A sales approach to data science
Key to the success of any project, is getting stakeholder buy-in. It does not matter if you are an external consultant or belong to an internal business unit. It helps to get everyone excited about the project, if it is to bear fruit. This applies to data science projects as well.
With any new project, the data scientist is doing one of two things: she is either trying to improve upon an existing process or she may be introducing something new to the business. In both situations, the data scientist is basically selling the stakeholder an analytical solution to their business problem. She is trying to get them to trust that her solution will not only meet their needs, but that she is the right person to deliver it.
To do this effectively, a data scientist must employ the skills and tactics of a good salesperson. Salespeople have great empathy. They have an intimate understanding of what the customer wants and why they want it. Data scientists that treat their stakeholders in a similar fashion, get to know their clients’ business inside out and therefore understand their problems much better. Having this knowledge allows the data scientist to offer solutions that make a real impact to the business.
Just like a salesperson will speak to each customer differently even about the same product, a data scientist needs to be able to profile her stakeholders and speak to them in the language they understand, touching on points that speak uniquely to them and to their challenges. For example, when tasked with a sales forecasting problem, it would be very difficult to get both sales and the inventory department excited about the project by employing identical messages. These business units work closely together but require product forecasts for very different reasons.
This skill does not come easily to many data scientists who can get consumed by the intricacies of a solution to a problem than by the problem itself. It is worth remembering that data science clients care little about the exciting new features in some revolutionary piece of code. They are only concerned - and rightly so - about achieving a more productive work pipeline.
For the data scientist, this is more about developing a shift in attitude and perspective. It helps to spend as much time as possible with the client to really understand what their business is all about.