Category: bi

Previous entries in Everything Analytics:

The Many Wandering Paths to Analytics
Landing Your First Analytics Job

Confusing Web of Titles

The analytics space is rapidly growing & evolving, and this fast growth has led to a convoluted web of job titles which are overlapping and contradictory. I’m here to help you sort through some of those job titles and even open your eyes to different types of Analytics job you may not have considered. Data Scientists get all the press – but there are many more roles out there which may be a better fit for your interest & skills.

You will rarely find consistency from company to company. In fact, I’ll start with a couple disclaimers:

Disclaimer 1: Keep in mind that my opinion on the separation between these jobs has no bearing on how the HR department of a company defines their positions.

Disclaimer 2: This list is not exhaustive. There are lots of substructures to these roles as well as other data-adjacent or niche jobs which exist.

Keep In Mind When Applying

Make sure you absolutely understand the job description and ask many clarifying questions during interview rounds to fully understand what you’ll be doing. If you aren’t thorough in evaluating the job, you may not end up with the work you thought you’d be doing.

Example Job Titles

Data Scientist
(Related: Statistician)
Data Analyst
(Related: BI Analyst)
Data Engineer
(Related: BI Architect, BI Engineer)
Business Analyst
(Related: Technical Project Manager)
Machine Learning Engineer
(Related: Software Engineer)

Reporting Structure

While there is no one-size-fits-all structure, there are general trends:

Data Scientist/Data Analyst/Business Analyst

These roles may report to any part of the business, depending on how centralized the data organization is. The more centralized, the more likely they are on the same team. Sometimes they may be their own team entirely, rolling to the CEO independent of any other C-Suite leader. Other times they may roll up through the COO or CTO.

Other times, they may be decentralized and be scattered across the company with no specific structure.

Data Engineer / Machine Learning Engineer

Typically these fall under the CTO. Data Engineers may be under IT, or may be their own division. MLEs typically fall into Software Engineering — see below for more discussion.

Detailed Breakdowns

Data Scientist

Overview: This is the most-publicized job title out there and therefore is the broadest; it can mean many things at many places.

Data scientists are forward-looking and focus on predictive analytics. They certainly can do descriptive analytics, but their value comes from modeling/classification/etc.

Due to the emphasis on modeling, data scientists typically have advanced degrees in statistics, applied math, information science, or similar.

Example task: Predict how much stock of each item a company should order from its manufacturers in advance of the holiday season.

Data Analyst

Overview: While Data Scientists are generally forward-looking, data analysts are generally backward-looking and more entrenched in the business. Their job is to help the business understand what has happened up to this point and provide data in a clear & concise way for decision making in the future.

Typically data analysts focus heavily on making visualizations and presentations for the business and bridge the gap between the business and the data.

Data Analysts are also more jack-of-all-trades. It’s common to do a bit of data science, analytics, engineering, and PMing in a single role.

Example task: Create a flexible Tableau dashboard for leadership to track trial conversion to paid users over time

Data Engineer

Overview: Data engineers work on the databases that the other members of the analytics org use to get information to stakeholders. They are responsible for bringing data from the business into some form of data warehouse in an accurate, timely and secure fashion.

This means the typical customers of data engineers are the data analysts/scientists at the company. They also may work directly with different parts of the business as they want their own data ingested automatically into the larger data warehouse.

This role is typically more technical and code-heavy in order to move massive amounts of data around at scale. There is less interaction with the business than other parts of the analytics organization.

Example task: Mirror Salesforce data in a schema in Snowflake, updated every 5 minutes, for analysts & scientists to analyze/visualize

Business Analyst

Overview: Business Analysts are sometimes called a “project/product/program manager”, or PM. No matter the name, they are distinct from Data Analysts/Scientists in an important way. They coordinate and organize data projects across the business.

This role typically doesn’t exist early on in a data team’s existence. Usually individual analysts/scientists take this on until the burden of project managing starts outweighing time spent actually doing analysis. Eventually, the role of Business Analyst comes along.

