During the COVID lockdowns, a data scientist I know set out to land her first data leadership role.
Stuck home with nothing better to do, she spent her weekends and evenings boosting her already impressive technical skills.
By the end of the lockdowns she had passed:
- 8 cloud certification exams;
- the TensorFlow Developer Certificate exam; and
- the notoriously hard OSCP (Cybersecurity) exam.
When the job was advertised, she just knew the promotion was hers for the taking. After all, how could the selection panel not be impressed with her skills?
She blitzed the technical assessment – scoring far higher than any of her peers – and was certain the interview went fine.
Guess who got the job?
Spoiler alert: Not her.
In fact, the feedback she received was “too technical”.
Instead, the job went to someone who could barely code, but had been with the business for years and knew exactly what stakeholders wanted.
Welcome to the paradox of data science careers where the technical skills that opened the door won’t help you climb the ladder.
To understand this paradox, we need to look at how data scientists are trained.
What Data Science Education Really Prepares You For
Data science degrees and bootcamps have one primary goal – to help students land their first job. This means focusing on the technical fundamentals, with business skills taking a back seat.
But in doing so, this gives students the false idea that technical excellence is what data science is all about.
Technical skills are just half of data science, with business skills comprising the rest.
And in the absence of a mentor to explain this reality, it can lead to four common career limiting mistakes.
Left uncorrected, these mistakes can leave you stuck building basic dashboards while others get to tackle the most exciting projects, or worse, prevent you from getting the promotion you deserve.
In this article, I’ll explain not just what these mistakes are, but exactly how to fix them and transform your career impact and earnings potential.
The 4 Hidden Mistakes That Keep Data Scientists From Earning What They're Worth
1. Growing Technical Skills Instead of Business Savvy
"Your subject matter expertise gets you to the table... but it's your process expertise that gets you hired."
Have you ever noticed that data scientists have a tendency to collect technical skills and certifications as though they’re Pokémon – much like my friend in the introduction?
I’m not judging. I used to do the same.
I think it’s because, once you embark on a career built solely on technical expertise, it then becomes necessary to maintain that expertise.
I’m constantly hearing the following from data scientists looking to advance their careers:
“I’m focusing on my technical skills for now, and I’ll learn the business skills once I get the promotion.”
But it’s a race you can never win.
With new data science and AI technologies being released every other day, there is always another skill to learn or course to complete. Eventually, you find yourself learning skills that aren’t even relevant to your current role – just in case.
The cloud certifications my friend earned? They were across AWS, Azure and Google – to be prepared for whatever might arise.
What she was really doing was buying an insurance policy for imagined future opportunities, while ignoring the opportunities right in front of her nose.
I get it. Growing your technical skills feels safe and you can tell yourself you’re “doing something” for your career. But at the end of the day, you’re just procrastinating from growing the skills that might be outside your comfort zone but that businesses actually value in data leaders.
What actually drives career advancement are skills such as:
- The ability to translate business challenges into data opportunities;
- Communication skills to influence non-technical stakeholders;
- Strategic thinking that connects data work to business outcomes; and
- Project prioritisation based on organisational impact, not technical interest.
Ask yourself who creates the greater business value: A data scientist who can build cutting-edge models far beyond what the organisation can implement, maintain or understand, or one who can deliver practical solutions to real business problems using the technical capabilities that already exist?
Leadership knows the answer – and they promote accordingly.
To paraphrase Blair Enns, your technical skills get you to the table, but it’s your business skills that build your career.
What to Do Instead?
Technical skills aren’t Pokémon. You don’t have to “catch ’em all”.
Breaking free from the certification collection mindset isn’t easy, but it’s essential if you want to grow your data science career.
Learn the technical fundamentals of data science by all means – you can’t build a career without them. But then deliberately shift your focus to developing the business skills that leverage your existing technical skills.
For example, rather than optimising your model to increase the accuracy by an additional 0.1%, spend the time translating your model results into business metrics (such as dollars and cents) instead.
If you need any additional technical skills along the way, learn those as you go – guided by business needs, not what looks good on your resume.
What are the technical fundamentals of data science? That’s something I’m going to discuss in a future article. Sign up to my mailing list to make sure you don’t miss it.
2. Taking Orders Instead of Taking Charge
"If you're a waiter taking orders, that's fine... But if you want to do interesting work and really solve problems and do amazing things and advance your career, then you need to acknowledge that your stakeholders don't always know what they need. And the only way to help them figure it out is to have these conversations with them."
It was the mid-2010s and data science was brand new – to my organisation and the rest of the world, too.
To raise awareness of our new data science team, we held a drop-in session.
We commandeered a corner of the office kitchen; handed out flyers with examples of our work and a list of what we could achieve – highlighting our shiny, new technical skills, of course; and asked any passers-by what we could do to help.
At the time, we thought we were showcasing our expertise. In retrospect, we were inadvertently positioning ourselves as the data science equivalent of McDonald’s.
We might as well have started asking stakeholders: “would you like a regression with that?”
