The Five Foundational Data Skills
The building blocks of great data work at every level
I believe there are five foundational data skills that are highly relevant for all types of data work across all levels of seniority.
They matter when you're a junior data person, and they matter just as much when you're a Chief Data Officer.
Coming up with these has taken many years of synthesizing lessons from all the great data people I’ve worked with and managed over the years.
Until now these ideas have only been a part of my professional development program, but I'm sharing them more broadly because I think they can be very valuable to the data community.
They're helpful when you want to evaluate your own skill as a data person, or if you're a leader looking to evaluate your team.
The Skills
Here they are, and then we'll go through each of them in detail:
Strong Mental Models
Deep (enough) Tech Skills
Executive-level relationship building and communication skills
Active, Supportive and Outcome Focused approach
Data, project and stakeholder management skills
Let's go through all of them.
Strong Mental Models
Mental models are the internal representation inside your brain of the way a system works in the real world. For the purpose of data work, I like to break this overall skill into three different acumens:
Business Acumen
Systems Acumen
Organizational Acumen
Business Acumen
Business acumen is a mix of how well you understand your specific business, as well as your general business skills.
Do you know the key levers that influence why customers buy your products and why they stay or churn? Does your sales team want to bring you into sales conversations? Does the marketing leader want to get your input into the tracking strategy for their latest campaign?
There are levels of business acumen (and all the other skills in the five foundational data skills).
The level of business acumen you expect from a junior data engineer is far different from the level of business acumen you expect in a Chief Data Officer.
However, you can compare people to the expected levels for their positions and it can help you understand where people are strong vs. where people need improvement.
You should look at both components - general business skills and specific knowledge of your business and industry.
Systems Acumen
This is how well people understand the specific systems in your company - not their general level of technical skills.
Do you only know the database you spend all your time in, or do you understand all the various systems in the company and how the nuances of those systems contribute to the way the data looks.
How frequently can you solve your own problems in the database vs. needing to rely on a software team somewhere to respond to your email and tell you whether it's the `orders_new` or the `orders_latest` table that has the current calculations finance is using for board reporting?
(obviously it’s actually the `orders_latest_new` table you didn’t notice…)
You can have amazing technical skills and terrible systems acumen.
You might be able to write the most amazing machine learning models - but not understanding your systems means you will struggle to get any work done.
And on the opposite side, you might have 20 years experience and know every nook and cranny of your company's systems, but have never written a line of python in your life and wouldn't even know where to get started with writing a machine learning model.
Organizational Acumen
As a data person, your impact will come through other people changing what they do based on data. And getting groups of people inside a company to change is not easy.
Organizational acumen is your understanding of your company and how to get things done inside your company.
For example:
Do you know whether the CRO and CMO get along or hate each other?
Do you know what basic questions are landmines that you should never bring up in a leadership meeting?
Do you understand which people on the product team have power and which people don't, regardless of title?
Do you know the engineers and can you get them to do occasional small things when you need?
If you don't know something, do you know where in the organization you can go to get it?
The Acumens
Between those three acumens: Business Acumen, Systems Acumen and Organizational Acumen, you can measure your understanding of the various components needed to be successful in a data job.
If you have an accurate mental representation of your business, your systems and your organization, you will be HIGHLY effective.
Deep (enough) Technical Skills
You need to be good technically to be a data person. Data work is all about combining technical skills and business skills, and without those technical skills data work doesn't exist.
Different data roles will have different requirements for technical skills. Data analysts will need different skills than data engineers who will need different skills than data scientists or data PMs.
Not only across roles, different specific jobs within the same role will have different requirements for technical skills.
For example, a data scientist who specializes in building models for financial fraud detection will have different technical expertise than a data scientist who focuses on computer vision and object detection.
Being world class in computer vision when your job is fraud detection does not help you do better. However, when your deep expertise is aligned with the specifics of your job it makes you far more capable.
This is why I call this "Deep (enough) Technical Skills."
One of the most common failure modes for data people when it comes to professional development is that they keep thinking they need to learn more technical skills, even though those skills aren't necessary for their jobs.
In the meantime, all the other areas they could devote their time to that would make an actual difference don't get much focus.
Instead of learning yet another technical tool you don't use, either focus on developing the non-technical skills that are more likely to be holding you back, or spend your time deepening the SPECIFIC technical skills you use everyday to become a true expert.
When you're judging yourself or your team for this, benchmark your skills against other people with the same job / title. Data engineers will generally have stronger software engineering skills than data analysts, but if you want to be promoted, you need to have standout skills against other data engineers.
Executive-level relationship building and communications
Data isn’t an end to itself. It’s only useful if it helps someone make a different decisions than they would have made without that data.
