Is the hype surrounding the promise of Machine Learning merited or are we getting ahead of ourselves? Is artificial-intelligence no longer just for the digital-only business models of Amazon, Google, and Netflix? As a finance executive, what do you need to know to maximize the opportunity for both you and your business?
In this article, I will answer the questions our clients most frequently ask me. What does it take to get started? What role can I play?
First, Machine learning (ML) is based on algorithms that can learn from data without relying on rules-based programming. With the terrific growth in data volumes and computing power, we are now able to train algorithms to learn with very high degrees of success to predict outcomes. As a finance executive now is the time to learn and explore the potential for you and your business because the competitive significance of business models turbocharged by machine learning is poised to surge disrupting every enterprise, business model, and process.
What does it take to get started?
There are plenty of relatively straightforward use cases for machine learning in business. An obvious example is in sales and marketing to improve the targeting and acquisition of new clients. Another well-tested use is to predict customer churn. However, as a finance executive, you are positioned in an area of strategic influence and should think of ML just as you would, for example, using mergers and acquisitions as a strategy to drive growth. ML should be seen as a tool to craft and implement a strategic vision.
To fine-tune or even create a strategic vision, you will need subject matter experts on your team. Back to the M&A analogy, if you are embarking on a growth strategy fueled by M&A, you will hire M&A experts. ML is no different. To get started considering expanding your team to include "Quants," who are experts in the language and methods of ML as well as what I call "Digital Evangelists."
The Digital Evangelists play a crucial role in cross-discipline areas such as data analysis, ML, decision design and execution. Perhaps this is you? You need someone on your team who can take what the Quants have discovered, or lead them to reframe complex data patterns into actionable insights that your team or other managers across the enterprise can execute.
What about the data silly?
Yes. Access to volumes of useful and reliable data is required for effective machine learning. So a successful data strategy starts with identifying gaps in the data and then determining the cost-benefit of filling those data gaps and breaking down data silos. Again, this is where finance can lead because creating trusted financial data as a single source of the truth will enable thoughtful and uncompromising assessments of the evaluation of new ideas and strategic ventures.
All too often I've seen departments hoard information to politicize access to it, which is destructive but somehow allowed. As a finance executive, you can be the voice of reason that challenges this status quo and presents a vision where data will be democratized across your business for the benefit of all ML initiatives.
How can finance leaders really play a role in ML?
This is another question I've been asked many times. Like some many changes brought about by digitization of businesses, the finance leaders role will be to influence, encourage and enable it. Of course, ML should not be given a blank check. But, finance leaders perhaps need to learn from their marketing counterparts about A/B testing and follow the approach of controlled experimentation to deliver critical learning points for the future.
Finance leaders usually thrive when control and predictability are present when encouraging an ML strategy you can take the same approach. At Qubix we recommend a product roadmap approach to ML where each evolution of the plan builds on the previous and enriches the next iteration. We would recommend three phases the follow the potential of ML to your business.
Three Phases for ML
First, getting the data in order is a foundation, we refer to this as the 'platform' stage or the 1.0 release of ML. Getting the platform right leads to the high potential next phase of 'prediction,' think of this as ML 2.0 which can have a profound impact on the underlying performance of the business. Then you can progress to ML 3.0, the 'prescriptive' stage where man and machine work together to take pre-emptive decisions or actions that will have the potential to have a remarkable impact.
Switching back to ML 1.0 a frequent concern that I have seen often paralyze finance executives is their belief that data has to be perfect. In our experience, most companies have sufficient information to obtain new insights even from incomplete, messy data sets. As long as the optimal ML algorithms are used, the quality of data doesn't have to be perfect. Further, in our experience, adding new data sources can often be of only marginal benefit compared with what can be discovered from existing data sources.
In future posts, I'll explore each of the phases of ML, from 1.0 through to 3.0. We will look at the benefits of being able to answer the question, "what will happen?" rather "why did that happen?" Furthermore, we will connect the role of finance, and the data sources in finance to the overall strategic path ML can follow for your business.
As a final thought no matter what previously unseen insights ML and advanced analytics uncovers, only human managers can decide the essential questions, such as which critical business problems a company should be trying to solve.
The winners will be neither machines alone, nor humans alone, but the two working together effectively.