If you find yourself asking stakeholders “what would you like me to build?”, rather than “what business problems are you trying to solve?”, then you’ve fallen into this trap, too.
Becoming a data science order-taker is an easy trap to fall into because most professions work this way – from plumbers to tax accountants and lawyers.
Before becoming a data scientist, I worked as a pricing actuary and my role was clear – come up with prices for new insurance contracts. It was in my job description, so I knew exactly what to do.
But whereas an insurance company knows they need prices for their insurance contracts and can clearly articulate that requirement; most stakeholders don’t actually know what they need when it comes to data science.
They might know that they have a business problem, but they lack the knowledge and experience to envision how data can create a solution – a gap that the data science role exists to fill.
In my experience, asking non-technical stakeholders for their data product needs tends to lead to two types of requests:
- Ones that are trivially small and barely worth the time – such as simple reports and Excel formatting requests; or
- Ones that are so complex as to be impossible given the data and tools at hand – such as predicting insurance claims with perfect accuracy up to 5 years from now.
At our drop-in session, we got both types of request. And we spent the next 6 months bouncing between underwhelming projects and doomed moonshots.
The result? Six months of under-utilised skills and a wasted opportunity to prove we could do more.
We failed to deliver on the more challenging requests while establishing a reputation as Excel support.
When author of “Stakeholder Whispering” Bill Shander says “your stakeholders don’t always know what they need,” he’s describing exactly what went wrong with our drop-in session approach.
We acted like order takers and expected our stakeholders to come to us with predefined data science problems to solve; when really, we should have acted like the strategic partners we wanted to be and diagnosed their problems for ourselves – asking the right questions to understand their challenges, then proposing data solutions they might never have considered.
It’s the data scientists who position themselves as strategic problem-solvers who reap the rewards of greater income, impact and influence.
So, ask yourself, where would you rather be: the drive-through window or the decision-making table?
What to Do Instead?
The high-impact projects you dream about aren’t just going to land on your desk. If you want to advance your career, you need to take control and uncover them for yourself.
Step away from your desk and start asking questions, such as: what strategic objectives are your stakeholders trying to achieve?
Understand the business context, your stakeholders real concerns and how business value flows.
Then use that knowledge to identify the highest-value areas for improvement and propose data science solutions that can get the business there.
This approach isn’t just a one-time effort, but a whole new way of doing work. And it’s the mindset shift that this approach involves that sets the data science leaders apart.
What other questions should you ask your stakeholders to uncover their true needs? That’s something I asked Bill Shander in an episode of “Value Driven Data Science” and he shared his six-questions framework for doing just that with me. You can listen to our complete discussion HERE.
3. Building Shiny POCs That Never Make It To Production
"Most analytics and AI projects fail because operationalization is only addressed as an afterthought."
After learning so many impressive data science skills, it’s only natural to want to put them to use. And when those exciting opportunities aren’t landing in your lap because you’re focussed on technical skills instead of business problems (mistake #1) and you’re waiting for orders instead of uncovering needs (mistake #2), there’s a predictable next step…
You build a proof-of-concept (POC) to showcase your technical prowess.
From cutting-edge AI projects to complex neural network/random forest ensembles, the technical sophistication buried in Jupyter notebooks on data scientists’ laptops would impress even the most seasoned practitioners.
But these projects rarely ever make it to production – instead becoming digital ghosts that contribute nothing to your career advancement or the company’s bottom line.
Why?
Because many POC projects suffer from two fatal flaws:
First, they don’t actually solve any pressing business problems – they solve technically interesting problems instead. They’re the equivalent of the portfolio projects data scientists are advised to build when looking for their first job.
Second, even when they do address real business needs, they often require significant operational changes that the organisation is unwilling or unable to implement. The solution might be brilliant, but if it demands rebuilding entire workflows and processes, it’s unlikely to ever see the light of day.
I know a data scientist who recently ran into this exact problem . He had built a sophisticated anomaly detection model in a Jupyter notebook that, if put into production, he was sure would reduce the number of defective items by 15%.
But he couldn’t get the business support he needed because implementing it would require changes to three different systems and retraining an entire department. After months of work, the model just went to waste – along with his opportunity to demonstrate value to leadership.
Eric Siegel highlights the disconnect between model development and production in “The AI Playbook”:
“An ML project is a business endeavour, not simply a technical one that can be handed off to data scientists to take on alone. After all, a model is going to directly change business operations, so the project requires a wholly collaborative process driven by business needs.” (p.42)
This disconnect doesn’t just waste time and resource – it’s also actively holding back your career.
Every impressive model that stays trapped in your laptop is another missed opportunity to advance your career.
But how can a data scientist bridge this gap between technical possibility and business reality?
What to Do Instead?
Instead of building POCs, try building minimum viable products (MVPs) that prioritise implementation over technical sophistication..
Start with the simplest possible solution to a real business problem and focus on getting that into stakeholders’ hand as fast as you can – even if you have to deliver insights by email, Excel spreadsheets, or writing numbers on a post-it note and attaching it to their computer.
