Managers often say, “What gets measured, gets managed”. In corporate life, we discuss deliverables and key performance indicators (KPIs). The same goes for any health and fitness goal, without any sort of measurement system, it will be difficult to realize small, medium, and large gains.
A lack of data prohibits us from reflecting on the journey — whether we are just starting or have been on the grind for some time. In my behavior analytic world, routine data collection is a cornerstone of my practice.
In the gym world, we all experience many of the same interventions, whether on purpose or naturally. These often include starting new gym routines, working with a new coach or trainer, updating your workout program, or changing gyms and class times. With so many variables changing, I need to evaluate the effects of some intervention, and I can’t do that without measuring something in my routine.
This was the draw of CrossFit to me. Most gyms have consistent coaches, programs, class times, and most importantly, standard benchmarks that you can test. If you can do a workout faster, finish more rounds, or lift more weight, then you will know that some aspects of your strength or fitness improved.
I have collected fitness data for over a decade and experienced my fair share of navigating certain (e.g., moving cities) and uncertain changes (remember the pandemic?). This blog is not about those different workouts (e.g., 1 rep-max back squat, “Murph”, 10 rounds for time) or evaluating program changes over time, but how my data collection system evolved over the last decade. I went from not working out and collecting zero data to collecting some data, eventually leading to routine data collection, hardcore analyses, and developing many dashboard views.
My data collection journey is broken down into distinct phases:
- Learning To Lift
- App-Based Collection and Monitoring
- Evolving To Exports and Dashboards
- A New Program, Custom Spreadsheets and Field Data Collection
Phase 1: Learning To Lift
Time frame: January 2014 – June 2014
Duration: 6 months
Primary Data Collection Method: Little-to-None
Data collection may not be a priority when starting out on any fitness journey. This is okay because much of your time is dedicated to learning new movements — pushups, air squats, mobility routines — and developing the routine to work out.
When I first drank the CrossFit kool-aid, there were seemingly endless movements (e.g., air squats, Olympic lifts, double-unders) and workout types to learn (e.g., AMRAPS, EMOMs, ROUNDS for TIME) all the while trying to survive each class. In this phase, I worried less about tracking any data apart from RSVP’ing to class, got into the habit of entering a score or two and checking the gym’s leaderboard periodically.
Props to each of the gym software out there. They make it easy to do these basics.
Phase 2: App-Based Data Collection and Monitoring
Time frame: July 2014 – September 2021
Duration: 7 years
Primary Data Collection Method: App-Based
After 6 months or so, I got a little in shape and the fun began as I contacted a lot of reinforcement for my hard work (aka “noob gains”). The second phase involved regular data collection in the gym’s app. As with any popular gym tracking app these days, you can log your workouts and strength numbers. The coach programs workouts in the app, so that following each workout you enter what you need (e.g., weight lifted, time, rounds completed).
These activities offer many people various forms of reinforcement: tracking your benchmarks and daily workouts, seeing your scores relative to others, and competing against your teammates.
People often ask me which app or device is the best, and the answer is always the same:
“The best app is the one that the athlete will use.”
I started my fitness journey back in 2014. I joined several gyms as I graduated, and moved, some gyms shut down, some gyms moved, graduated again, and I moved once more.
Yay, school!
The challenge with training for so long at different places is that each gym uses different software to collect membership and training data.
I was a member at the following gyms (with the years and data collection system):
- Indian River CrossFit (2014–2015, WodHopper)
- CrossFit 352 (2015–2017, No App — custom spreadsheet)
- NorthFlo CrossFit (2017, No App — custom spreadsheet)
- University of Florida Recreation Center (2017-2018, No App — custom spreadsheet)
- Gator CrossFit (2018 – 2019, SugarWod)
- Three Kings Athletics (2019 — Current, Wodify)
I have been a member at 5 different CrossFit gyms that used 3 different apps. For the gyms that did not use an app, I collected my training data on custom Microsoft Excel spreadsheets. I may have used templates from across the internet, but for the most part, I created and managed my own.
Lost data! While it is ideal that I have data for each workout, there are a few instances when this didn’t happen. Some days my training data were incomplete because I dropped into other gyms over the years and wasn’t in their app. Other times I lost data when the app crashed and failed to save, or I simply forgot to log my workouts. The pandemic naturally broke routines and deprioritized logging, logbooks were lost, or spreadsheets went missing on my computer.
Later in this post, you will see how analyzing my data and evaluating my gains over time became quite the technical challenge when I’ve been a member at different gyms that use different software.
Phase 3: Evolving To Exports and Dashboards
Time frame: September 2021 – January 2024
Duration: 3 years
Primary Data Collection Method: App-Based & Exported Data
In 2020, my professional interests led me to business intelligence software and analyzing large sets of data in Microsoft Power BI (PBI). PBI is a nerdy analytics tool that lets you graph countless ideas from different bits of data. I had a Eureka moment when I realized that I could analyze client data in PBI. At that time, my data analysis skills were as weak as my push press, and I analyzed and graphed my own and client’s data in Microsoft Excel. If you are familiar with applied behavior analysis, then you may know how arduous graphing data daily can be.
