Twice weekly, I share proven strategies to help data scientists get noticed, promoted, and valued. No theory — just practical steps to transform your technical expertise into business impact and the freedom to call your own shots.
The Decisions ML Can't Make
|
Data science exists to support decision making, but within any organisation, there is a hierarchy of decisions - low stakes, high volume decisions at the bottom; high stakes, low volume decisions at the top. Executives care most about the decisions at the top. There’s valuable work at the bottom of that hierarchy. But automating routine decisions is about clearing the path so executives can focus on the high-stakes decisions. It’s not about walking it with them. For years, data science has been synonymous with ML, but ML excels at the automation of high-volume decisions. When volume is low, it falls apart. From an executive perspective, these are the decisions that matter the least. Recently, I spoke to a property appraiser about the use of data science in his field. He made the following comment that summed this up: “You know, valuation models are now nearing the point where they might potentially be able to replace appraisers one day soon. But for the complex properties, the ones that stake up to build big risk, clients are unwilling to do that because the cost of getting this wrong is just too high.” His solution was to use data science not to replace his expertise, but to enhance it - by building statistical tools that support his job and provide him with evidence to back his recommendations. If your goal as a data scientist is to support executive decisions, you need to follow his lead. Executives want to reduce uncertainty around decisions that matter too much to get wrong, not to hand them off entirely. The path isn’t building better automated pipelines. It’s becoming the person who reduces that uncertainty. Talk again soon, Dr Genevieve Hayes |
Data Science Impact Algorithm
Twice weekly, I share proven strategies to help data scientists get noticed, promoted, and valued. No theory — just practical steps to transform your technical expertise into business impact and the freedom to call your own shots.
Around age 19, having just graduated high school, I got my first job in "data". I could barely use Excel and thought Python was something you'd find at the zoo, but a friend of the family hired me to tutor their teenaged son. My maths was good enough to help him get through high school maths. Around age 29, having just finished my PhD, I landed a role managing an insurance pricing and analytics team. I'd never heard the term "machine learning" back then, but I had spent a good chunk of the...
You’re already using optimisation every day as a data scientist - but you probably don’t think about it. Every time you train a machine learning model, you’re running an optimisation algorithm to find the best parameters. But there’s a second type of optimisation that most data scientists never even touch - and it’s where the real business value often lives. It’s called decision optimisation, and it can transform your ML predictions into actionable decisions. Here’s the difference: ML...
Building a data science model for an academic paper is one thing. Building a model that has to work perfectly during the Cricket World Cup with millions watching is something else entirely. There’s no room for the kind of errors that might be acceptable in research settings or even standard business applications. And if you get it wrong, you get emails… Lots of emails… From people all around the world… The type of emails that begin with: “how dare you you make my team lose?!” Worse still? You...