It is so easy to go into analysis paralysis. Where you keep digging and digging into the data. Slicing and dicing it in a myriad of non exhaustive combinations. Visualizing what you find in many slightly different ways, to tell the story in a novel way. Adding new data sources, looking at different time periods.
And then, to add onto this, there are always more questions. What if we look at this? What if we try that? What if?
In my experience it helps to drive towards practical real world action. With action, generally, comes timelines and due dates. Which gives you some sort of end date and thus limited time to focus your analysis, prioritizing the analytics avenues with the most potential.
When you get stuck into the data, there is inevitably all kinds of interesting off ramps luring you in. Staying away from these requires great discipline. Stick to what is relevant for the current ask.
Question the questions
Data insights can become addictive as entertainment, or what can be called fun facts. In my experience there are always more questions than answers. Which is good, in a way, as it shows that our collective thinking is expanding.
We can however not answer all of the questions in one go. And I would argue that not all questions necessarily need to be answered. Hear them out, carefully consider them YES. Then prioritize.
It is also worth understanding where the question is coming from. On the one hand, questioning the question will challenge whether it holds merit. At the same time it will help you understand more fundamentally what the stakeholder is looking for.
Rather than only being a one way street, where you as a data analyst just field a stream of questions and try your level best to keep feeding answers back, it should be a conversation. As you receive requests, you should make sure that you understand the why behind each one of them, trying your level best to get to the root of the matter.
What to do when
Data analytics in itself is not going to change anything. It is only through taking action off the back of the insights gained that impact can be achieved. Having an understanding of what is possible in terms of practical implementation is thus key.
Have a discussion upfront with the stakeholders about how they foresee bringing what you present to life. Obviously the nature of your insight might ultimately change that, but it is still good to have directional ideas in the beginning.
With some form of execution plan in mind, there will generally be associated timelines. Which does introduce pressure into the work, but also provides a limited time window of analysis, forcing a level of focus. This is what you want for your data analysis, boundaries.
Boundaries, and the focus that comes with that, keeps you from getting stuck in an endless loop of data work. When you've got a tangible outcome, with a date by which it needs to be done, you're forced to deliver within that time frame. Just like with a piece of art, at some point you will need to put the brush down, and stick with what you have. At least for now. The nicely packaged product can always be picked up at a later stage to feed into a seperate project.
Focus focus refocus
Data analysis is interesting. Before you know it, you could very well have wandered down a rabbit hole that does not actually directly serve the project scope. Although it might somehow be related. Staying the course requires a great deal of discipline and mindfulness, coupled with regular feedback.
Additionally, your stakeholders are also very likely to get sidetracked. Mostly because another related business request has surfaced. Sometimes there will be a need to review the scope, which is why ongoing engagement is so important. Open and honest discussions will be needed around the implications of the change, to agree on what will be possible within the timeframe and which elements of the delivery will need to be deprioritized.
Tough conversations often need to be had. Given limited time, there is only so much that can be analysed. Different stakeholders might also have different opinions regarding what is most important. The more projects you are exposed to, the better you will be able to gauge approximately how much time a specific task will require to be completed. It is never perfect, but an estimate is better than nothing.
Most importantly, always build a buffer into your plans. You definitely don't want to still be stuck deep in the data the night before due day. Leave enough time to properly tie up all of the loose ends, double check your numbers, and especially to build a coherent story for presentation. If the message is not landed properly, all of the work in the background will be wasted.
Where does it end?
Data analysis can be a never ending story. Without a logical end you can potentially keep analysing forever. And still not add any tangible value.
Towards ensuring that your work does come to a conclusion, with real impact and outcomes:
- You need to actively engage your stakeholders in questioning the thinking behind their queries towards working with them to get to the root of the challenge and crystallizing the analysis goals.
- Action, action, action. From the beginning it is key to start having discussions about how you foresee the insights being brought to life into decisions and changes to how we do things. Insights can change hearts and minds, but the result of that needs to be implemented somehow.
- Data analytics is like a laser pointer on steroids. Getting distracted is par for the course. Remaining focused in your delivery thus requires a great deal of effort and discipline.
There is not necessarily a natural end to data analysis work. By asking the questions up front, driving towards tangible action and focusing the delivery it is however possible to produce pockets of impactful insights.