11 steps to successfully implement AI in your gym
July 22, 2021 •
So, you’re interested in the notion of an AI gym. You’ve heard a lot about the power of AI – how it can increase gym membership sales, shape your gym referral program, refine your fitness marketing strategies, increase gym revenue, and far more besides.
The thing is, you’re not sure where to start.
We get it. The idea of an AI gym is exciting, but it feels somehow futuristic.
The good news is that it’s actually very much of the here and now, with gyms around the world already implementing AI and seeing exceptional results.
And it all starts with the data that already exists in your business, if you know where to look.
This data will be the fuel for the insights, innovations and results that stem from AI, and you want to be harnessing as much of it as you possibly can. Believe us, there will be no business wishing it had less data in the coming years.
With that in mind, if you want to turn your gym into an AI gym, your first steps – and the first six steps in this blog – must focus on your data practices, ensuring you’re capturing the data that’s required to power AI.
Step #1: Conduct a comprehensive data audit
Asset evaluation is a great place to start, with an excellent exercise being to look at every area of your business – every system you use – to map where data is captured.
You might call it a data audit (we call it an asset audit due to our understanding of what can be done with it!), but list the data capture points in two columns: the online and offline.
Be comprehensive, as you won’t want to find out in 12 months’ time that you missed one of importance. Look at the tools used across the business – not just proactive collection such as membership applications and PT bookings, but also passive data sets such as entry systems.
For the offline, review what becomes digital, noting whether every item does indeed make the jump and what, if any, captured information gets left behind. Note any islands of offline data that are not digitised, so you can be aware of them later for evaluation of their value and relevance.
Finally, consider any automatic data collection carried out by gym software such as entry systems; often, these sources are missed as they do not have human involvement.
Read more about the valuable data that already exists in your business here.
Step #2: Ensure your gym reporting software catches everything
Give each data source a condition value: a score of 0-10, with 0 being ‘currently a mess’ and 10 being complete and perfect. Your focus here should be on the manual entry systems, as it will be the human element that lets you down.
However, even for automated systems, check what data capture level you’ve chosen. Have you set it at a default level, by which you’re only capturing the data you need right now? If so, is there an option to expand this?
A good policy is simply to collect everything, even if you don’t have any current defined requirement for all of it right now, just so you don’t later regret any omissions. After all, you never know what could be of value in the future; imagine finding out, a few years from now, that some incredible insight is possible courtesy of a new tool – but not to you, because you chose not to collect the relevant data. If there’s no additional cost associated with doing so, turn it all on!
Step #3: Fill the gaps in your data sources
Any data sources scoring below 7/10 should be identified, and a plan put in place to both rectify the historical gaps and educate on what’s required going forward.
There may be a lot of work here, but there are freelancers on sites such as Elance who have the skills to fix data issues such as these, and who are often able to supplement with expanded data too.
Step #4: Connect the islands within your gym CRM software
Ask yourself: Can automation play a role? Do you currently require fields to be filled manually where this could actually be done automatically?
For example, does the personal training sign-up form include gender or age, when this data is already held on the member record or other element of your gym CRM software?
In all likelihood, once you stop seeing your different data sources as isolated islands, you’ll realise much of the data you need to fill the gaps you’ve identified already exists in the business.
Step #5: Put someone in charge…
Establish responsibility. What gets measured gets managed, as they say, and this is no different. You would certainly have someone looking after the cash in the business; data should be no different.
A Chief Data Officer maybe a little too grand a title (though 25 per cent of Fortune 500 companies now have them), but you want someone who can own the business data performance, monitor your gym reporting software and report on data conditions.
In some businesses, the owner/principal would be best, at least until culturally it becomes an accepted behaviour. The senior marketer is also a good choice, as they are usually the current beneficiary of good data practices and will have cause to access most of the data sources within the business.
Step #6: … but also ensure everyone takes ownership
Implement a data policy. Do not get this confused with GDPR or customer confidentiality, though you can undoubtedly integrate these if you wish. Your data policy should describe the importance placed on data within the business, explain how diligent you expect staff members to be, and encourage them to report any data-related issues.
Everyone in the organisation needs to understand the value of what’s being collected, and their role in ensuring it’s done correctly.
In our experience, the best way to do this is to make clear the benefit to the business of them doing so, rather than focusing on rules or reminding them that it’s part of their job. They probably all know they should be doing it, but it’s the difference in diligence around it that will determine whether or not you’re working with quality data in the future.