A high school maths teacher and aspiring data scientist I know exemplified this approach perfectly. He identified a specific data need among faculty, developed a simple but valuable report, and then emailed it to his fellow staff on the first day of each month, directly from his computer. No fancy deployment, no complex infrastructure – just direct delivery of insights that solved a real business problem.
This approach accomplishes two critical things: it confirms whether your solution address a real business need and it demonstrates value right away – both essential for career advancement.
Once the business need is validated and stakeholders have experienced the initial value, you can iterate and enhance your solution based on actual stakeholder feedback, building the case for more substantial organisational change as the business impact becomes undeniable.
This is exactly the sort of collaborative approach advocated for by Siegel, but with the minimum of friction along the way.
A simple solution might not seem so impressive to your data science peers, but it is infinitely more valuable than a sophisticated one that never leaves your laptop.
4. Letting Your Wins Die In Silence
"Numbers have an important story to tell. They rely on you to give them a clear and convincing voice."
But what if you do finally succeed and deliver a data science solution that solves a real business problem? Surely that should be enough to guarantee career success?
If only it were that easy.
Data scientists frequently make the mistake of believing that their work speaks for themselves, when in reality nothing could be further from the truth.
This silence happens for several reasons.
Many data scientists are uncomfortable with communicating their wins, through presentations and written reports, preferring to leave it for their manager to do.
Others become bored once the problem is solved, and immediately chase the next challenge instead of optimising the value of their success and demonstrating its full business impact.
And some simply lack the business vocabulary to translate their technical achievements in to terms executives understand.
The cost? Your wins remain invisible to the decision makers who control your career trajectory, while colleagues who are able to communicate the business benefits of their less impressive results command the promotions that should rightfully be yours.
I’ve seen this play out time and again in job interviews. When asked to give an example of previous work, candidates who simply say “I built a classification model” rarely ever advance. But a candidate who tells me how their classification model increased customer retention by 12% and generated $1.5m in additional revenue? They’ve almost certainly got the job.
Same work. Completely different outcome.
Of the four mistakes, this is the greatest of them all because it can easily negate months of brilliant work.
To paraphrase the old proverb, if a data scientist saves an organisation $500k and nobody is around to quantify it, did the saving even happen?
What to Do Instead?
Management guru Peter Drucker famously said, “what gets measured gets managed”.
If you’re serious about managing the growth of your career, then you need to start measuring the value of your work and communicating the impact you’re creating.
This isn’t bragging or “someone else’s” job. Translating your technical accomplishments into the language leadership understands is the crucial last step that closes the loop between technical solution and business value.
As a data scientist, measurement should come naturally to you. Before starting your next piece of work:
- Identify the specific business metrics your solution is expected to influence;
- Establish a clear baseline of metrics before implementation;
- Design a simple comparison study to quantify the before-and-after impact; and
- Translate these technical measurements into business outcomes like revenue, cost savings or time saved.
For example, instead of saying “I built a customer churn model with 86% accuracy”, this strategy puts you in the position to say “My customer retention solution identified at-risk accounts worth $3.2M in annual revenue, allowing us to retain 74% of them through targeted interventions.”
Once you quantify your impact in business terms, craft a narrative that explains to stakeholders why this should matter to them by answering these three essential questions:
- What? (i.e. what is your solution and its measurable results?)
- So What? (i.e. why do these results matter to the business?)
- Now What? (i.e. what should the business do now to make use of these results?)
If you want to learn more about data story telling, I interviewed Dr Selena Fisk on this topic HERE.
The data scientists who advance fastest aren’t necessarily those who create the most sophisticated models – they’re the ones who effectively communicate the business value of the work they do, no matter how simple that work may seem.
So What?
Technical skills are an essential foundation of any data science career – but they aren’t the only thing necessary to excel. And like everything in business, investment in technical skills follows the law of diminishing returns.
The most successful data scientists aren’t necessarily those with the deepest technical knowledge, but rather those who can translate whatever skills they possess into business solutions that matter. These are the data scientists whose careers and compensation consistently outpace their technically brilliant peers.
Whatever technical skills you currently possess, they’re likely already sufficient for your next career leap. What’s missing isn’t more algorithms – it’s strategic application.
Instead of signing up for yet another certification, focus on these three career accelerators instead:
- Develop a deep understanding of your organisation and the challenges your stakeholders are actually trying to solve;
- Deliver simple, implementable solutions to today’s pressing problems rather than theoretical answers to tomorrow’s imagined challenges;
- Quantify and communicate the business impact of your work in language decision-makers understand and value.
This strategic shift is what transforms your perception from “just another data person” to “indispensable business partner” – the kind executives turn to first and compensate accordingly.
Are you ready to stop collecting skills and start accelerating your impact?
If so, but you’re not sure exactly how to make this transition, I can help. My Data Science Impact Accelerator program is specifically designed for technically-skilled data scientists who want to maximise their business impact and earnings potential.
In just 12 weeks, I’ll guide you through developing the exact business skills covered in this article – from uncovering high-value opportunities to quantifying and communicating your impact. To learn more about how you can transform your data science expertise into real business value and financial rewards, reach out to me HERE.
