Why Should You Freelance?
Freelancing is becoming more and more popular. According to Upwork’s 2019 Freelancing in America Survey, 35% of the US workforce freelances at least part-time. And it’s not just part-timers either. The full-time freelancing crowd has increased from 17% in 2014 to 28% in 2019. Freelance data scientists are only a small part of this crowd, but they are also a growing trend.
But popularity is hardly a good reason to change your working style. Here are some better ones:
- Make some supplemental income – When you are salaried, a 40-hour workweek and a 50-hour workweek generally pay the same. If there is an advantage to your career in working those extra 10 hours for your company, go for it. Otherwise, 10 hours of freelancing is a good way to rake in potentially thousands of extra dollars a week.
- Temporary support during a job transition – Nobody wants a gap in their resume. More importantly, nobody wants a gap in their bank account. Freelancing is a good way to pad both of those during a transition. Just keep in mind that the freelance gigs may not come in immediately, so it helps to start reaching out to potential clients in advance of when you need them.
- Build a career change portfolio – Employers are all about managing risk in the hiring process. Hiring someone for a full-time job is inherently riskier than hiring someone for a 2-week freelance project. Once you’ve proven yourself in the field, you can always go for the full-time gig if you still want to.
- Dip your toes into entrepreneurship – Are you not sure whether you want to venture out on your own? Freelancing is a great way to try out entrepreneurship in a quicker and less involved way than most other methods. You’ll also be racking up connections and industry understanding that will benefit in any future business ventures.
- Just become a full-time freelancer – I love being a freelancer. I wasn’t sure that I would, but it’s been the best experience of my life. Is that just Stockholm Syndrome? Hopefully not, but I’ll let you know.
Is Data Science a Good Fit for Freelancing?
Short answer: yes.
The supply of good data scientists still hasn’t caught up with demand. This makes it much easier to find good job opportunities, both in full-time and freelance work.
Many companies have the resources to hire a full-time data scientist, but there are just as many others that don’t.
Even amongst those that have the resources to hire one, many companies don’t need data scientists, even if they think they do.
Some of these companies need data engineers or data analysts instead of data scientists. Others would benefit from a data scientist but don’t need one for 40+ hours per week.
For those companies, they can either find someone internally that can handle some basic data science alongside their other responsibilities or they can find a part-time or freelance data scientist. Since data science knowledge is not widespread, the latter option is almost always easier.
Unlike some other technical jobs, several parts of data science are hard to commoditize. The complex nature and heavy communication needs (see the next section) also make it hard to offshore most data science.
All of this spells out great potential, especially if you are US-based and looking to do some freelance data science.
What Do You Need to Get Started?
If you have experience as a data scientist, you are already in a pretty great place.
If not, you’ll need to spend some time becoming a data scientist.
Especially if you have no experience in data science, you want to niche down like crazy.
Do you have experience in a particular industry? That’s a great thing to include in your niche.
Do you have experience with data visualizations, forecasting, or revenue models? Each of those can become part of a niche.
I know this seems counterintuitive, but having a more narrow niche will actually make it easier to find freelance jobs. Moreover, it’s easier to become an expert in a smaller niche, so your prices will rise faster.
Here are some example niches that are at a good granularity.
- Revenue modeling for SaaS businesses
- Demographic analysis for eCommerce stores
- Geographic data visualizations in Tableau
Keep in mind that you aren’t stuck in a niche forever. I’ve written in the past about how to break out of your specialization, and I’m still a firm believer that hyperspecialization is bad for your future.
If you can’t find a niche right away, that’s also not that big of a deal. You can always pick a slightly broader niche at first and narrow it further as you find what jobs you enjoy and can get hired for.
You may also consider defining your niche based on job listings that you find. If you see several listings going unanswered for waste management demand forecasting, maybe that’s a niche worth digging into.
Once you’ve picked your niche, make sure you have the skills necessary for it. One of the biggest differences between joining a data science team and doing freelancing is the onboarding process – or lack thereof.
Your client will likely not have data science experience. Often, they won’t have statistics or programming knowledge either, but they will likely know their domain. You’ll have to bring the rest.
When a client hires a writer, they can tell whether the result appears correct. When they hire a programmer for a mobile app, they can try out the app to make sure it works.
However, they may not have the level of knowledge necessary to tell whether your data science project has been done correctly or to understand the results.
Because of the knowledge gap between client and freelancer, being technically proficient is not enough. You need to become a master storyteller – able to convey information in a way that is both compelling and informative.
That being said, don’t even attempt to be disingenuous with your results. Nothing will ruin your credibility faster than being caught trying to take advantage of a seemingly ignorant client.
You may have already caught this from the past few paragraphs, but you are probably going to be a bit of a data consultant here and there as a data scientist freelancer. You can choose to be okay with that or just look for clients that have a better idea of what they need.
