4 Tips for Statistical Consulting – Learn from my Mistakes


I decided to perform my first independent statistical consulting project. What is statistical consulting? Statistical consulting is engaging with a client to provide statistical advice and/or services. This client found me through a recommendation from a professor. The client was an Ed.D. student completing their dissertation. I assumed the project would be quick and easy, but I was wrong. I made mistakes that made this task more tedious for me. The following are four tips that I wish I could go back and tell myself before I took on this project:

1. Know your worth

I experienced statistical consulting for companies and graduate students before, but this was my first time independently servicing a client. Additionally, I assumed this project would not be much work. This led me to accept a financial offer for my services that was less than I would have preferred. Never—and I mean never—settle for a price that does not leave you completely satisfied. It is not worth it! You’ll be kicking yourself while you are underpaid and overworked.

Guide to Pricing as a Statistical ConsultantHow do we develop our prices as a statistical consultant?

If I asked you to give me a quote for your services, chances are you would charge less than you should. Statistical work is valuable, so price your services accordingly. Salaried statistical consultants (with bachelor’s degrees or higher) are paid around $125,000 in the United States. You should treat this as a baseline for your prices. This means If you are using an hourly model for pricing, you should charge no less than \$60 an hour. You can increase your price if you have a master’s degree or experience. Use this hourly rate to determine the price of your services on a project-by-project basis. If a project will take you an estimated 10 hours to complete, you should charge \$650 (at \$65 an hour). Alternatively, you can use a package model to determine your price. For example, you can have a data visualization package where you provide four figures or tables for a client. Let’s say it takes a maximum of 8 hours for you to clean a dataset and prepare four figures or tables. Then the price for the package would be \$480 (at \$60 an hour). This gives you pricing packages where the client can select what works for them. Either way of pricing your service is great just make sure you know your worth before you set prices with a client.


2. Ask questions, then ask more questions

Before agreeing to the project with my client, we had an initial meeting to talk about the project and whether I was interested. I made the mistake of not asking enough questions which caused some issues after the fact. Always strive to understand the project, timeline, and the client’s expectations.

What questions should you be asking?

Asking the right questions can save you time and make sure you are both are on the same page. The key to asking good questions is to never assume anything! You want to know about the project background, timeline, and the statistical analysis they want you to perform. Your client probably will not have a strong statistics background, so don’t be afraid to question their statistical methods and how it relates to their research question. Determine if they have collected data, ask them about the attributes of their dataset (sample size is important). Outside of the research project details, make sure you are clear on what materials they want you to provide and how they want those materials to look. If at any point in the project you need clarification, ask the client as soon as possible.

3. Establish firm boundaries

When working on the statistical consulting project my previous mistakes came to home to roost. Due to data collection delays, the project was postponed until the start of a new semester. I was not paid for the extra work that I complete. I redesigned the method and redid the statistical analysis because of an error by the client. For weeks, the client was contacting me day and night to ask questions about my work after I completed the project. These factors caused me to struggle to get my assignments in on time. These issues could have been avoided by setting clear boundaries from the start.

How to determine your boundaries as a statistical consultant?

Talk to the client about your boundaries before you agree to work with them. When will you be available to be contacted? Do you prefer email or phone contact? If they ask you to do more work than previously agreed upon, refuse or request a payment increase. Be honest with yourself about how long a project will take and overestimate this time for the client. If you struggle with this, it may help to double the amount of time you estimate. Whatever your boundaries may be, make sure you enforce them. Don’t be afraid to tell a client no or terminate the relationship if necessary.

4. Create a client agreement

I did not have my client sign an agreement because I thought it was a small project, and it would be no big deal. This was not a good idea. For example, I had to redesign their statistical methods halfway through the project. I did not sign up for that, but the client in their mind believed it was only natural for me to do without a pay increase. A written agreement can help alleviate misunderstanding and give you a legally binding document to rely on if things go wrong.

Categories for a client agreement

What should you include in a client agreement?

There are certain things you should include in a client agreement. These items may change depending on you as a consultant and the type of client. In your consulting commitment section, you should include an explanation of the role of a statistical consultant and expectations for a client. This is a great place to state your boundaries. In the services section, you can describe what specific services you will provide for the client. You may want to think about the number of meetings, and revisions to your work as well. The terms section consists of the timeline and a termination date for the project. Information on payment and your refund policy can be found in the compensation section. The confidentiality section provides the client with a clear idea of your level of confidentiality. After preparing your client agreement, send it and have them sign. Keep this document in a safe place. If you have any issues with your client breaking the agreement, you can use the document as a reminder for your client. Worst come to worst, you have a legal document in the unlikely event that legal action is required.

Final Thoughts

I highly recommend giving statistical consulting a chance if you are interested. It can be rewarding if you are mindful of the potential pitfalls. Although I struggled with my first independent statistical consulting project—due to a lack of knowing my worth, asking questions, firm boundaries, and a client agreement— I learned a lot and have grown as a statistician and a person. Statistical consulting, as a Ph.D. student, can offer some financial flexibility, experience for a future career, and teach you how to find and manage clients.


Are you interested in becoming or are you a statistical consultant? What are some of the mistakes that you’ve made?

Share your experiences in the comments below!

About Chanel Lee

Hello, I'm Chanel Lee. I am a statistics Ph.D. student at Rice University who owns my own business. I provide services as an academic, career, and leadership coach, statistical consultant, and speaker. For more info on me and my services visit https://www.chanelclee.com/.
This entry was posted in Advice, Grad student life, Statistics. Bookmark the permalink.

2 Responses to 4 Tips for Statistical Consulting – Learn from my Mistakes

  1. Abbe Herzig says:

    When talking with a client about their project, go beyond the specific statistical analyses or visualizations they ask for. Make sure to ask them about their research questions. Often a client will ask for a particular analysis or visualization that may not be really what they need or want—that can create extra work for you later. Know what questions they want answered, and then discuss with them your expert opinion about what methods will answer their questions.

    • Chanel Lee says:

      Exactly! I assumed the client knew what statistical analysis they needed since they told me. Only after I realized they had me do the wrong thing. It’s better to sort it out in the beginning.

Comments are closed.