Business Analysts are not expected to code or be as savvy on the technical side. Rather, their job is to identify problems, gather requirements, allocate resources and coordinate expectations between the data team and the business. This is no small task as many technically minded individuals are great at doing an analysis when there’s a clear question, but struggle to work with non-technical individuals across the organization.

Example task: Sales wants standardized KPI dashboards across their worldwide teams available for next quarter’s SKO

Machine Learning (ML) Engineer

Overview: This is a bonus position added in, largely since I see Machine Learning discussed commonly on Analytics forums and many of you may be wondering how it fits into a data org. The short answer: this role doesn’t fit into the data org per se.

Specifically, this role is commonly found on the Software Engineering (SWE) org and is more of a Software Engineer with an ML focus than anything else. This role is most similar to a data scientist and usually is more involved with implementing models within the production code of the company to solve whatever problem has been identified.

Example task: Predict which users on the website may want to know about Feature X, which will prompt an informational pop-up

In Conclusion

There are all sorts of roles to explore and this list is by no means exhaustive. As I mentioned at the start, the names above may be conflated with each other at any given job you apply to. Regardless, this gives you some guardrails around what sorts of roles are out there — from non-technical to technical and everything in between.

Entry Level Position – requires 5 years experience

-Every analytics job posting

This is Part 2 of the Everything Analytics series. Find Part 1 here.

Too few applicants with experience

As I mentioned in The Wandering Path to an Analytics Career, there is a ‘Great Filter’ in Analytics. It looks something like this:

Lots of people want to break into an analytics or data science career, yet not many are able to. This leaves a glut of competition for entry level positions, and not enough qualified applicants for mid-level to senior positions. Once you get your first few years of experience, you’re golden! You have your pick of many options within the data world – but you have to get past the Great Filter.

This rings true for me in an anecdotal sense – I have experienced this as a job seeker, interviewer and in discussions with data hopefuls. Given this is a blog devoted to data, I wanted to quantify the interest in a analytics positions just posted on LinkedIn. Unsurprisingly, you’re swimming upstream if you’re just blanket applying to analyst jobs – dozens to hundreds of applicants within a day or two of posting. See below:

For a fantastic & further in-depth analysis, I highly recommend reading the “Glut of New Data Scientists” section of this blog by Vicki Boykis.

Applicants Focus on the Wrong Things

As I’ve combed through resumes, cover letters and LinkedIn messages for the past five years, I’ve noticed applicants consistently missing the mark on what will set them apart. They consistently point to technical ability:

                Technical skills (SQL, Python, R)

                Mathematical skills (Statistics, algorithms, modelling)

                Certifications (Data Science Bootcamp, vendor-specific courses)

Those things are all great, but they don’t differentiate you from the pack. Everyone has some nominal experience in these things, you likely don’t have experience in all the tools the company needs (What if they use Looker instead of Tableau?) and even if you didn’t much of this can be taught on the job.

When I interviewed at my current position, I wasn’t asked one technical question. When the Director of Analytics stopped to see if I had any questions, my first one was “Why aren’t you asking me any technical/SQL questions?”  I’ll never forget his response: “If you’re missing any technical skills, we’ll teach you.” Wow.

This seems counter-intuitive. Isn’t data analytics/data science more technical? Don’t you have to code?  Of course you do! But those aren’t the most sought after skills; they’re a means to an end.

What Top Applicants Demonstrate

Analysts that shine on applications and interviews show they can persuasively communicate complex ideas using data. The job of a data analysts is to work with a stakeholder to generate business value. That doesn’t happen through coding – that happens through understanding the business, understanding the problem (even if it’s not directly stated!), breaking that complex problem down and communicating what the data says to do. Technical ability is solely leveraged to get there.

This is in the realm of “soft skills” is learned quickly on-the-job and is tougher to gauge for someone with no experience. How effectively can you work with non-technical stakeholders? Will people like working with you? Can you distill an ambiguous question into an actionable insight?

Top applicants can point to experience showing they can handle these scenarios and that’s why they rise to the top.

