Prices Evaluation — (Picture by Samir Saci)

Subsequently, mounted prices have greater than doubled, leading to a complete price range of 8.7 M€/month (versus 5.68 M€/month for Situation 1).

Have we reached our goal of Emissions Reductions?

Emissions have dropped from 5,882 (Tons CO2eq) to 2,136 (Tons CO2eq), reaching the goal mounted by the sustainability group.

A bar chart showing CO2 emissions in tons by market (Brazil, Germany, India, Japan, and the USA) with sources split into production and transportation emissions. The USA has the highest combined emissions, with transportation dominating.
Emissions per Market (Situation 2) — (Picture by Samir Saci)

Nevertheless, your CFO and the merchandising group are fearful concerning the elevated value of bought items.

A stacked bar chart showing the breakdown of the cost of goods sold by market (USA, Germany, Japan, Brazil, and India) into production, transportation, and fixed costs. India and Brazil have the highest COGS due to high fixed and production costs.
New COGS for Situation 2 — (Picture by Samir Saci)

As a result of output volumes don’t take in the mounted prices of their factories, Brazil and India now have the best COGS, going as much as 290.47 €/unit.

Nevertheless, they continue to be the markets with the bottom buying energy.

Merchandising Group: “As we can not enhance costs there, we is not going to be worthwhile in Brazil and India.”

We’re not but performed. We didn’t contemplate the opposite environmental indicators.

The sustainability group would love additionally to scale back water utilization.

Situation 3: Decrease Water Utilization

With the earlier setup, we reached a mean consumption of 2,683 kL of Water per unit produced.

To satisfy the regulation in 2030, there’s a push to scale back it beneath 2650 kL/Unit.

Two charts: on the left, a donut chart displaying water usage distribution by country, with Japan leading at 38.8% and the USA at 33.5%. On the right, a bar chart showing water usage per production location, with India using the highest at 3,500 liters per unit.
Water Utilization for Situation 2 vs. Unit Consumption — (Picture by Samir Saci)

This may be performed by shifting manufacturing to the USA, Germany and Japan whereas closing factories in Brazil and India.

Allow us to see what the mannequin proposed.

Icons of three types of factories: a small factory, a medium-sized factory, and a large factory with chimneys, representing various production capacities.
Manufacturing community for Situation 3 — (Picture by Samir Saci)

It seems just like the mirrored model of Situation 1, with a majority of 35,950 items exported and solely 13,000 items regionally produced.

A Sankey diagram showing production flows from countries (e.g., Germany, USA, Japan, Brazil, and India) to respective markets, with unit quantities labeled for each flow, highlighting production-to-market supply chains.
Move chart for the Situation 3 — (Picture by Samir Saci)

However now, manufacturing is pushed by 5 factories in “costly” nations

  • Two factories within the USA ship regionally and in Japan.
  • We have now two extra vegetation in Germany solely to provide the USA market.
  • A single high-capacity plant in Japan will likely be opened to satisfy the remaining native demand and ship to small markets (India, Brazil, and Germany).

Finance Division: “It’s the least financially optimum setup you proposed.”

A stacked bar chart showing the costs of goods sold (COGS) analysis by production location. Includes variable costs in red and fixed costs in blue, with total costs highest in the USA at 2.4M€/month.
Prices Evaluation for Situation 3 — (Picture by Samir Saci)

From a value perspective, that is the worst-case situation, as manufacturing and transportation prices are exploding.

This ends in a price range of 8.89 M€/month (versus 5.68 M€/month for Situation 1).

Merchandising Group: “Models bought in Brazil and India have now extra affordable COGS.”

A grouped bar chart illustrating COGS per unit across markets, broken into transportation, production, and fixed costs. Brazil and India show the highest COGS due to higher transportation and production expenses.
New COGS for Situation 2 — (Picture by Samir Saci)

From a retail viewpoint, issues are higher than in Situation 2 because the Brazil and India markets now have COGS consistent with the native buying energy.

Nevertheless, the logistics group is challenged as we now have nearly all of volumes for export markets.

Sustainability Group: “What about water utilization and CO2 emissions?”

Water utilization is now 2,632 kL/Unit, beneath our goal of two,650 kL.

Nevertheless, CO2 emissions exploded.

A bar chart showing CO2 emissions by market and source, separating transportation (green) and production (blue). The USA leads in emissions at 2,500 tons, mainly from transportation.
Emissions per Market (Situation 3) — (Picture by Samir Saci)

We got here again to the Situation 1 state of affairs with 4,742 (Tons CO2eq) of emissions (versus 2,136 (Tons CO2eq) for Situation 2).

We will assume that this situation is satisfying for no events.

The issue of discovering a consensus

As we noticed on this easy instance, we (as information analytics consultants) can not present the proper resolution that meets each celebration’s wants.

Three world maps illustrating sustainability scenarios for supply chain networks. Each map represents different setups for factory locations, logistics routes, and corresponding environmental impacts.
Situations and impacts on groups — (Picture by Samir Saci)

Every situation improves a selected metric to the detriment of different indicators.

CEO: “Sustainability just isn’t a alternative, it’s our precedence to change into extra sustainable.”

Nevertheless, these data-driven insights will feed superior discussions to discover a last consensus and transfer to the implementation.

A diagram with a sustainability team, analytics models powered by Python, and three supply chain maps showing factory locations, logistics routes, and impacts, demonstrating an integrated decision-making process.
Knowledge Pushed Resolution Design — (Picture by Samir Saci)

On this spirit, I developed this instrument to handle the complexity of firm administration and conflicting pursuits between stakeholders.