Near enough IS good enough when it come to data.

I love making up stats...

Thanks for waiting! I’ve been AWOL that last few weeks but we are back.

This week I am exploring a touchy subject: How accurate does your data really need to be when informing the business?

My vote: Not as much a you may think. Let’s explore ⬇️

❤️ I love making up stats.

95.3% of data professionals are absolute perfectionists.

After all, we think of data as a science and strive to be as exact as we possibly can. Often to a fault.

The problem is that as data professionals in corporate and commercial settings, our analyses don’t have the luxury of waiting for perfection. Business is increasingly autonomous, distributed and asynchronous. For our work to drive impact it needs to be done, and done FAST.

👌🏼Perfection Paralysis

The quest for absolute accuracy will slow down your ability to deliver. You’ll have less opportunity to impact business decision making, have less exposure to crucial business conversation and will overall be less influential both in terms of the work you deliver and as a leader.

Striving for perfection and 100% assuredness in our work will results in us failing to achieve our key objective as data professionals:

👉🏼To strategically advise the business.

We can’t do this if we’re slow to react.

☝🏻 Directionally Indicative

You don’t need to know “for sure”, you need a directional indicator.

Unless we’re running statistically sound AB tests on every thing we do - which no business is - it will be very unlikely that you’ll have a confidence level anywhere near what is expected in the CRO discipline. And when it comes to business recommendations - you don’t always need it.

To make agile decisions, we need to get comfortable working with a degree of uncertainty and base decisions on data that is available and accessible.

🔄 The Trade Off

There will always be tension between perfect accuracy and timely, actionable insights. The goal here for data professionals will be to know what can be expedited and what needs a more thorough approach.

On the whole, so long as we have directionally indicative signals, most pieces of work won’t require us to to ‘go all out’.

We need to reflect on how the data we produce is likely to be used by stakeholders and whether or not the level of effort put into the work will be reflected in the level of appreciation of value extracted from it’s creation.

👋 Hi There,

For those who are new here, my name is Tristan (Tris) Burns.

My mission in life is to help data leaders develop strategies to become more influential so that they can transform the data culture within their organisations whilst growing in their careers.

I love nothing more than connecting with and helping data leaders around there world!

Whenever you are ready, there are a few ways I’d love to help you: