Prescriptive Analytics

Data is the fuel, Analytics is the driving force, but the prescriptive practice is the one that can generate the scenarios for an optimized business decision and to amplify human intelligence.

Prescriptive analytics recommends the best decision based on one or a few pre-defined objectives. Once the business is mathematically modeled, it is possible to establish which path to take towards the desired business goals, such as:

  • Improve net margins
  • Higher asset utilization
  • Better ROI
  • Lower operational costs

The main steps

It is all about the business decision that will need to be made; it might be about sales, fulfillment, costs, margins, people or even a combination of those. The way this challenge is translated to an analytical equation is the key success factor for a great response from the model, but the restrictions it needs to comply with and the data that will be input are just as critical.

DHAUZ employs a multidisciplinary team, which includes business professionals, data engineers and optimization specialists, to review our client’s decision-making process and incorporate and optimized scenario analysis and action recommendations in order to improve their P&L.

1

2

3

4

5

6

Understand the business challenge and available data

Set objectives and identify restrictions for the analytical design

Develop mathematical models and optimization algorithms

Test, final adjustments and clients validation

Solution implementation, hyper-care and KPI and benefits tracking

User support and continuous improvement

1

Understand the business challenge and available data

2

Set objectives and identify restrictions for the analytical design

3

Develop mathematical models and optimization algorithms

4

Test, final adjustments and clients validation

5

Solution implementation, hyper-care and KPI and benefits tracking

6

User support and continuous improvement

The most common questions

What is the best method to solve an optimization problem?

Different optimization problems require different optimization models, actually, the same problem, in a different business context or decision making process, will be best solved by another unique method. What matters most is to be able to identify, for each specific case, the best suitable algorithm and method that will provide the best possible solution, consistently and within the adequate time frame for the decision process.

Is it worthy to invest in developing optimization solutions even if it’s meant to deal with very complex problems?

Without a doubt - yes! The proper algorithm will always recommend a better solution than what humans could put together without the support of a mathematical model. The people involved with the decision have an essential role, which is to validate and implement such recommendations; they can also refine the proposed scenario to accommodate lessons learned from experience and end up implementing the resulting solution.

Can the implemented models evolve and adapt to process changes?

It is essential that the optimization models are reviewed and adapt to changes in processes and business context. To enable this, it is important that, from the start, we understand the business challenges and keep constant contact with the end users. Whenever a need for change is identified, we can make adjustments to the model and algorithms to guarantee that the results will continue to be accurate and relevant.