Microsoft Excel is great, and has its place, but has its limitations — it can be very, very time-consuming. I am not an Excel wizard but found immense value in the fact that all I have to do today is keep my spreadsheets updated and hit PBI’s magic “refresh” button.
I still logged data in Wodify at Three Kings Athletics, but I soon figured out how to export all of my historical data to graph and analyze in PBI. This was a game-changer. I recorded a YouTube video some years ago describing how I did this.
Dipping my toe in this business intelligence (BI) world, led me to combine data from multiple sources (each gym) in Power BI. I’ve renamed this work fitness intelligence. This created a unique opportunity — I could copy all the spreadsheets into one giant spreadsheet, but again, that would be very time-consuming. What I learned in the BI space is that you can clean and structure different files, organize them behind the graphs, and make your workout data refresh like magic.
The point of combining the data is that it allows me to answer questions like:
- How many times did I lift each year?
- What was my best “Murph” time since 2014?
- How much has my back squat improved over the past 10 years?
If all my back squat data were in 5 different places, then I would need to create a graph for each one. That doesn’t sound very fun because it would be difficult to compare charts with different X-axes, and you still have to make 5 charts! A realistic alternative is to create a master spreadsheet, but again, you would be spending hours upon hours organizing your data to answer even the simplest question.
The solution? Upload each spreadsheet into one cohesive data model. This is the power of using a tool like PBI. Loading each spreadsheet and linking them together by common data points like date or movement unlocks your fitness analytics world!
Here is a screenshot of what this “data model” starts to look like when we have workout data from SugarWod, Wodify, WodHopper, and other personal Excel spreadsheets that I’ve curated over the years:
The benefits of this approach are too many to name, but here are the most important ones:
- Data are easy to update
- Analytics are on-demand
- Ease at creating new visuals
- Find insights that may have been hidden in plain sight
When the data are organized like this, you can make charts like these that are beautiful, interactive, and insightful:
My PBI and spreadsheet system works great and is still the core of my fitness analytics. However, a new natural contingency occurred. I abandoned following the traditional class workouts and joined a group of peers with custom gym programming.
The wrench in this plan?
Workouts were no longer programmed in Wodify and cannot be exported.
Phase 4: A New Program, Custom Spreadsheets and Field Data Collection
Time frame: January 2024 — Current
Duration: 6+months
Primary Data Collection Method: Exported Data
Today, my data collection world evolved. I am back to entering data manually, in spreadsheets. This seems like a step backward, right? If apps are so easy to use, why did you make it harder on yourself? Well, at the same time, my training program evolved too. My workout partners and I went off script, no longer followed the programmed class workouts in the app, but had a separate training program.
How do we get these workouts today? One of our workout partners is an elite athlete, who receives specialized coaching and shares the workouts with the team. No app required.
I still needed to collect data, so what should I do? It wasn’t easy to figure out, but I had to run my cost/benefit analysis and reassess what was still important for me on my lifelong journey of health and fitness.
I wanted to continue logging the fitness essentials: strength movements and benchmark workouts. Bench, squat, Murph.
My Workflow Today.
A New Spreadsheet. To collect strength data, I mimicked the format (or structure) of the exported Wodify data collection sheets into a new Microsoft Excel spreadsheet. This step was necessary because in Phase 3 I set up a system of dashboards on the Wodify exports. If spreadsheets do not share the same structure, then they can’t talk to each other. With this new spreadsheet, I could easily add it to my model from Phase 3.
Amazon Fire In The Field. Now, how would I log these data? Before I had an app and an iPhone. With my behavioral expertise at the ready, I knew that collecting data is most effective when it occurs in the context that matters (at the gym) and immediately following some behaviors (the lift or workout). Factoring these in, my learning history, and current systems, I decided to buy an Amazon Fire with its sole purpose to collect data at the gym. I removed all the junk apps and put it in my gym bag.
A Failed Google Form. Before I landed on the custom spreadsheet and tablet solution, I attempted to use a Google Form to collect my lifting data. It did not work out because of its clunkiness. For one movement, I had to open my iPhone, find the bookmarked site, enter all the form data, and submit. This process repeated for each movement or workout. The data entry experience produced frustration when entering simple 5×5 back squat data, so I moved on to developing the solution above. I thought using a Google Form would be simple enough because you can easily plug it into PBI behind the scenes. But that data entry process was not a winner.
What’s Next In Data Collection?
Time will tell. As I reflected on the concept of this article, that time allowed me to assess how something as simple as collecting fitness data grew into quite the monster. An important monster at that. I will likely maintain this process and tweak it a little until it becomes no longer valuable, which won’t happen anytime soon.
AI, perhaps? Today, data analytics can be enhanced by various artificial intelligence (AI) tools. You can ask ChatGPT to write a fitness program, analyze data, and create charts for you! Could AI enhance what I do today? Yes. What will that look like? I am not sure. Any AI tool would have to seamlessly integrate into my workflow today and make my processes just a little better. I caught a YouTube video the other day about OpenAI analyzing Summer and Winter Olympics data; AI performed poorly when counting the correct athletes. AI will not replace the entirety of what I do because, behaviorally speaking, I (the human) need to be a part of the data collection process. It helps me grow, get better, and know what to work on next.
Meanwhile, I will push on with the current system that I have in place.
