Statistical Consulting: Where Words and Numbers Meet

If you’re reading this, then chances are you’ve been on sites like Forbes lately and I can guarantee you’ve seen something about how data science is the hottest new buzz word in the business world. With a quick Google search, you’ll see a six-figure median salary. Also, there will be countless sites offering to train you with just one course! Although this may sound too good to be true, I’m here to tell you it may not be as far-fetched as you once believed. We’ll walk through the steps to begin your transition from the office desk to how to become a statistical consultant. We’ll focus on a broad idea of data science and statistical consulting, then break down the sub skills. You’ll need to remember the key to success is an entrepreneurial mindset.

Statistical Consulting and Data Science…so like graphs and stuff?

With a bit of research into this growing field of part time statistical consulting, it’s likely you’ve come across some vague explanations related to words like

Concept about machine learning to improve artificial intelligence ability for predictions

Don’t get me wrong, those words are a big part of data analytics. However, to start getting work you don’t need to have all the qualifications to land a senior position. All statistical analytics boil down to the ability to take a whole lot of data, be it numbers or words, and find the story that it is telling. Now, if I were writing a business pitch, I wouldn’t quite say it that way, my proposals tend to include something like

Rendered Image

Sounds fancy right? That’s just it though, when working as a statistical consultant, clients are looking for the “what to do” not the “why to do.” In the end, you need to remember they hired you to understand the data and the story it tells. When reporting to clients the goal is to provide actionable insights – based on the story the data tells.

How I Built My Company

I’m sure at some point everyone reading this asked their math teacher “When am I ever going to use this?” (disclaimer: I’m that math teacher who provides way too much detail in my answers to this question). I’m still lucky enough to be able to teach in the high school mathematics setting, mentor new staff, and work with our school’s data. However, one day after hearing that question for the hundredth time, I thought to myself

“Why aren’t I actually applying more of the mathematics I’ve spent so much time teaching?

I then started researching what it would take to start my own company and “practice what I preach.” If you’ve had jobs looking at data/graphs and drawing conclusions, then you’ve already got the basis of the skills and practical experience you’ll need to start consulting (hint: you don’t necessarily need a degree in statistics).

How Do I Start? Do I need an LLC?

First piece of advice, anybody who provides “guidance” when it comes to numbers and money is in a situation of liability. Statistical analysts, in-house or not, are responsible for collecting data and market research that leads to big financial decisions. This means that should an occasion arise where decisions are made based on your “advice” and it doesn’t go as planned then you are liable to be sued. By creating an LLC or other business entity you help to separate the personal “you” and all the assets “you” have from the statistical analysis business you’ll be creating. I run my business from home, no point in renting an office to do what I can do in my b-office (basement office), right? The process varies a bit state by state, I live in Illinois and from start to finish it took me a few hours to get everything submitted. I recommend for anyone looking to get into consulting to do this since the investment in an LLC or similar business entity creation isn’t terrible and it helps protect your personal life.

Okay So I Run a Business Now…What Services Does It Offer?

Great question! So, this is where we dive into what the world of data science and statistical consulting looks like. I’ll say in my experience whatever you think you’ll be spending most of your time doing, you’re probably wrong. Data analysts do a lot more data than analysis. If you want to be a good analytical consultant, you’re gonna eed some knowledge of software programs that can help you out. This is where research comes into the picture. There are many programming languages and statistical programs out there, I have yet to meet anyone that has mastered more than a handful. Why? The fact is that what you can do in one language or software is likely reproducible in many others. For example, if you’ve been looking into data science chances are you’re under the impression you need to make a choice about studying R or Python (here’s a great resource for R programming).

raster round icon of web shield with R letter programming language - isolated flat design illustration long shadow
Vector illustration of an icon of the Python programming language. Logo in the form of two snakes. Flat icon on white background

Personally, I use both, along with Excel and the occasional Tableau workbook. Wait, I need to learn all four just to get started? Of course not! Excel is something most people working in the business world are familiar with. Tableau is essentially the super user version of Excel. Python/R are programming languages with great pre-built packages for work in data science. My suggestion would be for anyone looking to get into the industry is first become familiar with Excel. Then choose either Python or R as a starting point for your coding journey. 

