Should you use Scrum for analytics projects? This question never comes up in meetings. Why? Because it is usually predetermined by the company culture, background of the decision makers, whoever sits above analytics in the org chart or simply because they don't know any better. Software development companies impose agile on analytics simply because everything else they do follows this methodology as the right way to do things. Same with data analytics teams that are part of the IT department (questionable choice). Analytics is still treated as IT-something. After all, computers and software are involved, so it must be the same thing, right?
When the question does come up, it's usually asked by data analysts desperately trying to fit the methodology to their workflow.
What Do We Even Mean by Analytics?
The question about SCRUM and analytics is not well placed because it puts all sorts of analytical tasks in one bag. Analytics now is a hugely broad term.
It is used for getting the top 3 vendors from a clean Excel spreadsheet. It is also used for producing a complex multi-level forecasting. Analytics is used for an extensive multi-dashboard suite just as readily as for a simple 3-chart page. Analytics is also getting an answer to a complex business question that includes various data and context from multiple departments. Believe it or not, some people refer to setting filters in a spreadsheet also as analytics. The term is also used for implementation of an advanced ML and for writing an XLOOKUP. The examples could go on.
Insisting that this universe could be managed with the same method is lunacy. Yet, it is a fact in way too many organizations.
The management methodology should be decided with the specifics in mind if you care about the outcome, its quality and the team mental health.
For example, SCRUM works very well for building and maintaining BI products. It does not work for answering the CEO's question "What of all the 42 activities we did caused X to be higher than last year?".
Of course, there are people who will object and affirm SCRUM works everywhere and it is a matter of skill and knowing the method to apply. Those people either make money out of it or are in a desperate need of some upgrade of their knowledge of the world.
Overthrow SCRUM Keep Agile in Mind
Although SCRUM is not the answer to any analytical project, true agile approach, unbounded by frameworks and made-up metrics, is at home in most analytical tasks.
The agile mindset fits analytics naturally because analytical work is iterative by nature - you rarely know the full shape of the answer before you start looking. Short cycles of exploration, quick checks with the stakeholder, change direction when the data tells you to - this is just how good analytical work flows. No sprint ceremonies required. Agile without the SCRUM costume is just good analytical thinking with a name.
The good news is the cure is simple: match the method to the work, not the other way around. The bad news is that requires thinking, and thinking is harder to put on a Jira board.
Next time someone insists your analytical project needs story points and a sprint backlog, ask them one question: what exactly is the product? If they can't answer, you have your answer. SCRUM is a fine tool. So is a hammer. Neither belongs in a surgery.
No comments:
Post a Comment