Data Science is a Team Sport! The Value of Creating Analytical Communities
By Kim Nilsson, CEO, Pivigo
If only I got an analytics insight every time I heard the phrase “we are going to hire one data scientist and see how it goes”… Some companies are just hell-bent on going down this route, presumably as they are uncertain about the potential outcomes of a data scientist’s work, and therefore hesitant about investing a greater amount. We will come back to what options exist for this dilemma later on, but for now, let us just establish one ground truth: hiring one data scientist does not work. Data science is a team sport.
Let us first think about how data science projects work. Each project will require a wide range of skills; programming and engineering, math and statistics, machine-learning, database querying and more. The likelihood of finding all these skills in one person is small, and the price to pay if you do is high. Individuals that cover all these skill sets are also likely to get frustrated soon, unless given sufficiently challenging problems to solve. A more cost effective and future proof strategy is to hire a team of individuals with complimentary skills, that feed off each other and deliver projects faster and more creatively.
Let us then think about where this team sits within an organization. A core data science team will typically sit within a larger analytics team of some sort, which will also include BI analysts, analytics managers, and the IT department. This team, in turn, is likely to be one business unit out of many within a company, further surrounded by employees across the business, compliance officers, and the strategic direction from the board and C-suite. Finally, let us not forget that a company is not an island; it sits in a wider environment consisting of other stakeholders such as customers, investors, and regulators.
Thinking of this onion of layers surrounding the data science team, it becomes obvious that there are very large network effects possible, given that the majority of the company and its resources are located outside of the data science team itself. If your data efforts can draw on these resources, and engage the rest of the company, progress will be more rapid, more cost effective, and importantly more in line with the actual needs of the business.
The solution is to create analytical communities. What is an analytical community, and how do you create one? There are many methods available to those who wish to create their own communities:
Identify analytical individuals.
There will be many employees in your business that either have an interest in analytics and/or certain levels of skill sets. It could be the marketing executive with an A level in Maths, or a supply chain expert who codes in her spare time. These individuals can be identified through a combination of searching through the HR database, and simply posting on the company websites asking for interest.
Train and upskill.
For those that have an interest, give them time and access to training resources to upskill. Not everyone needs to be professional coders, or machine-learning experts, getting everyone to understand basic logic and statistics, and how to use tools such as Alteryx, Dataiku, and Tableau will already make them valuable members of your analytics community. Great learning resources exist online, and companies such as Corndel, in combination with the foresighted Apprenticeship Levy in the UK, deliver excellent training for those wishing to upskill significantly.
Create digital meeting spaces.
Find ways for your analytics community to meet. Most likely, the best way will be online and there are many ways to organize large teams in online communities. Typical tools for this include Slack, Yammer, Microsoft Teams, etc. These types of forums allow individuals to communicate both one-on-one and in lose teams, and often allow sharing of information and documents in specific channels or project spaces.
Run internal events.
If it is possible to get your community to one location, running an internal event, such as a hackathon, is a great way to get everyone excited and motivated, to get them to know each other better, and to progress with projects. Some companies bring in external partners for these hackathons, e.g. a charity, others prefer to work on their own projects. A day or two is usually enough to get some work done, and get the team excited.
Encourage and facilitate cross-team projects.
Whether via an online community or internal events, get the community to work on temporary teams, delivering value to the business. Ideas can both come from the data science teams, but also from individuals within the community itself. This way you are drawing from resources you already have, as well as working with individuals who truly understand your business inside out. No extra cost and no need to explain the context.
Run external events.
Individuals who have great skills and who are interested in your business is not just limited to your employees. Every once in a while, consider also running external events, such as meetups or hackathons. This will help draw ideas, inspiration, and skills from a much wider community. It can help drive projects forward, and could also result in identifying individuals to hire.
Seed the internal community with external experts.
Bringing in external experts and consultants for a period of time can seed the internal team with ideas and expertise, and these people can mentor and help build the internal community. You may even decide to keep the contractors, if they work well enough with the new team, but even if not, they will help accelerate any internal efforts.
Ultimately, for these initiatives to work, commitment and blessing is required from the top levels of the business, as it will take some time and resources away from other parts of the business. But, the positive benefits are manifold. The company draws on internal skills at a very low cost, and it can scale analytical efforts quickly. It will bring interesting insights from the rest of the organization into the data science team. It can help motivate individuals in their day jobs, by allowing them to work on analytical side projects and allowing them to upskill. And importantly, the greatest challenge for any organization on their data science journey is to get buy-in from the whole organization for data science efforts, and to change the culture to embrace a data driven mindset.
Coming back to the initial question of proving the value of data, a final piece of advice is to always break down any larger initiative into smaller components and projects that can be delivered in an agile fashion. That reduces cost, reduces risk, and increases the likelihood of success from the ability to course correct when the scope starts to wander. With the right project set-up, and the right people working on the project, you will be harvesting data insights and value in no time!
Get in touch if you are interested to know more about Apprenticeship training via Corndel, or if you are interested in how Pivigo can support with external experts, event organization, and community building.
Pivigo s a part of the investment portfolio of Merian Ventures, a venture firm founded by Alexsis de Raadt St. James that invests in women-founded and co-founded tech innovation in the US and UK.