Step #7: Determine why you want to become an AI gym
Now you have your data in check, you’re getting close to the point of pressing ‘go’ on your plans to become an AI gym.
Before you dive straight in, though, there are some important questions to ask.
The first is: Do you have a problem to solve?
We see too many examples of AI solutions being asked to look for a problem. Likewise, we see operators enamoured of a concept (by which we mean the buzz around artificial intelligence) without ever being likely to see value from its implementation.
As we’ve discussed already, there are plenty of areas of potential, but you need to ask yourself some important questions before you take the leap, namely: “If I had this insight/value/prediction, would it be valuable?”
Step #8: Dedicate time to the project
Next ask yourself: Do you have the time for implementation?
The results from properly applied data + AI projects can be exceptional, but the initial pay-offs over the first six months are marginal. It’s normally 12 months before we see the true justification.
To get to that point, there must be a willingness not only to dedicate the time to implement the platform, but to then engage with it – and potentially change many of your workflows once armed with the new insight.
During the early days of Keepme, we found that after delivering a Keepme Score on every member – effectively giving our customers a window into who would leave and who would stay – fewer than 20 per cent would do anything with it.
Let’s consider that for a moment. These operators could now see, often with a runway of more than nine months, which of their members would not be with them when it came to renewal if they did not take action. Yet 80 per cent chose to do nothing with that information.
With retention a hot topic for every operator in the sector, we were perplexed, so we reached out to find out why. What became clear was that the insight didn’t fit their process. Meaning, being told that Member A would be gone in six months if they didn’t take action was valuable, but they didn’t have a workflow to do anything about it.
Step #9: Make a plan of action
So the next step, of course: What are you going to do with your results? How will you use the insights from your AI to increase gym revenue?
You may feel this is a simple question, but with the above scenario in mind, it’s clear the answer to the question ‘what are you going to do with your results?’ isn’t always obvious – not even once equipped with a valuable prediction.
The answer must be this: You must have (or if you don’t, then put) a process/workflow in place to take advantage of your new insights.
- If I know I’m going to miss gym sales targets in three months because of poor performance in my social media campaigns or other gym marketing channel, what can I do to remedy that?
- If I know which of the members who joined in the last three months have the highest probability of leaving, what will I do to change the likely outcome?
An AI implementation will give you access to a whole new level of transparency, but to gain value, you need to be very clear about what action will be taken as a result of it.
Put another way, the value is never in the insight. The value lies in the action that’s taken as a result of the insight.
And that value can be significant. Find out just how significant in our recent blog, How AI can increase gym membership sales at your club.
Step #10: Identify opportunities for gym automation
In the first white paper by our CEO Ian Mullane – The Fitness Future: Rules of Engagement – Ian mentions the power of AI and automation. He is clear in his view that processes fail mainly because they have a human element within them.
With this in mind, the next step in preparing for AI implementation is to ask yourself: Will I add automation, not just in the data collection (step #4) but in the use of that data?
“My advice,” confirms Ian, “is that AI + automation is where the real results are.”
When you decide to deploy AI, you’ll need to first understand what insights you’ll receive, then work out what actions you want to be taken with those insights.
Then pause and consider this: Can (any of) these be automatically actioned, rather than requiring a team member to be engaged?
This can be as simple as automatically generating a call list – all members who have become at risk in the last seven days, for example – and sending it to Member Services, who then call those members as part of your gym member retention program.
At the next level, it can be the automatic serving of digital engagements that present customised offers to those members most likely to purchase them.
We’ve seen these and many more implemented by fitness operators, as well as across other services industries.
Step #11: Find the right partner
If you have a strong technology department stacked with Python developers and data scientists, then availing yourselves of some of the Open Source AI tools available – such as TensorFlow from Google – will get you started.
For the 99.9 per cent of operators who don’t, the last big question before you take the leap into becoming an AI gym is: Who will you work with?
Before you go rushing off to find yourself a data scientist, may we suggest that you first take a look at what’s already available for the sector. The best solution will always be your data fed into fitness operator-specific applications.
This doesn’t mean you’ll end up with some cookie-cutter model similar to everyone else, as your data will drive the results. However, using a platform that’s designed for the sector will ensure the downstream effects – i.e. the value that flows from the model – will be relevant and usable in your current environment.
Right now, that means Keepme – the first to provide these services to the fitness industry in an AI platform that’s specifically designed for gyms. We’d be delighted to have a conversation with you to explore how you could become an AI gym.
To read more about implementing AI in your gym business, download our fantastic new white paper – Everything You Need to Know About Data & AI – for free, here.