If you are new to data science or haven’t done the specific task in your desired niche before, you are also going to need a portfolio.
There are a couple of ways to build up that portfolio.
- Put the free in freelancer – Find companies in need and offer to do some work in exchange for testimonials and inclusion in your portfolio. This option works really well for freelance writers, but it can be tougher for data science as people often believe that free work is low-quality work. If you emphasize that the testimonials do provide value to you, you can potentially overcome this belief.
- Build your own projects – There are quite a few public datasets that you can use for this. Basically, you need need to perform the task as if you were doing it for the client. Then create whatever visualizations, slides, or other documents would make sense to demonstrate your competency.
You’ve got your niche, you’ve learned the appropriate skills, and you have your portfolio. Let’s find you some clients.
Where Can You Find Freelance Data Science Jobs?
There are several options here. I’m going to break them into two major categories: freelance platforms and finding your own clients.
The great thing about these websites is that clients come there looking for one thing: freelancers.
Thus, getting the job just means convincing the client that you are the best freelancer for the job.
If it sounds too good to be true, that’s because it generally is. Since these clients are easy to find, everybody finds them.
You’ll be competing against a much larger pool of data scientist freelancers on these sites. This means you’ll really have to stand out.
Because the law of supply and demand still exists, it also means that you should expect a lower rate of pay. And that’s before the website takes their cut!
I do recommend dipping your toes in here, though. These sites are easy to set up, and they typically handle payment processing and the like – simplifying your first foray into freelancing.
These are some of the more popular sites for data science freelancing work:
- Upwork – They are the big player in freelance jobs for a reason. If you want the simplest possible starting point, this is a great option. I wrote an article about how to get your first writing job on Upwork. Data science is obviously not the same, but much of that advice still applies. Upwork lets you search by skills by keywords and skills. I find the latter option to be important for finding data science jobs in a given niche.
- Fiverr – Instead of you searching through client listings, Fiverr has you define set projects that clients can search for and purchase. This makes it all that much more important that you have a well-defined niche in advance to avoid creating too many different projects. There is a whole Fiverr data science section, so that’s a good place to start for figuring out what projects you can offer on their site.
- Toptal – Their entire schtick is that their freelancers are the top 3% in their field. You will have a harder time getting started with Toptal than with the other freelancing sites, but it’s a good one to aim for as you advance.
Find your own clients
If you are willing to work harder on finding clients and risk chasing down invoices, finding your own clients is the best way to maximize your income.
Obviously, it will not be as easy to find clients this way. You’ll need a more multi-pronged approach to keep the funnel filled.
Get your favorite spreadsheet program out. You’ll want to start jotting these down to keep track of which ones you have contacted.
Here are a few of the best sources to keep in the mix:
- Indeed – The first thing to do is check specifically for freelance gigs in your niche. Don’t be surprised if you don’t find anything that way, but it’s worth a look anyway. After that, look for any full-time vacancies in your niche. Write all of these in your spreadsheet. Your goal is going to be to find someone to contact in that department (preferably the hiring manager). You’d be surprised how many of these companies are amenable to using a freelancer instead of someone full-time. After all, they may be having a hard time finding a good candidate, and they may not have realized that freelance data scientists were even an option.
- LinkedIn – There are a few pathways for this one. First, it’s a job board just like Indeed. Second, you can check out company pages much like what I describe in the company websites bullet below. Third, you can beef up your profile to let people find you. Don’t forget to add a great LinkedIn headline!
- AngelList – They only do startups, but that is a great place to start until you can get some bigger name clients. You might even find that you just like working with the little guys.
- BuiltIn – They are specific to a select set of cities, but they have some great options if you live in one of those places.
- Company websites – Put together a list of companies that could be potential clients, and go check out their sites. The rest is pretty much the same as the job boards: check their career or job page. While you are there, see if you can find the email address of someone likely to be hiring for your niche (or at least someone that can put you in touch with that someone).
- Cold (or maybe warm) emails – By this point, you probably have some contact information. Reach out and see if they need your services. If you’ve never successfully used cold emails before, look up a guide for cold emails. Most people do them wrong. Don’t be one of those people.
- Let them come to you – Build your own website. Take that website, and add some SEO content in a blog to start ranking on Google. Paste that website on your LinkedIn, your other socials, and everywhere else you can. It’s a slow process, but getting people to come to you pays off in the long run.
Are you excited to do some freelancing?
I hope so because now you know how to become a freelance data scientist, so the next step is all you.
Go! Get out of here! Find your niche. Master your skillset. Build a portfolio. Get a few initial clients.
Like freelancing but not sure about data science? There are plenty of other freelance options out there. Spend 20 minutes browsing Upwork, and you’ll find all kinds of options.
Freelancing is competitive, but you’ll find a way to make your uniqueness work for you.
I love to hear success stories, so feel free to come back and let me know how it all goes. I’m looking forward to it!