So, in order to have an actual business impact, you need to work with the people in the business who are doing the actual work of making and selling your product, and help them effectively use data.
To do this effectively, you need executive-level relationship building and communication skills.
Communication
Data is often technical and complicated. You need strong communication skills to translate whatever insights you found to other people.
How many times have you seen brilliant data work be wasted because the person who did the work couldn't figure out how to get the business people to understand it?
Relationships
There are two types of relationships that are crucial to build as a data person.
Strong Trusted Relationships
Weaker and Wide Relationships
Strong trusted relationships
If the data points to a great new optimization the company could do with the marketing budget, it's actually the marketing team that needs to make the change.
If you uncover a brilliant product insight or build a new churn prediction model - it's up to the product and engineering teams to actually build things to take advantage of those insights.
The people making these changes are sticking their necks out. If it turns out you gave them bad advice and they don’t hit their goals as a result, they’re the ones that are going to get fired, not you.
In order for these people to trust you - you need to build a strong relationship and earn that trust.
Weaker and wide relationships
More than almost any other role, in order to effectively do your job as a data person, you need to interface with people across the company and externally.
Very often, you need internal data or other technical work / business context from other teams inside your company.
You are also faced with new tasks all the time, where the number one step should be talking to other people at different companies who have done that before to learn lessons from their experience.
Building your network internally and externally only after you need it is too slow to be effective. Sadly many data people don’t even realize how much they’re missing by not focusing on this.
Executive Level
Data people disproportionately work with senior people and executives.
The relationships you build should be of the style executives build with each other as trusted advisors, and you should be comfortable working closely with people many levels above you on the org chart.
Active, Supportive, Outcome focused approach
This style of data work is highly effective in ensuring your knowledge, technical skills and relationships are channeled in a positive way to drive business results at your company.
Active
Active means you need to actively be moving things forward, you can't wait around for people to tell you what to do and how to do it. If you want data to actually drive business impact you need to make it happen.
Supportive
Supportive means that the way you impact the business is through supporting other business teams and you should lean into that.
Yes - you might be tasked with supporting the data for marketing. But if marketing misses their numbers it's the CMO who will be fired not you. Therefore, your job isn't to do whatever you personally think the marketing team needs, it's to support the CMO in executing their vision as best as possible.
90% of the time you will be on the same page, but 10% of the time you won't, and if you try to override the CMÕ, during that 10% of the time you will 100% ruin the relationship. (lessons from experience...)
Outcome Focused
Outcome focused means you ensure the work you're doing is actually making a difference in the business.
You can build the world's most amazing machine learning model, but if it never makes it to production because you couldn't get it on the product roadmap, was it really any good?
Data teams can easily spend all their time on back-end work to build out an amazing data infrastructure while the business barely makes use of that infrastructure.
You need to be outcome oriented and ensure that your work is actually driving real business results.
Data, Project and Stakeholder Management
This is how you keep your work as a data person organized. Data work is always a bit chaotic, but if it becomes TOO chaotic it's a problem.
Data Management
Data management is your ability to manage data as an asset. This includes data governance, security, and all the other things you need to know to make sure that your data can actually be used effectively.
Project Management
Project management is essential because as data people we are often our own product and project managers. We need to be able to plan out and execute our work and coordinate with all the other relevant teams that need to be in the loop.
Stakeholder Management
Stakeholder management is important because data people often have MANY people who can give them work and all have their own priorities. Effective stakeholder management is a must.
Like the NFL Combine
Analyzing someone through the lens of these foundational data skills isn't a substitute for an actual performance review, but it's important additional information that is generally a leading indicator of actual job performance.
I like to think of these skills like the NFL Combine, which is a scouting event for football players. They measure your speed, strength, agility, and performance on many drills for your specific position.
Being the fastest wide receiver doesn't automatically mean you'll score more touchdowns, but there's a reason the fastest receivers get drafted and the slowest ones don't.
If you could hire a data person with great foundational skills vs. weak foundational skills, you would want to hire the person with great foundational skills every time.
And by identifying particular skills as areas for improvement, you can help yourself or data people on your team find the areas where investing in their skills will make the largest difference.
These Skills Cover All The Bases
The five foundational data skills are designed to be all-encompassing to what you need for successful data work.
Knowing What To Do → Mental Models
Execution Skills → Deep enough tech skills & executive level relationship building and communication
Effective Approach → Active, Supportive and Outcome Oriented
In Control → Data, Project, Stakeholder Management
Being Advanced Means Being Advanced At The Basics
I believe that being a high level data person isn’t about knowing countless tricks, it’s about high level execution of the basics.
When you get to the next level, suddenly you’re comparing yourself to a new level of data people with more advanced foundational skills.
And it never ends! I’m still working on mine!