There are a ton of great and affordable courses online on sites like Udemy that can help you master a variety of skills without the outrageous costs of graduate degrees. Now, that you’ve chosen which language you’ll be working on, it’s time to figure out what type of services you’ll want to offer. My suggestion is that you’ll want to check out sites like:

These are sites that offer a ton of postings and job alerts for freelance work in the field of statistical consulting and analytics. This is where you can read through a ton of job postings and look for the common words, phrases and services that companies are looking for.

Jack of All Trades or Subject Matter Expert

This is where it gets challenging, when you’re researching jobs, you’ll see a ton that want a very high level of knowledge about a specific piece of software. Yet everyone seems to be talking about a different software. When I first started this was pretty discouraging. I felt like I’d need to spend five years of lonely nights preparing to complete jobs that might not even be around next week! This is where the resiliency of the entrepreneurial spirit comes into play.

I realized that my brand-new profile on UpWork may not inspire confidence in someone advertising a $2000 one time gig involving developing a predictive model for textual user input. However, I bet my current work samples would convince people I could make a mean bar graph and handle cleaning up data. Full disclosure, when I started this journey, I was finishing up my Master of Science in Data Science. As you’ve probably guessed that very costly education left me with about 10% of the skills and statistical techniques I would actually need to make it to where I am today. 

If I could go back in time, I would have spent a lot more time using much less expensive resources. Which are available on sites like Udemy as opposed to buying them anyway to supplement the gaps in knowledge graduate school lectures left me with. So, circling back to whether you need to be a jack of all trades or a subject matter expert to start consulting in statistics. The truth is you’ll probably follow a path like this:

Novice in a few things–>Learn about many things–>Subject Matter Expert in Something

Focusing Your Studies and Skill Building

Okay so off the top of my head here’s a noncomprehensive list of the various fields within the broader scope of data science and applied statistical consulting that I currently work with:

Data CleaningData AugmentationData Analysis
VisualizationNLPCloud Architecture
Predictive AnalyticsMachine LearningPsychometrics
App DevelopmentData JournalismNumeric Sorcery/Witchcraft

The last one is how I explain to my children what my business does for people, they easily understand my career as a teacher but the whole image of daddy typing on a computer and using phrases like “multivariate statistical modeling” on phone calls remains a bit confusing for them. Hence, I explained my idea of numeric sorcery/witchcraft as a way of taking a bunch of seemingly random numbers and words and using them to create an evidence based meaningful story that helps my clients accomplish what they need to get done. As you can tell my bedtime stories are top notch! 

The point is that there are countless subfields within the job title of data scientist and statistical consultant. If you’re looking to build your skillset and find areas of expertise, my advice would be to start from the bottom up. No matter what branch you want to specialize in, you’ll never avoid the base level skills of:

Data CleaningWhen it comes to messy data, trust me there are companies out there running legacy software that creates CSV files which no normal human can gain insight from without cleaning them up first.
Basic VisualizationsBasic visualization creation skills are key, not only for delivering early proof of concept results to clients, but also for yourself as you begin to work on more complex modeling and machine learning problems. 
CorrelationLooking for correlation when it comes to determining which variables have a relationship that is statistically significant, that’s a concept we teach in high school yet never loses importance in the world of business and consulting.

Whenever I train and evaluate predictive models, I use the same box plot visualization code snippet I wrote forever ago because it helps to me quickly understand which models can be ruled out before I waste my time hyper tuning something with only 20% base line predictive accuracy. Once you feel confident in these three general skills, then it’s time to start working on specialties. Those are the phrases I listed above, areas of skill that set you apart from your bidding competitors. Do some research on them and others and start working on some pet projects to showcase your skills. One of my first major clients I contracted with was looking for a general data scientist with expertise in public health or psychometrics. They needed to validate the tests their company offered. It turned out my early interest in psychology paid off, as I knew that psychometrics was the mathematical extension of statistics applied to scorable tests and competency measures.

Are you interested in sports? A friend of mine made his entire “pitch portfolio” based on a program predicting who would win the Super Bowl. It could analyze data and make predictions by looking at teams, individual players, weather patterns, you name it! The cool part about data is that no matter what your interests are, the tools for analysis are similar!

