Opinion
Whether you are looking for new career or just want to make some extra money on the side, Data Science freelancing can be very lucrative if you know what you are doing. Unfortunately, most aspiring freelancers fail to get off the ground. Below are some tips that were crucial to my success in this field and that I hope will apply to you as well.
1. Don’t get bogged down by logistics
Initially, I was overwhelmed by everything I "needed" to do. Finding a catchy name that wasn’t taken, building a website, designing a logo, getting a business license, and the list goes on and on. I finally decided to just jump in. Personally, I don’t love my business name or logo and my website is still not finished. One day I will fix these things, maybe, if my work ever slows down.
Honestly, I have seen very little return from these things. Clients rarely ever click the link. What they are really looking for is a solid proposal and proof of a good track record (both discussed below).
2. Choose a platform
When you are just getting started, don’t waste your time trying to find clients yourself. I spent weeks cold calling businesses, applying for job listings as a consultant, and reaching out to people in my network with very little success. I found much more success using a Freelancing platform. A simple google search will show there are many of them. I personally use Upwork. Most take a percentage of your earnings. It is worth it! Remember, your core competency is data science, not sales. Rates can range between 5–20% and usually drop as your put more time in. This can feel like a lot but it is a small price to pay for the time saved trying to find clients the old-fashioned way. I promise they are not paying me to write this by the way, it just really worked for me.
3. Create a solid proposal template
Having a solid proposal makes a difference. Alter it based on the specific project. There are many tutorials out there that can help with this. I personally like to lightly highlight my experience after talking about similar projects that I have worked on and the success I had with them. Sometimes I share very good reviews on those similar projects. If I have a previous project I can share, even better. Experiment to see what works for you.
Remember, this isn’t a resume. The only thing the client cares about is if you can do the job at a reasonable price. If you can prove that, they will pick you over someone else regardless of what school you went to.
Make sure to sell them on your value as well. When I was first getting started, I made sure to include something in my proposal about working at a major discount because I was new to the platform but had a long history of successful projects. Now, I will typically point out that I produce more accurate models with lower overall costs due to my experience and coding efficiency. Clients like to know they are getting a top-quality product at a good price.
4. Find your first client
Landing the first client is by far the hardest. It took me a few weeks and hundreds of proposals. I promise it gets easier! Once I had that first 5-star review, the second client took half as long. The third took half the time as the second. Fast-forward a year and I now have clients coming to me without me even looking. I actually get more work requests than I have time to work on.
Your first couple of clients are for building reviews, not making money. I started out only charging $50 an hour. With Upwork, you will get a "rising star" badge after a few successful projects. This helps because many clients are specifically looking for these bargain freelancers.
Occasionally you will find a client that gives you all the data in the job posting. On three occasions, I solved their problem and sent them the solution without being hired first. If you go this route, work fast! Many clients immediately go with the first qualified freelancer they talk with. It is definitely a risk, but I was awarded the contract on two of them which worked wonders in jumpstarting my profile.
Finally, if a client engages with you but decides to go with someone else, don’t be afraid to ask for feedback. I have received a lot of useful suggestions and critiques that helped me out in the long run.
5. Organically grow your rate
After your first few successful projects, start incrementally raising your rate. If you stop getting work, bring it down again. With the premium version of Upwork, you can see what the bid ranges are for a project. Seeing this information is worth every penny because you can see when you might be able to go higher than normal or if you will risk losing the contract (or shouldn’t even bother with it).
The real key for me turned out to be the profile rate. Once I had enough work coming in and didn’t have enough time to work on other projects, I would bump up my profile rate to slow down the requests. Think of it as a "what would they have to pay me to consider working way more hours than I want to right now?" To my surprise, it didn’t slow things down. Then I raised the rate even higher because I really didn’t have enough time to take on more work. And again, the work kept coming in… After a month or so of being VERY overworked I was able to settle on my equilibrium rate; the rate at which work consistently comes in at a pace that is comfortable to me. If I ever decided to do this full time, I might lower the rate to get a greater number of contracts. It’s all a balancing act.
One final note on this: freelancing work isn’t always (ever?) steady. I think it is due to many companies having a similar annual budget cycle. I have months when all my past clients come to me on the same week and months where I get very little interest. Since my projects usually span a few months it hasn’t affected me as much but it is something to be aware of as you adjust your rate.
6. Find (embrace) your niche
When I first got started, I wanted to do all things AI. The problem is, every data scientist wants those jobs. So my first several jobs where in a legacy skillset where I was superior to other freelancers: advanced Excel VBA modeling and forecasting. From there, I was able to branch out to Python ML models that would integrate cleanly into the client’s Excel tools. Now, I can land a pure (non-Excel) AI job, but they don’t pay as well as the Excel ones. Why? Because everyone wants to do AI, but very few can integrate AI into Excel. Would I prefer to work completely within Python? Yes! Enough to drop my rate? Nah.
Now I’m not saying that you should master Excel and do what I am doing. Please don’t! It would make me very unhappy if everyone reading this decided to crowd me out. What I am saying is that you should find something you can do that is unique. For example, Tableau added the ability to integrate with Python. If you are an expert with Tableau, embrace that. Maybe you are an expert in tools that are rarely used these days like Matlab. Any weird, legacy, or even obsolete skill you have has clients out there desperate for your help. If you can mesh technologies that were not designed to go together, even better! Once you have gained their trust, pivot to the projects you actually want to work on. Take a look at proposals out there to get an idea of what clients are looking for. And if you really want to do well, don’t fight your niche like I did when I started out, embrace it.
7. Go forward and conquer!
Freelancing has been very rewarding for me, and not just in a monetary way. I now see the world in terms of opportunity costs. Do I really want to binge watch another Netflix show tonight or make an extra $300 towards that trip I wanted to take? Sometimes I do go on a binge, but I find I do it much less now because I finally understand the value of my time. Not just with freelancing; I now spend a lot more time with family or learning new skills. For me, freelancing created a paradigm shift that has allowed me to be a happier and more productive person.
So what are you waiting for? Stop reading this and jump on a platform! (unless you want to connect or clap first that is). Here is my Upwork profile in case you need inspiration: https://www.upwork.com/freelancers/michaelmalin
About me
I am a Data Science freelancer with over 10 years of experience. I am always looking to connect so please feel free to:
Please feel free to comment below if you have any questions