Login

Why AI has transformative power for fitness operators - Keepme Blog

Home  »     »   Right here, right now – the transformative power of AI

ai-fitness

Right here, right now – the transformative power of AI
March 12, 2021 •

When you hear the term AI (artificial intelligence), what does it conjure up in your mind?

Do you imagine it to be something futuristic: technology that will one day power world-dominating robots, self-driving cars and so on?

You certainly wouldn’t be alone in this view, but to pigeonhole AI as something for the future is to miss out on the very real opportunities it presents to health club operators today.

AI is, after all, a significant force in our lives already: from deciding what we need to see in our news feed, to suggesting an artist we may want to play, to giving us to-the-minute arrival times in navigation apps, it touches us in myriad different ways every day. Its role and reach will continue to grow, but we are – all of us – already reaping its benefits in the shape of enhanced, more personalised experiences throughout our daily lives.

So, what does this mean in a fitness context?

When you look at the new entrants to our sector, whether they be existing tech players like Apple or the likes of Peloton, you can be assured of this: AI is at the heart of their product development, customer experience and business operations. 

It can, and should, be at the heart of yours too.

Let AI do the heavy lifting

The starting point is data, but please don’t allow this statement to set alarm bells ringing. You may have been told that AI isn’t something our sector will be able to take advantage of due to shortcomings in operators’ data. Please know this is fundamentally untrue.

Admittedly, clubs’ data capture may not always be purposeful or intentional, done mainly through necessity and not always comprehensively. Nevertheless, the data already exists in your business to improve and refine your offering in a multitude of ways.

Indeed, from our conversations with more than 100 club operators around the world in the last 18 months, without exception the data they needed to transform their businesses – from personalisation of the member journey, to increased secondary spend, to better targeted products and services – was already there. 

You would be amazed at what can be done with incomplete datasets, too: it’s now entirely possible to clean and enhance them, filling gaps and making them fit for purpose. 

By then applying today’s powerful AI and machine learning tools to this customer data, you’ll unveil a level of insight that, in many cases, makes it blindingly obvious where improvements can be made and opportunities taken. The insights are right there in the data. They always have been, in truth, but with AI now come of age, it no longer requires a staff of data scientists or business intelligence folk to get to them. The tools simply request you feed them data, tell them what you want to know, then leave them to do the rest.

Let’s take a look at some specific examples of how this – the application of AI to your customer data – can be used to transform your business.

Personalise the member journey with AI

With the availability of AI and machine learning, we have the capability to understand exactly where the customer is in their journey, as well as what engagement is most likely to be well received at that particular time. 

Let’s start by looking at the customer’s probability of continuing with the club at the next opportunity they have to make that choice – i.e. member retention. How can an operator gain an understanding of a customer’s current mindset? And, crucially, how can this insight be secured in a timeframe that allows the operator to change the outcome?

Feed a basic set of standard member data into an AI model and it will generate huge volumes of additional items that we call ‘features’. With a basic member dataset of 45 items – items that might include the member’s age, gender, join date, lead source, membership type and attendance data, for example, depending on what data the club has access to – a machine learning model can generate 8.8 million decision trees to predict many outcomes. No single decision tree is statistically significant, but when combined, you’re presented with a highly accurate prediction of likely member behaviour. 

In turn, this real-time indicator of retention probability – accurate at any point of the customer relationship, whether they’re in week three of a 12-month contract or on day 23 of a monthly pay-as-you-go agreement – means a user journey can be crafted appropriately, with personalised engagement. 

Benefits of this AI-enhanced approach include: 

  • Member engagement in a time window where you can still change the outcome.
  • Confidence you are not waking sleeping members.
  • The ability to customise an offer to minimise attrition risk, including harnessing dynamic pricing.
  • An enhanced understanding of…
    • Your demographic strengths and weaknesses: member-level data allows you to identify where age and gender are playing a part in attrition. 
    • Human factors in attrition: site-level data can highlight discrepancies between clubs offering exactly the same class, for example, but achieving very different levels of attrition.
  • Confidence around who to approach with opportunities for non-dues spend.

Even better, it can all be automated. Welcomes, attendance prompts, birthdays, milestone achievements… Every stage and every action can be set to happen based on simple criteria. Each engagement is customised to the communication preferences of the member (including GDPR) and, where there are options, optimised to the required outcome. Personalisation – in the form of personal preferences, gender- and age-specific images – is automatically built in.

Such automation will become a central part of every operator’s strategy; the often-heard “it’s too impersonal” is a hugely outdated understanding of what’s now possible.

Grow non-dues revenue with AI 

Your AI will also be capable of looking at your membership base and optimising your non-dues spend.

Say you offer personal training for purchase as individual sessions, or in packs of five or 10. Your AI can first identify the customers who will purchase from you. It can then segment them further into members most likely to purchase each of the particular packs. This ensures not only that you drive revenue from your membership, but that you do so without leaving revenue on the table. 

