Project management, the uninitiated tend to assume, is project management. You draw up the plan, assess the risks, and then deliver. A project manager can shift from project to project without any difficulty. But that’s not quite true, especially in business analytics.
Managing business analytics projects is not quite like other project management jobs. While all projects have variations, business analytics projects tend to involve large amounts of uncertainty. The requirements change frequently and dramatically, as do the possible solutions. As a result, a study published in MIT Sloan Management Review a few years ago explained that good business analytics project managers tend to have very particular skills and attributes, not always shared with other project managers.
Attributes for success in business analytics
A focus on delivery, and a belief in ‘doing’ and experimenting – This might sound like good project management in general, but it is more subtle. Traditionally, project managers focused on getting the plan right, and then delivering it. Analytics project managers tend to plan, but recognize that the plan will change hugely. They are therefore focused on what works, and particularly on trying things out, in a process of ‘intelligent experimentation’.
An understanding of the importance of being able to use the project deliverables – Analytics project managers understand that a deliverable is only valuable when it is being used, and especially, being used to deliver value. They therefore favor prototypes that can be used immediately, with a replacement rolled out when new features are ready. The process is seen as iterative.
A commitment to engaging with users – Stakeholder management is one of the most important aspects of any project, but analytics project managers take it even further. If things cannot be over-communicated in an ordinary project, this is truer by several orders of magnitude in an analytics project. A model that is not understood will not be used. And one that is used wrongly will deliver poor results. Co-development with users is vital to ensure that the model adds value from the start.
An ability to ‘translate’ between IT and business, and make things happen – Many people in business have a horror of the IT department, and the time it takes to make things happen. Good analytics project managers are able to balance the demands of the business for speed and practicality against IT’s fear of projects going wrong, or technology being ineffective, and speak the languages of both business and technology. Their focus is on managing relationships across both areas.
Learning and developing skills
While these skills may currently be scarce, they can all be learned. Learning, however, requires a commitment, and access to resources. There are plenty of online resources available, and conferences offer both formal presentations to learn from the experience of others, and also more informal discussions at coffee time, with the bonus of face-to-face interaction. But attending a conference can seem like a big commitment, especially if the event lasts several days. It is also rare for one to turn up at just the right moment when we are grappling with a particular issue.
This may explain why meetups are becoming increasingly popular. They are an agile and dynamic form of engagement and learning.
Meetups, for those who have not yet encountered them, are social networking with a difference. Where much social networking is online, meetups, as the name suggests, start online, but with the aim of meeting in person. They are community-based, bringing together people with similar interests within a geographical area. They therefore facilitate local networking.
Meetup groups tend to meet regularly, and for fairly short sessions—an hour after work, perhaps, or half a day. They are usually designed to discuss a particular issue, such as a problem that is exercising several members of the group, with invited speakers and plenty of conversation. They can, therefore, be an ideal way to learn more about how other people are tackling a problem that you face.
For data scientists, this kind of opportunity may be like gold dust. We have all heard about the global shortage of data scientists, but for the individuals concerned the problem goes beyond the recruitment issues faced by companies. When you are the only data scientist in your company, it is hard to discuss detailed analytics problems with your colleagues. The neutral environment of a meet-up makes it easy for the participants to have peer-to-peer conversation at eye level. Also cross-industry best practices are valuable and helps to look beyond one’s own nose.
Meetups, with their focus on networking and engaging with others within your community, allow data scientists to overcome this issue, and build a network across a region or location. Sharing ideas may start at a meetup, but it certainly does not have to end there.
I am currently supporting the dataviz meet-up in Frankfurt a.M., Germany. We are hosting a full-day meeting in Heidelberg on the 6th of May, 2017. You can find our meeting calendar here.
