- Megan Yates
Who You Gonna Call?
Or How to Decide When You Need a Data Scientist Everybody wants success. With revenues in many industries under pressure, more and more companies seeking to transform their fortunes, now contract or hire data scientists. The problem is, companies don’t always know how best to utilize data scientists let alone how to successfully integrate them into their business. In this blog we’ll explore some points to help you decide when you need a data scientist. 1. You Know What a Data Scientist Does - and What They Don’t Data scientists don’t arise from one discipline. The specialty was born out of a need to solve business problems using complex data. Most data scientists have a good mix of solid statistics and probability, experimental design, advanced modelling and computer science. But apart from these technical skills, they have a drive and curiosity to analyse deeply, solve problems and learn. This combination of skills translates into being able to collect large amounts of data, process and analyze them to solve business problems and drive tangible change through pro-active interventions. Common techniques used by data scientists include forecasting, segmentation, machine learning and natural language processing. Knowing a data scientist’s capabilities and understanding how they can be used to leverage your data is the first step in knowing if or when you may need one. You also need a good grasp of how data science skills differ from other data-related roles like analysts performing business intelligence (BI) and reporting roles. Typically, BI analysts look at historical trends while data scientists will focus on predictors and building algorithms to anticipate future events. Not understanding what data scientists are best at can mean giving data scientists projects and tasks that don’t fully utilize their potential - and ultimately don’t add value to your business. 2. You Have Data The power of data science is best realized with data - and preferably a lot of it. Put effort into data science when you have collected some data. Unfortunately data is often collected for reasons other than analyses, for example, compliance. Therefore, getting advice from a data scientist during the design phase of a product or service can be crucial. Basing a product or service on a strong data strategy, and guided by a data scientist who knows what data will be needed in the future, will go a long way to realise data opportunities and data-driven value for your product or service down the line. 3. You Have a Problem To Solve - and you can articulate it clearly Get a data scientist when you have something specific to solve and not because everyone else is doing it. A former executive of a mobile network operator told us recently how the company churned out dozens of new products in an attempt to keep up with the competition. Most of these products subsequently lost the company money. So start off by formulating the problem and take time to do this properly. This means you’ll have gone through the process of identifying something that will significantly impact your business. Ruthless focus on a relevant and important problem will encourage efficient use of resources and help derive business value more quickly. 4. You Want to Be Less Reactive If you feel that the business reports you receive showing historical trends and past performance aren’t helping you make decisions, or if your performance figures are sometimes a surprise, then it might be time to enlist the help of a data scientist. A data scientist can help you build models and algorithms that reveal the significance of various predictors in observed trends and also predict future events. Knowing what is likely to happen facilitates much smarter, strategic planning. Your business will gradually change from one where fire fighting is the norm, to one of calm, measured, proactive planning. 5. Aggregate Metrics Aren’t Giving You Value Imagine you attend a presentation on your customers where you learn that 53% of your customers are male, 20% are using Product X and 35% pay late. While this information is interesting, your quickly realise you have no idea what to do with it. Firstly you’re not alone, and secondly you could benefit from having a data scientist close by. A data scientist will help you understand trends within groups of customers who behave similarly and, most importantly, will allow you to predict particular behaviors, such as product usage and changes, churn, payment and response to messaging, and all this at a customer level. Finer detail in the approach allows you to anticipate an individual customer’s next moves and to develop tailored services and communications around this knowledge. 6. You are Willing to Experiment A business guided by data with data scientists at the helm will require running experiments. And the outcomes of these experiments won’t be known ahead of time. This can be a difficult concept for cost-conscious managers to embrace. As a business leader however, you’ll need to be prepared and amenable to trusting your data scientists and sanctioning experimentation. Feedback from the experimental outcomes - whether positive or negative - will contribute to a much better understanding of your customers and and drive incremental improvements to your business.