Andrew would buy a five-pack, but he’s just as likely to purchase a 10-pack, so serve him that offer. Kate, meanwhile, is highly unlikely to want anything but a single session, so don’t scare her off with a five-pack offer. Bank the one-session sale.

Add the previously discussed automation and you have AI working 24 hours a day, seven days a week to find members who want an offer, before serving them with the correct one in a piece of communication that at the very least is relevant to their age and gender, and delivered in the optimal engagement format for them. All with no human intervention. 

The potential is massive – one customer adopting this approach increased non-dues revenue by 43 per cent in six months – and the use case highly flexible to encompass many different areas. 

Harness dynamic pricing with AI

We previously mentioned that AI provides operators with the ability to customise an offer to minimise attrition risk, including harnessing dynamic pricing. Here’s how this can work.

Say you’re an operator providing a multi-activity facility: alongside fitness facilities, you also offer an outdoor pool and racquet sports. These outdoor activities are somewhat seasonal and, in the late spring and summer months, you duly see a surge in membership joiners. You then see a rise in terminations as autumn draws in. 

Introduce AI + automation, however, and you can now offer memberships that suit customer requirements. Your AI will predict which of your newly-signed members will prove themselves to be summer-only, and which are likely to extend past that. 

For the summer-only members, an offer is dynamically served to them at the optimal time to propose another tier of membership – one that reflects the reduced facilities in the winter months. Then, as winter passes and spring commences, the automation will serve the revised Summer Upgrade membership, seamlessly and without the need for staff intervention. 

Gyms already offer a number of different membership plans that incorporate entitlements around facilities and times of day when members can access the premises. With AI + automation, each member could be provided with a dynamic membership price – each quarter, say – that best reflects their use of the facilities. 

Adoption of this approach would move us away from the binary member/non-member relationship and provide a more customer-centric approach – one that will drive lower attrition and build longer-term relationships as members see the commercial part of the relationship centred around their specific requirements.

And again, this is possible right now: one of our US customers introduced precisely this style of dynamic pricing to its business in 2020. 

There are many more ways in which AI can optimise pricing, too: it is completely plausible that gyms might manage pricing dynamically taking into account site locations, facilities and the competitive environment around them. 

Indeed, pricing is an area with so many variables – the majority of which will materially affect the outcome – that allowing an AI model to monitor them all, looking for the optimal outcome, makes far more sense than relying on gut instinct or market experience.

Tailor the product for retention

Neither should the impact of AI on operations be underestimated. Consider this as a deliverable example: 

  • Your AI sees member and class attendance and the schedules you have in place. 
  • It also understands the drivers behind your member retention performance in quite literally a million different ways. 
  • You ask your AI to optimise your class timetable for member retention. 
  • Your AI will take into account every variable and generate new timetables. 
  • With each variation, it will provide you with the expected retention improvement from implementation, allowing you to optimise the schedules for member satisfaction. 

In many ways, even the above examples are pedestrian when you consider the extent of the technological capabilities available today. The possibilities to streamline and add value to a health club business are almost unlimited, with AI leveraging a club’s data to unlock new insights and opportunities to boost member engagement and satisfaction, as well as to attract and recruit the right members in the first place – individuals who, matching your optimal member profile, are more likely to stay with you and drive a higher lifetime value. 

A time for action

If any of the scenarios in this article sound futuristic, or at least not immediately attainable, let us finish with this: every aspect of what we’ve outlined is deliverable today – in part or wholly depending on your requirements – using commercially available products and expertise. 

And the results speak for themselves. 

Every single operator we have worked with to implement AI has improved retention within a few months, with one Australian operation boosting it by 9 per cent in the first six months. 

Another Australian operation took just 30 days to see a difference in cancellation rates, thanks to the fact that – drawing on 13 years of data – AI was able to risk stratify its entire member base immediately, with the club then able to turn this into targeted strategies.

Meanwhile, one US customer reported its highest ever member engagement rate with its digital campaigns – newsletters, renewal offers, NPS participation and so on – once AI was harnessed to ensure the right content was sent in the right format at the right time: average open rates rose from 47 per cent to 72 per cent. Another US customer logged a 52 per cent increase in NPS completion by members following AI assistance. 

One last example comes from an Australian operator that, by running a non-dues revenue campaign underpinned by AI, recorded its best ever result: an impressive A$35,000 upsell in just two weeks. All of this by targeting members – selected by AI and based on habits, frequency, likes, age profile and so on – with a tailored offer that made them feel the club genuinely knew them and what they wanted.

All of these opportunities exist, now. It all comes down to what action you choose to take.

Find out more in our free white paper, The Fitness Future: Rules of Engagement.

-->