3 Main Pathways To Successful Mathematical Optimization Implementations
Author: Edward Rothberg, PhD
Enterprises across the business spectrum are accelerating their adoption of AI technologies, with 83% of companies reporting that they boosted their budgets for AI over the past year and 75% stating that they plan to embark on new AI projects in the next six months.
As the CEO of a mathematical optimization software firm, I’ve witnessed this recent surge in interest and investment in AI firsthand as an ever-increasing number of businesses are looking to launch mathematical optimization implementation projects.
The question is: How can these companies successfully steer their implementation projects from concept to completion?
Every mathematical optimization project starts with a vision of how this AI software technology could be used in an organization to address a complex business problem (such as supply chain planning or workforce management) and enable better decision-making and business outcomes.
But many companies aren’t sure exactly how to transform this vision into a reality. To do this, they must be able to figure out how to integrate mathematical optimization technologies with their existing IT systems, as well as their processes and people, and how to implement a mathematical optimization application in their organization.
Although every company’s journey to a mathematical optimization deployment is unique, there are certain well-trodden pathways (or, in other words, best practices) that you can follow to help you navigate the implementation landscape.
In this article, I will highlight the three main pathways to successful mathematical optimization implementations and discuss how you can determine which approach is right for you.
Understanding The Three Main Implementation Approaches
When you begin to think about investing in mathematical optimization (or any AI technology for that matter), a fundamental question you must ask yourself is, “Does my company want to custom-build an application from scratch or buy a packaged software product?”
Your answer to this critical “build or buy” question will help guide you in selecting the right pathway as you progress on your implementation journey. Generally speaking, there are three main implementation pathways that companies take:
1. An Off-The-Shelf Software Product
This implementation approach involves investing in a ready-to-use software product that has been developed for the mass market with mathematical optimization capabilities (and a commercial mathematical optimization solver like my company’s solution) embedded in it.
A wide range of vendors offer off-the-shelf products with built-in mathematical optimization functionality, from software giants like SAP (which incorporates mathematical optimization as a key component in many of its solutions) to smaller players like River Logic (which produces supply chain planning solutions powered by mathematical optimization).
If you want to use mathematical optimization to tackle a fairly standard business problem that can be commonly found in certain industries (such as financial services, aviation and manufacturing), then this off-the-shelf product approach may be right for you.
Off-the-shelf solutions are also suitable for companies with smaller budgets and shorter project timelines because they’re often less expensive than their custom-built counterparts and are easier and faster to integrate, install and use.
The chief drawback of off-the-shelf solutions is their lack of flexibility and limited customizability. If you’re looking for a bespoke mathematical optimization solution that can provide a perfect fit for your company’s business needs, you may want to opt for a different implementation approach.
2. A Consulting Partner
The second main implementation approach is to engage a consulting partner that has a track record of successfully developing and deploying mathematical optimization applications for companies across various industries.
These partners, who may be global consulting companies like Accenture, Boston Consulting Group or McKinsey or boutique firms like End-to-End Analytics or Princeton Consultants, will help you identify the key levers of costs and growth within your organization and will work with you to create a customized, scalable mathematical optimization application (with a commercial solver inserted into it) that meets your business requirements and enables you to capitalize on your business opportunities.
This implementation approach is usually more time-consuming and costly than simply buying an off-the-shelf product, but leveraging a partner’s know-how to build a solution tailored for you may be worth it in the long run because it can deliver immense, lasting ROI.
3. An In-House Team
A growing number of businesses today are assembling in-house teams of experts in analytics, operations research, data science and related fields who have the capability to build and deploy bespoke applications powered by mathematical optimization solver technologies.
If your company possesses (or is willing to invest in hiring) such a team, then you would be wise to entrust them with the task of developing your mathematical optimization applications because:
- They have a deep and unique understanding of your business problems and objectives and can design applications that fulfill all your requirements and fuel ongoing, optimal decision making and bottom-line growth.
- They can continually look for new opportunities to expand and enhance the use of mathematical optimization across the enterprise and thereby ingrain a culture of optimization in your organization.
Although the costs associated with hiring in-house resources can be substantial (especially for smaller firms with limited IT budgets), the potential benefits of having these experts within your organization are immeasurable.
Deciding Which Implementation Approach Is Right For You
There’s no right or wrong mathematical optimization implementation approach. It all depends on what’s best for your business. The decision of whether to build or buy and which implementation pathway to take should be based on many different factors, including the nature of your business problem, your project budget and timelines, and your level of in-house technical expertise.
It should be noted that many companies adopt a hybrid approach — for example, working with a consulting partner to implement an off-the-shelf solution or engaging external experts to collaborate with an in-house team.
Additionally, many companies find that their implementation approach evolves as they may start with an off-the-shelf solution and ultimately end up with their own in-house team designing custom-built applications.
It’s important to remember that whichever pathway you choose, there’s a whole community of mathematical optimization users out there that can help advise and guide you as you move forward on your implementation journey.
This article was originally published on Forbes.com here.