At the heart of it, the purpose of analysis is to drive better actions through informing better decision making. In this article, I will focus on one of the most important parts of this process; making sure the output lands well and drives action.
When is the right time to think about how an analysis will land? At the very start! Start every analysis by asking two questions: ‘so what?’ and ‘who cares?’. If it is not clear what decision you are trying to inform and who is going to make that decision, start by getting clarity on those points. Understanding the business problem and the audience for the analysis will radically increase your chances of doing useful work. This is not a case of fire and forget, either; constant alignment throughout the project is required to ensure course corrections are made as requirements change.
Analysts and data scientists are by their very nature inquisitive, curious folk. This characteristic tends to lead to the occasional use of techniques that are new, complex or fun rather than appropriate for a given task. Skunkworks style operations and large, meandering exploratory analysis projects are commonplace in many data teams. There is nothing inherently wrong with new, complex or fun techniques once they are an effective, efficient choice for the task at hand. As much as I dislike the phrase - I feel it seems like an oxymoron - it is important to manage innovation appropriately and balance freedom to innovate with clearly and effectively delivering on important business goals.
This kind of resource wastage can also come from the business itself. Poorly formed questions, requests for analysis not connected to a clear outcome and zombie reporting (periodic reports with a purpose long forgotten) can all drain analytical capacity. It is the mandate of a data function to challenge other teams; challenge them not only to make better decisions, but challenge them to define which decisions they need to make in the first place. If a piece of work is not aligned to a business problem, be merciless in challenging its value.
Avoid emergency landings. Communicate the intent and progress of an analysis with stakeholders throughout the project and it will be far more likely to land in the right place and less likely to fly around looking for somewhere to land before crashing with little business impact. Doing this right usually equates with having a clear understanding of stakeholder priorities. If they are not clear enough, work together to set them.
Finally, be proactive; if you do not know why you are doing something, or if you are not sure if something will be useful, ask. Pilots do not wait for a warning light before taking corrective actions and neither should you. Seek clarity, challenge the value of the work and communicate openly. Do these things and you will increase the likelihood of a smooth landing every time.