Introducing data science in your company: A peek behind the scenes

Data in every nook and cranny in every company in the world. Being able to harness this data can bring huge potential and value to companies. We can firmly say that big data and data science influence the success and growth of a business. The transformational power of technology has numerous use cases across different industries. 

With the help of various tools, algorithms and machine learning principles, data scientists discover hidden patterns in the raw data. But the secret sauce is how data is used to derive meaningful insights and take data-driven actions. 

A shift towards Data Science in the company

As with other state-of-the-art technological advancements, introducing data science in a company is challenging. Young startups may be more open and adaptable, considering that the shift is at a smaller scale compared to traditional enterprises. However, disregarding the company size, a change in the culture is imminent when trying to utilize data to reach the full potential.

Introducing data science in a company is like being dropped in a forest, and it’s night time, and you only have a pocket knife with you, and the forest is on fire.

So, what do you actually need to do to successfully introduce data science?

When you are about to introduce something revolutionary in your organisation, it’s probably smart to talk to people that have already done it. For that purpose, we asked Adrian Badi, a former Senior Data Analyst at William Demant, current Business Finance Manager at Jupiter Back to share his experience with introducing data science in a big company. 

“Introducing data science in a company is like being dropped in a forest, and it’s night time, and you only have a pocket knife with you, and the forest is on fire”, starts off Adrian in his presentation at Data Innovation Summit 2019. The image may be exaggerated, but the feeling of diving into data science is certainly not.  

To make it easier for businesses that are yet to start with data science, Adrian Badi suggests a list of “ingredients” that you will need in your data science project:

• Business case – data science is not magic that can fix all the company’s problems. You need to select a concrete case you want to solve by applying data science. The problem has to be clear, concise and measurable. There are companies that are too vague in presenting their problem that it’s a real pain for data scientists to translate them into code.

• Data – data is the foundation and material upon which all data analysis is based. Where do you find this data? Every place you store any sort of information – CRM systems, transactional data, meeting appointments record – as much and as quality data as possible that can be captured, analyzed it and have models built based on it.

• The Mastermind principle – the term presents a group of most successful and brilliant people in the business. A term introduced by Napoleon Hill in 1925, them referred to a gathering of successful people where they discuss various problems. Your goal is to bring together a mastermind team, each member with experience in their particular field and relevant to the business case. The idea is that people working in a team that resonates and work as one mind will solve the problem faster than trying to solve it individually in separate silos.

• Success criteria – or KPIs in business terminology. As with every project, you need to establish success criteria in order to know whether your data science project was fruitful. Are you going to measure an increase in sales, less production time, fewer errors in products, less average time to close a sale, etc?

• Testing the MVP – After gathering all the data, you provide a product, a dashboard, a tool that people from sales, marketing, finance, manufacturing, etc, can use in order to implement the insights in order to improve their processes. The feedback they are going to receive from the end-users will help you discover user pain-points and improve the data science product.

• Plating – While the employees are testing the data science product, it’s crucial to plate your findings. When you are in the exploratory phase where you learn the data and try to find new insights, you want to keep the stakeholders in the loop. You cannot just disappear and come back with answers. This is especially important if you are working with people that are encountering data science for the first time. It helps if you just present whatever meaningful insights you uncover during this process that can prompt informed decisions.

• Keeping the momentum alive –  Be consistent in informing your stakeholders about all new insights you discover along the way. Promote and market every new finding so that all employees within the company. All employees should be familiar with the data science project, so they know what is going on and how it can help them get the most out of what they’re doing.

Challenges to introducing data science

The data science process comes with certain challenges. So it’s best to be prepared and know what awaits you on your path towards a data-driven organisation. Here are some of the challenges Adrian has faced with his projects:

1. Big companies have an established way of functioning and strict day-to-day operations. You will also meet people that don’t know what data science is. So you have to be prepared to work with what you have. Be flexible with your time.
2. You are going to have to allocate time to the mastermind principle to discuss the project progress. Be prepared to put everything else on hold that day in order to be concentrated on the discussion. And you’ll be surprised to find out it’s actually really effective when you’re not distracted by emails and phone calls.
3. When you are doing your proof of concent, you may come across some resistance and scepticism from people, especially if you are working with traditional organisations that are not used to data science. So be prepared to be challenged and defend your data. Remember that you should present the value of data science in a structured manner.
4. Deployment is a challenge by itself. Putting things into production is an issue for some companies. You need to find a way to automate the solution so you don’t have to manually execute the whole process time and time again.
5. People often associate data science and automation with the fear that their jobs will be replaced with an AI robot. Many processes can be automated, but data science is a field about people, and people will never get out of business, asserts Adrian. “You need actual people that understand the business and the processes – and that’s a skill that will never be [replaced],” he reiterates.
6. Companies believe that if they hire the best data scientists, there is no fear of failure. But, according to Adrian, the biggest commodity in the future will be curiosity, not data science. Data science can be learnt, but having the curiosity and the willingness to go out there and ask the questions, being energetic and enthusiastic is the right state of mind that organisations should be after.

What counts in the end

And at the end, challenges are and there always will be, but one secret ingredient helps you push through them all – excitement. “Just like negative feelings and criticism, excitement is a contagious feeling that spreads like wildfire”, explains Adrian ardently. When your team and mastermind are excited and enthusiastic about the insights you’ve found, you spread that energy to the people you are presenting to. And you help them understand that data science isn’t just numbers and statistics, it can add value and make people’s lives easier. It’s noble to help someone do their job better and ease their life. That’s what matters and data science makes it possible.