What are companies actually looking for when hiring for Data Analytics?

And why does it seem like the market is saturated when there are reports that this field is rapidly growing?

Source: US Bureau of Labor Statistics

Spend any amount of time on data Reddit or similar online communities, and you’ll see the question “Is the data analytics job market saturated?” or some variation.

Even though data science and other data roles are in high demand relative to other jobs, it certainly doesn’t seem that way to anyone trying to break into the field.

The reality is that Data Analytics is saturated when it comes to entry-level roles. For one thing, there aren’t that many truly entry or junior level roles to begin with. Despite that, there are hordes of people coming out of masters programs, bootcamps, and online courses who know some combination of Excel, SQL, Tableau/PowerBI, Python/R, statistics, etc, but have no relevant experience. So you have a ton of people competing for the few entry-level opportunities that will consider them.

But Data Analytics is not quite as saturated when it comes to the level of candidates that companies actually want. Talk to any hiring manager looking for a mid- or senior-level candidate, especially for hybrid/in-person roles, and they’ll tell you it’s actually very hard to find the right candidate, despite getting hundreds or thousands of applicants.

So what is it that they actually want? It’s not just technical skills and experience but a mix of competencies for the following. Read on for more details.

• Technical competence.

• Statistical knowledge.

• Analytical knowledge.

• Industry or domain knowledge.

• Problem solving.

• Good communication.

Technical Competence

Excel, SQL, Tableau and/or Power BI, maybe Python or R, plus industry-specific tools like Adobe or Google Analytics, Salesforce, GIS tools, etc.

You’ve heard over and over again that you need to know these tools. You’ve learned them via online courses or other means. But is it enough?

The reality that I hear from hiring managers is that a lot of candidates fail their technical assessments. I used to be one of those candidates until I really took the time to learn exactly what interviewers were looking for and then practice. Being good at SQL during a live interview assessment is not quite the same as being good at SQL on the job. For one thing, I don’t talk about my query out loud on the job. But talking through your solution is a key step of passing an interview.

There’s also the difference between your code running, your code returning the correct result, and your code running correctly and optimally.

And then there is the ability to answer questions about why you wrote your code the way you did. I’ve been writing SQL for years but after a recent conversation with someone, I stopped to reflect if I should be using LEFT JOIN as often as I do instead of INNER JOIN.

So how do you get better? Consult sources on best practices. Here is a rundown of the framework I use for SQL interviews. Also look at other people’s code. When you solve a problem on StrataScratch (or Hacker Rank or whatever you use), check the solutions that other people submitted. If you have a job, check out your coworkers’ SQL queries. I have learned a lot that way. It can be very enlightening.

Statistical Knowledge

I see a lot of folks learn SQL and then think they can jump into data analytics.

Not so fast. How do you know you’re looking at the right data? Analyzing it accurately? How do you know that your conclusions are good?

It pains me when certain data influencers say that you don’t need to be good at math to do data analytics. Yes, it is true, the computer does the math for you. Obviously I am not calculating averages by hand.

But do you know when to use median instead of mean (average)? Do you know why you shouldn’t report an average of averages as a population average? Do you know how to properly calculate lift? Why you should calculate confidence intervals (and add them to your visuals)? How to calculate outliers - and when to keep or remove them? How to properly run an A/B test? How to check for correlations? Or how to measure the relationship between your data and a target outcome? How to decide which independent variables to include in your model? When to normalize your variables for a predictive model? Which accuracy metric to select?

While your stakeholders might not care about the math - especially if they are non-technical - using proper techniques is important for good analysis and reporting results that are accurate. Just because two populations have a different mean doesn’t mean it’s statistically significant. You may have reported randomness as a conclusive fact.

Not sure where to get started? Start with the foundations.

Analytical Knowledge

Beyond knowing stats, you also need to know how to make sense of data.

Which metrics should be reported as a volume and which should be a rate or percentage? And how do you calculate those rates or percentages? What do you use as your numerator and more importantly, your denominator?

What should the success metrics be for your company or for a specific project or A/B test? You can’t use all of them, so how do you justify what’s in the top 3-5?

This is something that comes with experience, domain knowledge, and good critical thinking. It’s can be hard to teach, but it’s why the Google Data Analytics Certificate starts with foundations instead of jumping right into technical skills.

A rough framework is to always make sure you are clear on “what is the problem we’re trying to solve?” or “what does success look like?”

Industry or Domain Knowledge

We might all use the same tools and do the same math, but there are differences when it comes to useful analytics for tech versus healthcare versus finance versus other industries, and then marketing versus product versus sales versus supply versus HR versus other functions.

I was able to pivot from marketing analytics to product analytics because both of my roles were primarily web analytics. But I would struggle in an analytics role in healthcare or finance or supply chain.

My knowledge of product/web makes me a much more valuable partner to the teams I support. I know the industry jargon and best practices. I’ve solved similar problems. I can think about the big picture. I can handle vague questions.

A lot of your business partners won’t have a good understanding of the data available or the possibilities of analysis. If you have a good understanding of their work, you will be better at meeting them where they are. You can also build trust which is so essential to the role.

This is why a lot of folks working in this field pivoted from another career. I pivoted from marketing and my colleagues pivoted from accounting, finance, business development, customer service, and software engineering.

Problem Solving

Data Analytics is a relatively new function at a lot of companies. Meaning some teams are still figuring out what they can do and what problems they can solve. The business will have some vague ideas for what they want you to do, but you have to do a lot of work to figure out how to take those vague ideas and come up with good solutions using data. And sometimes the solution is a brand new idea that has never been carried out at the company before - and you have to figure out how to carry it out.

Additionally, most data teams are too small to spend much time with training and onboarding. So often you have to fill in those gaps yourself.

Good Communication

The teams you support might be non-technical and not very data literate. I’ve seen a lot of really smart Data Analysts and Data Scientists lose their audience because they were too technical and detailed in their presentations.

You have to be good at communicating complex ideas to people who are 1) hearing these concepts and ideas for the first time, and 2) do not come from a technical or statistical background.

Too many new Data Analysts think the data will just “speak for itself.” It will not. You have to meet your audience where they are and tell them the story of your project and your analysis, why it is important, and what you want them to take away. But in a succinct and impactful way so you don’t lose their attention.

Need help? A lot of people swear by the book Storytelling with Data, and there are also online courses available such as this and this that are part of the Google course.

---

Does that sound like a lot?

To be frank, this is why the median salary for Data Analysts in the US is $82k (almost 25% more than the median household income in the US) and the median salary for Data Scientists in the US is $104k (over 50% more than the median household income in the US). No one is just handing out these high salaries to folks who learned a couple of skills. I don’t say that to gatekeep but to be realistic. Companies have very high expectations for their Data Analytics teams - after all, they are overhead and costing the company money, so they are constantly under pressure to prove their value. Despite what other data influencers may have told you, this can be a tough field to master, and unfortunately, thanks to layoffs and a slowdown in hiring, it has only gotten more competitive to land a job.

So what can you do?

If you’re struggling to get interviews or an offer for a proper data analytics role, my advice is always to get whatever role you can. As I mentioned above, a lot of people in this field started their career somewhere else. You can get data analysis experience in a role that doesn’t have anything like “data” or “analyst” in the title. I got my hands on data for the first time when my title was PR & Marketing Associate and then continued to work with data when my title was Digital Marketing Manager. I had years of working with data (among other tasks) before I got a title with analytics in it. And all of that other experience made me a much more well-rounded Data Scientist which has helped not only my work by my internal relationships with colleagues.