Is your data stack really that important?
Expensive bootcamps claim to make you a data scientist in a few weeks. Is your data stack really that important?
At the end of almost every interview, candidates ask me, “What tools are you using?”.
With the proliferation of very expensive bootcamp style courses claiming to turn you into a data scientist in just a few short weeks, there’s a perception that to be successful in the field of analytics you’d need the right stack under your belt.
The truth is, I don’t care about your data stack.
In all honesty, I’ve named some software in a job description because that’s what we happen to use. And life is messy: that stack may not be implemented across the organization consistently. Or we may have a hybrid approach or some manual interventions, or, or, or…
We (society?) are so hung up on the names of technology being used when there are fundamentals to analytics that are far more important than the programs you have on your resume. Why? Well, it’s easier to package a bunch of software training into a 12 week program and call it a “data science” bootcamp. It’s a lot harder to train the masses on the necessary soft skills you’d need to actually differentiate yourself in this field. And it will take more than 12 weeks of 12 hour days to get there.
Remember, the term “data scientist” isn’t an officially defined occupation like “M.D.” or “attorney” and it’s up to the institution training you to define what exactly they’re offering. Be aware that data science is a very buzzy term today so when you’re shopping for the right education for you, look beyond the flash and into the meat of the offering.
Of course, a level of technical expertise is very important in this space. You will be required to have some technical knowledge. In today’s analytics environment you can’t compete without it. More important than the specific technology is the ability (and willingness) to learn a new software or stack. And software evolves, technologies come in and out of favour. This dynamism can render the question, “What’s your stack?” useless.
Ok, so now you know I don’t think your data stack is really that important. By now you know I’m more interested in your soft skills and the other ethical and leadership questions troubling our field. I also have two areas of focus that I think every analyst should consider:
1. Your capability to pick up on the unique issues facing the business.
In the field, we aren’t analyzing data as a thought experiment or a word problem on an exam. Data analysis doesn’t happen in a vacuum. Our practice adds value by understanding our business and anticipating how data could be used to help solve problems or optimize the operation.
In order to give the business good insights using their data, we ought to understand the data as it relates to the real world. Even if you’re the most junior member of my team scrubbing data, you need to know about how actions in the physical world turn into data.
We make widgets? Great, you’re heading down to the floor to talk to the people making them. Are we moving people using trains? Awesome, let’s get you walking around our network to learn from the teams making it happen. Before you can get to the nitty gritty of the data, you need to know what the heck we do.
When we are out there, it will feel a little irrelevant and maybe like we’re wasting our time. In my experience, that means we’re on the right track.
There’s plenty of time for data teams to sit in an office fiddling with models. Before you get there, you have to understand – or better, empathize – with the folks generating that data. You build credibility by showing up, and you can recall the operation more readily as you work through your models later on. And you’ll have someone in the field you can call to kick the tires on your initial insights. I call it Sneaky Stakeholder Engagement (trademark pending, haha!).
We need analysts who are willing and able to rapidly connect business to data. I want you physically getting out there and observing and learning in the field as much as possible. Not another staid meeting gathering requirements or reading a binder of SOPs. That fun stuff can come later. First, get out there and start building those relationships now. You’ll ask better questions during those requirements gathering meetings, I promise.
Do it IRL! Here are some tactics you can try to start learning more about the business directly from the people running it:
Pick a data owner who has been open and warm to your work in the past. (Make it easier on yourself, don’t start with someone who doesn’t believe in your work). Ask them to spend an hour in person with you walking you through their operation.
If working arrangements allow, schedule meetings as close to the source of the data as possible – it could be in an office near the shop floor, a lunch room, a safe location on site. Pick up your laptop and go!
Always reference the help you received from your data owners in any presentation you make. You want to emphasize the partnership between the business and the analytics practice.
2. Your ability to communicate technical items to non-technical folks
Here’s a scenario I’ve come across pretty frequently:
Interviewer: How do you modify your analytics presentation for non-technical audiences?
Candidate: I would modify my presentation by using less technical language and more graphic elements.
Let’s call this strategy out for what it is: a nice way to say we are dumbing down the content. You may as well say you’re writing the deck in crayon.
Just because someone isn’t proficient in your analytics tools, doesn’t mean that they can’t become experts in how data can be used to better their operation.
This is where our field trips above come in handy. Assuming we’ve done our homework, you’ve spent a bunch of time with the beating heart of our business, coming to know the people behind the data and their pain points. You learn empathy. You learn how your audience communicates. You’ll better be able to craft a message to them that takes them through their data journey in a respectful and meaningful way.
Dumbing down your presentation is lazy and borderline offensive. People of all walks of life have the capacity to learn about highly technical subjects without engaging in the finest details of said subject. Case in point: in December 2019, no one except for epidemiology enthusiasts knew what spike proteins, contact tracing, and mRNA vaccines were: now, these terms are part of our vocabulary.
Your job as a data professional is to take the audience through the journey of their data. Don’t be afraid to introduce new vocabulary and concepts to people – as experts in their own operation, you may learn a thing or two from your audience as they become a partner in the data journey.
Do it IRL! Here are some tactics you can try to start more effectively communicating your analysis to non-data folks:
Resist the urge to stuff your presentation with technobabble and jargon. It can be tempting to include a lot of technical detail, especially if you’re nervous and intimidated by your audience. There are a lot of reasons why more jargon is a bad idea, but in this context, it can really alienate people and create artificial distance between you and your audience. Keep the language straightforward and conversational.
Before, during, and after preparing your presentation, reflect on the business question that your audience is trying to answer. Give your presentation an honest assessment. Did you answer the business question with a business answer, or did you answer the business question with stats and charts? Will the audience be able to use this information to make a decision or do they need added translation?
Practice the presentation with a non-technical audience. Who your practice audience is depends on the context but can be a friendly member of the real audience, a provider of source data (who you’ve built up a great relationship with by learning their business), or another employee unfamiliar with your presentation.
So, is your data stack really that important? No!
It’s so important that you have some technical know-how in the field of analytics. We’re not performing analysis using an abacus and a slide rule, after all. But is it the be-all end-all? No! Your data stack is going to evolve over time as your career progresses. Technology changes and comes in and out of favour. The field of data analytics and data science are so much more than the specific software you have on your resume.
Even more important are other fundamental skills that I believe all good analysts should have: to quickly pick up business knowledge and communicate respectfully with non-technical folks. These are harder skills to teach and require deliberate effort and time to develop. It’s worth it.
Photo by Kevin Canlas on Unsplash , Isaac Smith on Unsplash , Carlos Muza on Unsplash
If you enjoyed this article, watch this space! I’m going to be posting more of my thoughts on developing soft skillsets for well-rounded analysts, along with practical tools that can help you in your day-to-day life.