Tough to Teach in Classes

Classes are unfortunately a poor place to learn and/or demonstrate critical analytics soft skills. Teachers ask you very precise questions and give you very precise datasets to see if you’ve understood explicit topics listed in the syllabus. In the real world, this isn’t how analytics works.

Sometimes you aren’t even told there’s a question. If you’re asked a question the person may mean an entirely different question. The data might not exist, or it might sort of exist, or you might need to make it yourself. Your presentations are to a potentially skeptical crowd who doesn’t care what methods you used to arrive at your conclusion.

You can see the pattern – it’s near impossible for a teacher to create this sort of ambiguous and dynamic setting in a classroom. Imagine not knowing what day a test was coming, or if there were even questions on the test, or if the questions on the test were the ones you were supposed to answer!

This stuff is learned on-the-job, hence the need for experience.

OK OK, I get it. What do I do?

Now we’re to the crux of the matter – is there anything to give yourself better odds at landing that first job?

Yes.  There’s one overarching tactic, with three options you can do today to gain experience that will make a difference in an application.

Option 0 – Network, network, network!

This applies to all three other tips. If you’re throwing your resume into the ether of hundreds of applicants, you’ll find less success than networking with the leverage of the tips below

Option 1 – Start doing analytics in your current job

This is where many of us (including me!) started. You know how there aren’t enough good analysts out there? Take advantage! That means your company needs data people. Your boss, or some other boss needs help. Discuss their data issues and see if you can take something solvable. Don’t overcomplicate this. Use Excel to start, it’s a great place to iterate and extremely flexible. Move from there – find a pain point someone has with data and try to solve it. Start small and build. This is phenomenally effective, and you can point to this experience when applying to jobs later on (or move to a data position at your current place!).

Option 2 – Work on a data project you are passionate about

 I see tips everywhere saying “take on a personal data project” but rarely see much helpful advice beyond that. My top recommendation is to think of a hobby or interest you have and create an end-to-end analysis. Do you love a particular sport? Try to predict something that will happen. Have a favorite hobby? Think of a dashboard you could create to display your time spent/skill improvement. The more interested you are in the data, the further this will take you and the more time you’ll put into it. The options here are endless, but should be regarding something you love.

This gets you experience across the entire analytics pipeline – finding/cleaning/enriching data, asking good questions, visualization. Tableau Public is a great way to publish your results and iterate. The options are endless, and you can demonstrate skill and passion in an interview pointing to a portfolio of data projects. It doesn’t matter what tools you use, though my only recommendation is SQL and some sort of viz tool be involved.

Option 3 – Take online courses to brush up on technical ability

I’ve spent a significant amount of time saying that technical ability won’t separate you. That’s true, but you do still need some technical ability or you may be disregarded as not technical enough. If you don’t know SQL at all, you can take some basics online as part of doing Option 2. Do some basics on Tableau or Python or whatever strikes your fancy. This certainly is the least helpful option for standing out, but it also is a prerequisite if you lack technical ability. Typically if you’re doing Options 1 and 2, you’ll end up needing to do this option anyway.

In Conclusion

Breaking into analytics isn’t easy. But there are methods to get past the Great Filter and get your first job. It’ll take hard work and some luck and the goal is attainable. Companies need passionate and smart people to make sense of their data, and you can step into that role. Make it happen!

If we treated careers more like dating, nobody would settle down so quickly.

David Epstein
Range: Why Generalists Triumph in a Specialized World

I consistently receive the same questions from people seeking an Analytics career: What classes should I take? What certifications should I get? Should I learn SQL, Python or R?

Behind those questions there’s a consistent assumption: “There must be a clear path to an analytics career.”

I’m here to challenge that assumption. There isn’t one clear path to work in analytics – most of us got there through a winding, wandering series of career moves. My story is one of many – ask someone in Analytics and you’ll hear something similar.

Typical Wandering Path

(1) Get a college degree or other training – not super relevant
(2) Work for a while in some job non-analytics related
(3) Recognize interest in analytics
(4) Start doing basic analytics at work (ideally) or on own time
(5) Leverage that experience into first analytics job

I’ll call out each step as it happened in my career journey.

