This post belongs to our series of “technology in sport” and on this occasion we want to analyze the impact that Artificial Intelligence can have on your sport or fitness business. We will start with a more theoretical approach to the technology and will finish with some examples to help you comprehend how you could apply it within your organization.
What is Artificial Intelligence?
Rather than us giving you a long definition, we believe this video may explain more accurately what Artificial Intelligence actually is:
In brief, Artificial Intelligence (AI) implies providing machines with “human-like” intelligence to perform tasks in similar fashion as we do, considering 3 types of actions:
Generalized Learning which means “reacting to a given task appropriately,” or put another way, “adapt.”
And based on behaviors the machine has learned in the past it could ultimately:
Reason
Solve Problems
Moreover, this illustration from PWC also defines Artificial Intelligence and some of the realms it can be applied to:
The difference between Artificial Intelligence, Machine Learning & Deep Learning
While we are not experts on the matter, we believe it is important to establish the basic difference between Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning (DL) before moving on.
As defined above, Artificial Intelligence is the act of computers doing things that we would consider intelligent among humans, helping solve simple tasks.
Machine learning on the other hand refers to programs that get better the more you run them. This is where predictive analytics would come in, for example. Within ML, we can distinguish between three types:
Supervised ML which implies that the program is able to learn and make predictions based on previous data input provided by humans.
Unsupervised ML, where the aim is to transform unstructured data into structured sets.
The common thing though, is to have a combination of the two, and this is referred to as “knowledge mining.”
A third type would be reinforcement learning, defined in this article published in towardsdatascience.com as follows: A system interacts with a dynamic environment in which it must perform a certain goal (such as driving a vehicle or playing a game against an opponent). The system is provided feedback in terms of rewards and punishments as it navigates its problem space.
Finally, Deep Learning is comprehended within the realm of Machine Learning although the main difference is that it deals with complex problems.
As a general summary, the following diagram from Bismart, a Spanish data management and analysis business, may help explain the difference between Machine Learning and Deep Learning:
Examples of Machine Learning applied to the sports industry
You may be wondering how ML can be applied to the sports industry so we wanted to give a few examples to help you understand the concepts a bit better.
Supervised Machine Learning Examples
Price optimization: Imagine a sport apparel business selling online looking to understand the correct price to offer a customer depending on the phase of the customer journey they are going through. In this equation, variables such as previous prices, volume or channel could be factored in to get to the optimal price.
Churn Rate forecast: Consider a fitness business looking to understand its monthly churn rate. Based on previous historic data, the model could make an attempt to predict this ratio for the upcoming months, which could be particularly useful during certain periods of the year such as summer or Christmas.
Forecast up selling opportunities based on previous customer purchasing data. For instance, it could help a sports club identify those fans that are more likely to purchase official merchandise. This is a fantastic way to tackle LTV, a key metric for any sport or fitness business as we previously discussed in this post.
You could even apply this to the realm of talent management as it could help predict employees who will remain more time with your organization, those who could have “star potential,” or who are more likely to impact brand culture.
Unsupervised Machine Learning Examples
Anything related to customer segmentation initiatives as raw data could be analyzed and structured in such a way that it provides a clearer picture of “who” your customer is.
Is your sport organization considering launching its own OTT? Then ML can help as it can be applied to developing recommender systems, grouping users in such a way that they are delivered similar content.
Unsupervised Machine Learning could also help analyze incoming calls into the contact center, determine patterns in doubts, complaints, issues, etc. and lead to making decisions that improve the overall customer experience.
Business cases of Artificial Intelligence being used in sport.
To finish this article, we want to share real examples AI being used in the sport and fitness industries:
Chatbots in the fitness industry
Prior to the start of the Covid-19, we were actually involved in a project that aimed at implementing a chatbot-based solution for a leading fitness operator. Although the business decided to postpone the project, we believe chatbots have huge potential to help sport organizations in the future as it can help across several areas.
In our case, we identified the following realms of interaction:
Answer customer queries such as “What is the price of the membership?”/ “At what time does Yoga class start” or “Do you have virtual classes?”
Carry out transactions such as making reservations for a given class or hire a personal trainer.
Converting sale leads into customers or even preventing membership cancellations.
Given the new habits that people have picked up during Covid-19, we are sure chatbots will become a general practice among industry players. In our view, customers are going to be more reluctant to wait in line at a gym´s reception or at the ticket gate of a stadium. They may prove to be a great solution here as they enable many “day to day” operations to be done from home, improving the operational efficiency of the business.
Now, you could be thinking that at this time, you do not have the financial resources to invest in a project of this magnitude. We actually started doing the project we talked to you about without having to spend a dime. On one hand, many AI service providers overlook the realm of possibilities that a sport or fitness business can cover for the development of a chatbot service. On the other, there are many startups in the space looking for “success stories” and “testing spaces.” All you would need to do is reach out to such companies and offer your sport organization as a potential case study for them.
Personal Training and other Fitness based solutions
During the Covid 19 pandemic, brands in this space have experienced tremendous success. This is where you would find brands such as Peloton, Tonal, Mirror or Zwift (to name only a few).
