Hi ,
"I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes." - Sci-fi Author Joanna Maciejewska
Are you delivering the data science results your stakeholders actually want or the results you want to deliver?
If you're not sure, then ask yourself if your results are being used. If the answer is "no", then it's probably the latter. And something is going to have to change.
Because not all data science outputs are created equal.
I learned this the hard way back in the early days of data science.
My team had been set the task of delivering "15 data science projects" for the financial year. That was exactly how it
was written in the organisational plan for the year - it was the equivalent of asking an artist to deliver "100 square metres of art".
But it never occurred to any of us to question this task. We just set out to get it done.
Armed with the skills we had learned from
our data science training, we spent the next year applying all that we could to the organisation's data - and then emailing the results to anyone we thought would care.
Guess what? None of them actually did.
That year, we delivered 15 technically impressive projects
that gathered digital dust while the real business problems went unsolved.
Our "success" was reported in the following year's annual report. But the success metrics missed the point.
Looking back on it, I'm mildly embarrassed to admit to ever doing something this
"dumb". But in those days, I genuinely believed that all data science projects were intrinsically valuable.
What I learned that year was that data science without stakeholder alignment is just expensive maths.
If I could go back in time and talk to my former self,
here's what I would tell her to do - read the rest of the organisational plan.
Undoubtedly, buried in there were the actual business priorities, and contributing to solving them was what the writers of the plan really wanted the data science team to do.
Understanding what your stakeholders actually want is the key to creating data science value.
Just asking "what decisions would you like help to make?" instead of "what analysis would you like me to run?" can often be enough. Applying the same technical skills with purpose can increase their value by 10X.
In my experience, stakeholders rarely hide what they're trying to achieve - the answers are probably sitting there, right under your nose.
The responsibility lies with data scientists to do our research and ask the right questions, so that we can deliver what our stakeholders truly need - not what we think that they should
want.
What about you?
When was the last time you verified your data science work aligned with your stakeholders' needs?
Talk again soon,
Dr Genevieve Hayes.
p.s. Thanks to friend-of-the-list, Dr Peter Prevos, for sharing the quote with me.
