“When farmers decide to use different fertilizers, Toros Tarim has to adapt to the new demand, even though their factories are operating at full capacity,” says Alper Şener, Data Analytics Head at IBM Business Partner Selco Consulting. “At that point, it becomes quite hard to decide which customer demands to satisfy and which demands to forego.”

For years, Toros Tarim planners used spreadsheets to develop optimal solutions for fertilizer production, based on the price of raw materials, customer demand and the production capabilities of the company’s three manufacturing plants. But this manual process was time-intensive and planners didn’t always have all the data required to see the overall situation.

“Planners could spend a week developing an optimal solution for the next month, but when they took it upstairs, management would say, ‘That looks good, but we need another scenario,’” says Şener. “Each time a new scenario was requested, it added another week to the schedule, and it could be three or four weeks before managers could make a decision. Meanwhile, market conditions kept changing.”

Further complicating the process was the fact that planners were not working with a complete set of data. “Each plant had their own production figures and product definitions and sometimes plant managers would use different subjective decision processes,” says Şener. “By using different data sets in different functional silos, it wasn’t possible to develop fully optimized scenarios. There was a huge opportunity for improvement.”

Based on its extensive experience with supply chain management, business analytics and information technology, IBM Business Partner Selco Consulting was chosen by Toros Tarim to digitize the company’s planning processes. “Selco consulted with us on software development and created a mathematical optimization model to find the best way to maximize profit, taking into consideration all company constraints,” says Hakan Göral, Chairman of the Board of Directors at Toros Tarim. “It took about a year and a half to build the model.”

Building a detailed math model was an enormous undertaking. Initially, it required extensive research to understand the many components of the supply chain that fed Toros Tarim’s three manufacturing plants. “The final model we built was comprised of seven million variables operating over one million constraints such as capacity usage, costs and fertilizer recipe alternatives,” says Şener.

After mapping out all the data required to run the model, Selco and Toros Tarim chose to run the model on the IBM® ILOG® CPLEX® Optimization Studio platform. “We use Microsoft Excel files to create input data for the simulation model and it takes about half an hour to update the data,” says Göral. “And when the data is ready, it takes around 10 to 25 minutes to run the model on IBM ILOG CPLEX software.”

Creating a comprehensive model and using the IBM software as a solution for all of Toros Tarim’s plants was a significant step forward. “One of the success points was having the company agree on a single database in a single data system,” says Şener. Everyone sacrificed their buffer zones and started talking about actual numbers, actual capacities and shared definitions.”

By using the IBM ILOG CPLEX Optimization Studio solution instead of manually preparing optimization scenarios on spreadsheets, Toros Tarim decision-makers no longer had to wait for days or weeks for the what-if scenarios they needed. “With a long process time, we were not able to run alternative scenarios while changing the constraints,” says Göral. “The shorter the process duration, the more alternative scenarios we can compare. Now we are able to find the best possible solution — or alternate solutions — to maximize company profit.”