The Dos and Don’ts for Data Analysts

If you scroll through Data Analyst job descriptions, or write one yourself, you can’t miss the repetitive lists of technical requirements that Data Analyst candidates must have — SQL, Excel, BI tools, R/Python, and more.

This correlates really well with the usual Data Analyst interview process, where the candidate is asked multiple SQL questions, and advanced candidates are usually those who are familiar with advanced SQL terms, such as Window functions and CTEs.

So being an SQL champion = being a top notch analyst, right?

No. Strong SQL is NOT the most important skill for an analyst, nor any other technical skill you possess in your toolbox.

Mentoring over 30 analysts has made me realize that the business needs the analyst to have an opinion, and to have an opinion, the analyst must focus their time & effort on understanding the business.

What everybody thinks we should be doing

Early in my career, I tried to master as many technical languages and tools as I could get my hands on.
The fuller and bigger the toolbox, the better my chances of getting the job.
Makes perfect sense, right?

It’s been 7 years since I began my Data Analyst career, and here’s how many times a stakeholder has asked me which query language I’ve been using:
zero.
Don’t get me wrong, of course it’s important — you can’t be an analyst without SQL and a whole set of other data querying capabilities. Without them you can’t even begin your professional career.

The thing is, although the tools are important, nobody (except for maybe your direct manager) cares or knows. Analysts aren’t evaluated according to their technical skills, because that isn’t what they deliver.

The technical part of an analyst role is similar to a building’s foundations — essential, but hidden. Nobody sees the foundations.
Successful companies set their eyes on the outcome. It’s not about the tools, it’s strictly about how YOU use them.

What we really should be doing

As analysts, our main goal is to help stakeholders answer their business questions. That’s why an analyst should not only understand what is being asked, but actually dive a bit deeper and understand why it’s being asked.

In my career, I’ve encountered endless situations where the question I was asked to answer was actually a by-product of the real case worth analyzing.
Only by mastering the ability to deconstruct a complex business question into the building blocks which form it, will enable you to infer what the real problem that stands behind each question is.
Although it sounds straightforward, it is actually quite a challenge, but in order to deliver meaningful results you must master it.

After you’ve identified the What & Why, the next phase is to know to Whom you’re going to present your findings.
Although the data remains the same, the listeners on the other-side are usually heterogeneous and come from diverse backgrounds, each with their own wants and needs.
CSMs mainly care about the impact on their top clients. Management would like to know how it affects revenue. Product wants to know how it’s going to affect the upcoming new feature, and your analyst colleagues are thirsty for every piece of information.

Since you don’t want to waste your time trying to please everybody, it’s important to ask yourself who the audience is:

• Which departments are they from?
• Which regions do they represent?
• How senior are they?
• Have they encountered this topic before, or is it their first time?

This will not only help you better understand what to focus on prior to starting your analysis, but also on how you should present your findings when you finish.

And now comes the fun part….

The actual analysis

• Query that dataset bottom up with your favorite tool
• Slice & dice it with any visualization software you prefer (I always go with Excel pivot charts)
• Chew over that piece of data again from a different perspective, or even act like a Data Scientist and create predictions on top of your findings, based on the most advanced machine learning algorithms out there.
  Why Data Analysts are actually also Data Scientists will be my next post, stay tuned!

That’s it, You did it! Summarize your findings and start practicing your presentation.

Actually, it’s almost finished, as you should also be very careful of one of the common pitfalls when dealing with data — everything sounds interesting and worth presenting, but this is rarely the case. Usually from a big pile of insights, only a handful are actually relevant.

So the only thing left to do (and the most important!) is to make sure your analysis has actionable results.
Remember that your audience are rarely Data Analysts — so leave out long calculations, delete data processing slides and save the edge cases for yourself.

Focus only on the MESSAGE you want to deliver, and make sure that the audience will do something with what you’ve just presented to them — if it doesn’t trigger any response, it’s likely to be less relevant.

What do we need to have?

In this post I mainly wrote about what data analysts should and should not do, but I would like to end it by going beyond do’s and dont’s and share How it should be done:

1. Know your industry
   The true secret of becoming a successful analyst is to gain a deeper, more profound knowledge of the BUSINESS acumen. Learn the jargon, go to conventions, read blogs, view webinars and speak with as many people from your domain as you can. Every query that you pull and every number that you crunch tells a story, but the ability to see beyond the numbers and to link your data to “reality” is crucial for deriving actionable insights.
2. Keep it simple
   It ain’t no secret that in our day-2-day we use advanced statistical models and methods to analyze our data, but we must demystify these concepts into something that even a 5 year old will easily understand.
   Reduce the use of buzzwords, use business terms instead of technical terms and always back up your findings with examples.
   The ability to explain yourself to people from different levels & domains is more of an art than a science, but this type of art is a must for us.
3. Be proactive, and leave your shyness at the door
   It’s almost impossible to find an answer without asking the question. If you want to be a great analyst, never fear asking the tough questions. Send a message to the VP/C-level, even if that executive doesn’t know you, you’ll be amazed how a short talk (or even one Slack message) can help you focus.

Final note:

Some say that “Without data, you’re just a person with an opinion”.

Let me rephrase it and say that ”Without an opinion, you’re just another analyst with data”.