Creating Your Elevator Pitch – Specificity with Efficiency

A skilled statistical analysis consultant makes a pitch to a potential client for good consulting freelance work in an elevator

I tend to think of every attempt at new client acquisition like the classic case of an “elevator pitch.”

When you’re utilizing a site like UpWork to submit proposals there are a few strategies that have helped me to find success. These are what we’ll conclude this article focusing on. There are two parts to landing clients:

We’ll start with your profile; this is where you want to showcase and link to projects you’ve done in the past. These don’t have to be projects you’ve been contracted for on UpWork or even paid to complete. Working through online tutorials helps you build examples of your skills for clients. They make for a great talking piece in pitching your skills on a new project. Your biography should be at most one or two paragraphs. It should focus on highlighting your experience and the skillset you bring to the table. If you’ve got a LinkedIn account this is a great place to link to on your profile as well. It will help to add credibility to yourself as a consultant. Lastly, there’s the ability to include a video introduction which I’ve found beneficial. It puts a voice to the skills you’re advertising and gives an added layer of authenticity.

A great data analyst has a great statistics and data science profile to help land great analysis and consulting jobs

Now, when it comes to creating your pitch, we’ll focus on specificity with efficiency. In a perfect world, you’d have the chance to spend hours preparing and putting together each pitch and analysis plan. However, if you’re like me and still working in another career, you need to establish an efficient system – a template if you will. When a potential client sees your pitch, you’re likely not the only one, and hence we want this to be direct and yet informative. Mine is broken down into three short paragraphs.

Paragraph OneThis is where you thank the client for considering your proposal, and directly reference the goals of their project so that they can see you’ve clearly read their posting in detail. If you’ve completed something similar this is a great place to mention that work briefly to further their interest in your proposal.
Paragraph TwoHere you provide a condensed version of your work experience and qualifications. You can always include your formal resume as an attachment when submitting a proposal, but this paragraph should highlight the things that make you unique as well as specific skills that relate to this project.
Paragraph ThreeLastly, end with some reassuring words that you’re open to collaboration and want to ensure the deliverable you provide not only meets but exceeds their expectations. Mention that you’d be happy to further elaborate on your qualifications and plans for this project through a phone interview.

Brass Tacks- Bidding on Jobs in the Freelance Marketplace

By this point you feel strongly about your skill set and you’ve found a few job postings on UpWork that you know you can handle – what do I bid? There are two types of job postings on this site. The first type offers a fixed price for project completion, the second calls for you to bid an hourly rate. Fixed price projects seem to go one of two ways, either you get lucky and finish in half the time you thought ($$$) or it might take three times as long as predicted. When bidding on fixed price projects you want to be sure to ask questions to know exactly what you’re getting into. After all, when you are advertising your services the last thing you want is a bad review because you “glorified” the skills you had to land a job.

When bidding on hourly jobs, be realistic and do your research. There may be another “data scientist” freelancer profile you seen saying they work for $150 an hour. However, chances are they are a bit further down the road than you are in terms of reviews and actual projects completed, and probably work full time. I hate to break it to you, but chances are the hourly rate you’re going to offer when you first get started is much lower than what your skillset deserves. Within a month of starting my business, my first two consulting contracts that I signed were for $30 an hour. One was working with medical billing visualizations and the other was for reading over an education thesis. 

Soon thereafter I worked on some fixed price projects, one averaged out to about $50 an hour while a few were in the neighborhood of $20 an hour. However, each time I completed a project my profile looked a bit more impressive. Within a year of starting my company I now offer a base rate between $50-$60 an hour. Although I do much of my proposals through additional sites such as LinkedIn and contract directly with my LLC, that’s not to say I won’t bid for jobs on UpWork that pay $25 an hour.

For example, as someone still spending time teaching in a classroom, I genuinely enjoy mentoring people looking to get into the applied mathematics and statistical fields. If I see a posting about someone wanting their dissertation read over or a student looking for help, I always throw in a fairly priced proposal.

Looking for more high paying gigs? Check out these awesome articles:

This article was contributed by John Ciolkosz, of C Solutions LLC