My Wandering Path

Initial Career (Years 1-4)

Coming out of college I shared the assumption that careers were linear. After all, life to that point was linear, so why wouldn’t careers be the same?

Except, my linear plans fell apart two days before my wedding in 2011. I’d studied International Economic Development (Step 1), interned in Latin America, become fluent in Spanish and was planning to move with my soon-to-be wife to Bolivia. In one phone call and several subsequent conversations, that potential life and career ended. I was sitting in a dead-end job I thought I’d be leaving and had to figure out Plan B.

At first it wasn’t obvious – what else should I do? I was a Customer Success Manager (Step 2) but didn’t really want to do that as a career. I’d worked in sales departments, but didn’t really want to be a salesperson. But then I had an epiphany – there was a part of each of my first few roles that I loved that never was part of my job description.

I was consistently making little analytics & reports (Step 4 – which ironically for me came before step 3!). I’d turn 2,000 customer emails into a digestible summary for the product team. I’d make Salesforce forecasts & dashboards for the executive team. I made a Google Sheet for my Rosetta Stone team to help management track & manage renewal rates for their teams. This stuff was fun! I liked it! (Step 3) But what now?

The Great Filter: Landing First Data Job

Have you heard of the concept of the ‘Great Filter’? It’s part of the Fermi Paradox, which ponders why there is no extraterrestrial life given the seeming high probability it should exist in the universe. Within the Fermi Paradox, it’s the step getting from non-living matter to living matter (abiogenesis). The Great Filter is a catchall for “it’s hard to get past this point.”

I argue there is a Great Filter for those trying to get into Analytics – getting your first job. In fact, I’m devoting my next blog to this topic, so consider this a lead in to next week.

Passing Through the Great Filter

I realized I had an uphill climb ahead. Perhaps this is where many of you are – how do you get a company to take a chance on you?

I asked lots of current analysts via informational interviews at local coffee shops. They all said “I got here via a pretty random series of events.” Sound familiar?

They gave me a breadcrumb trail, though: “You have to get enough experience together and communicate about it well enough to get someone to try you out.” Easier said than done, but I did have some experience already in my current role.

I applied everywhere. I was told by recruiters/HR multiple times “Hey, I guess you could be an analyst but I think your future is in sales based on your resume.” Leads fizzled out until one day I got a call.

The Meeting That Changed Everything

“Can you show up at the office in an hour? The CFO and SVP of Sales want to talk.”

I got that call from Jacob — the same Jacob here at DataDuel. He was working at Funko, a quickly-growing collectibles company north of Seattle. They didn’t have a position open yet, but there was interest in getting analytics going. Before going into the meeting, here is all I knew:

• The position was planned to be part-time Analytics and part-time something else until analytics skills were proven
• The position was planned to be a contract position (not great – I’d just bought my first house and wasn’t looking for a contract spot)
• There would be minimal support since no data team existed, so a self-starter attitude was needed

Given those three bullet points, I had three goals going in:

• Communicate my potential to be great at analytics if given the chance
• Sell them that I was worth a shot as an employee & not a contractor
• Demonstrate self-starter attitude to analytics from previous roles

I quickly threw on a dress shirt, re-learned how to tie a tie on YouTube and flew out to the car.

The rest is history and full of the content I’ll fill this weekly series with. The conversation went really well and they decided to take a chance on me a couple weeks later (Step 5!). While I got a full-time position, I took a 10% pay cut because I needed to prove myself. I knew the temporary sacrifice would be worth it – I just needed my first position to get past the Great Filter.

In Conclusion

There’s no one path to analytics – there are many. I’ve used my path as an anecdote for the infinite options out there.

The general path, though, is to start doing some analytics in any fashion you can, and leverage that experience to get your first position. It isn’t easy – there’s a Great Filter out there which prevents many from getting in.

Come back next week and I’ll dive into the Analytics Great Filter in more detail, and provide some practical options of how to overcome it.