Although there may have been an initial skepticism to whether these solutions would work in providing a personalized experience for users, in this Forbes article, Daniel Sobhani, CEO of Freeletics confirms that 85% users rate their workout as perfect, and it is all AI-based. He also points out:
The service predicts what exercises you’ll be able to do and want to do based on first getting a very short profile from you, then comparing you to Freeletics’ other 50 million users.
Create workouts out of sets of exercises while learning how you react and what results you get.
Adjust workouts based on available equipment ... because you probably have less gear at home than in your gym.
Similarly, brands such as Nautilus are betting big on this space, developing their own AI based solutions with the objective of generating new revenue streams and entering new markets. In fact, and as we covered in this post, Nautilus has mapped out a "roadmap to $1 billion," hoping to acquire 2 million users. One of their main drivers of growth would be their AI based service, Jrny, a virtual trainer capable of designing virtual workouts for members through an app or the brand´s cardio equipment.
Predict Player Performance
A realm that is particularly relevant for sport organizations is the player performance space. With a given set of previous data, a club could predict, to some extent, how a player would perform in a particular game or if there is potential risk of injuries at some point during the season. This is actually where companies such as OloCip, a firm founded by former professional footballer Esteban Granero and which we talked to you about in The Ballketing Letter #7, operate.
Marketing & Advertising: Ticketing, Fan Segmentation, Sponsorship & Advertising Results
As we described above, AI could also help create cohorts of fans in such a way that it would enable personalization of the fan experience for broad categories including:
The loyal fan: Your season ticket holder, for example.
The regular fan: Although not a season ticket holder, this fan attends a few games during the season.
The casual fan, which goes to one or two games each year
The first timer or “the tourist.”
The “sleeper” which essentially is that fan that used to interact with your organization rather frequently but has been “missing” for a few months / year.
The important aspect here is to notice that not all these fans are in the same phase of the customer journey. As we described before, in an effort to increase relevance and therefore, sales, the commercial proposition for each segment should change. The loyal fan may have more chances of buying official merchandise than the “sleeper,” for example but on the other hand, a discount on a ticket price may reactivate those customers, bringing them back into the funnel and taking it from there.
As within other industries, Artificial Intelligence is a valuable “Revenue Management” tool that helps deliver the right product at the right price and at the right moment. In this episode of Harvard Ideacast, they discuss how AI can be applied to gather and understand your fan´s or customers decision with the aim of helping drive profitable business decisions or even put in place dynamic pricing initiatives. In essence, think about applying this to your ticketing strategy (both one off or season ticket holders) to maximize revenue.
Finally, there are several companies such as Blinkfire or Horizm that offer AI based tools that aim to optimize revenues coming from sponsorship & sales of digital assets that are experiencing amazing growth. Horizm has signed deals with major sport properties such as Real Madrid, the Australian Open or more recently, with Chelsea FC.
Blinkfire on the other hand, develops a solution that enables clubs and brands to measure and quantify the economic impact of the exposure of a given brand. It pretty much works as follows:
This is a great technology that helps sport properties and brands optimize brand exposure in critical moments. For instance, in this podcast episode a Blinkfire executive describes that even the way in which player pose for the initial line-up information has changed due to technology like this. In addition, they provide a great framework to renegotiate sponsorship deals based on value of the brand exposure.
Automatic Content Creation
Finally, another solution AI could provide sport organizations is the possibility of creating automatic content related to a given event. This implies analyzing which pieces of previous content created more engagement (both live and virtually) and when the next event comes up, the solution will be able to create similar short form content that aims to perform just as well, without you having to carry out a manual selection of that content.
Within this space, in The Ballketing Letter #7, we also described Edisn AI, which is an “AI powered fan engagement platform with state-of-the-art player recognition and contextual content delivery.” The opportunity they identified is that while content consumed on digital devices is customized (ie. Netflix), broadcast content is still linear, which has led to a decrease in attention spans. Edisn AI provides a solution in which they make the screen interactive to foster fan engagement and ultimately, monetization. For instance, a fan will be able to check stats from any given player during a game or even purchase official merchandise directly. In addition, those engagements will help develop different fan profiles, create customer segments, and enable delivery of personalized content.
These are just a few examples of where AI may have a major impact across the sport industry but as shown in this interesting framework from PWC the scope of opportunity is much wider:
Artificial Intelligence, the next frontier for the Sport Industry
Hopefully, this article has provided greater insight into what Artificial Intelligence is and how it can help your sport or fitness business in the upcoming future. While other technologies such as Blockchain may still take some time before they are widely adopted, AI is already being used across several realms and we are confident that solutions are bound to improve in the short-medium term, making it the next frontier to explore in the sports industry. Our recommendation is to embrace it and figure out if it is something that would add value to your organization whether it is in your sales & marketing tactics, operational processes, talent management or even from a “cost management” perspective.
We will be back with more articles on “technology in sport” in the future with the objective of helping you understand how they will lead you to take your business to the next level.
Meanwhile, keep safe.
Comments