What if you showed up at the gym and your phone told you exactly what workout to do to achieve optimal gains?

Under Armour partnered with Watson in January 2016 to make predictive, personalized fitness and nutrition reality

If you’re thinking, “Wait, Under Armour is a clothing company.” You’ll want to keep reading.

In 2013, Under Armour took the plunge and began its transition from an apparel to a technology company with the $150 million acquisition of MapMyFitness[1]. Then, in 2015, they acquired Endo Mundo for $85 million and MyFitnessPal for $475 million. Following this series of acquisitions of connected fitness companies worth nearly one billion dollars, in early 2016, Under Armour announced a partnership with none other than IBM’s Watson.[2]

In the initial communications, IBM and Under Armour declared the following:

“Under Armour’s Record app, combined with a Cognitive Coaching System will serve as a personal health consultant, fitness trainer and assistant by providing athletes with timely, evidence-based coaching around your sleep, fitness, activity and nutrition, including outcomes achieved based on others “like you.”

At the time it seemed the IBM Watson partnership was the perfect next step for Under Armour as they worked to capitalize on their recent acquisitions. Watson would help Under Armour turn their user data from ~225 million downloads[3] and 10 billion digital interactions[4] into a predictive, personal trainer and nutritionist for everyone. Their initial announcement revealed their plans to deliver customized exercise programs and food logging through visual recognition within the next year.

This bold move fit within Under Armour’s mission statement of making “all athletes better through passion, design and the relentless pursuit of innovation”[5]. Unfortunately, the digital transformation powered by machine learning at Under Armour has been met with declining top-line sales in the United States, management turnover, and saturation of the athletic apparel industry[6], making their foray into connected fitness slower than they originally hoped. In fact, their timing for a big move couldn’t have been worse.

Since January 2016, the company hasn’t shared an update on their long-term plans for the Cognitive Coaching System, and instead has rolled out tech-based apparel. At the Consumer Electronics Show in 2017, they unveiled pajamas that help you “sleep better and recover faster.”[7]. In 2018, they announced the third generation of the HOVR shoe series. The HOVR features a tracking device embedded in a 3D molded sole. The 2018 HOVR launch coincided with an update to the MapMyRun app that allows runners to assess their stride length and cadence[8]. These new products, while interesting, are variations of the previous version and are not in-line with the disruption Under Armour set out to accomplish just two years earlier.

Over the medium-term, I fear Under Armour will continue to focus on incremental changes that drive bottom-line performance, rather than staying focused on truly disruptive innovations. And what the fitness industry needs is disruption. Emmet Williams with Club Industry[9] imagines a “solution that prescribes a body-weight workout in your bedroom with Dwayne Johnson, aka The Rock, knowing that you missed your resistance training boot camp yesterday. Then imagine you are prompted for breakfast being an egg white omelet because your DNA profile suggests you are not optimized on your macro nutrient goal this week.” Under Armour was on its way to create this type of disruption, but it is unclear if they are still on that track.

Without much information on how their Watson partnership is working, Under Armour might consider acquiring a company that specializes in machine learning that could be completely dedicated to this project. Another alternative is to sell rights to the data generated from their connected fitness platforms. In the growing world of health start-ups, there must be a company that would love to get their hands on this data. Each of these options ease concerns about Under Armour’s ability to attract the tech talent required to create the disruptive innovation the industry so desperately needs. Finally, the business of retailing is a difficult one that requires focus and continual adaptation. Taking their eye off retail to focus on technology could derail the brand and the core engine of Under Armour, so a partner in addition to or in lieu of Watson might be just the trick.

What I wonder is, could they prove me wrong? Could a retailer truly transition themselves into the world of tech and create the next generation of fitness products and if so, should they? Finally, what other companies might be positioned to disrupt the fitness industry if Under Armour can’t achieve their goal?

(718 words)



[1] The Motley Fool, “I Still Can’t Believe Under Armour, Inc. Spent $710 Million on Connected Fitness,” https://www.fool.com/investing/2017/04/28/i-still-cant-believe-under-armour-inc-spent.aspx, accessed November 2018.

[2] “Under Armour And IBM To Transform Personal Health And Fitness, Powered By IBM Watson,” IBM press release (Las Vegas, NV, January 6, 2016)

[3] “Healthcare 2030: Obscure Technologies that will take us beyond wearables.” Chris Van Hoof, Medical Design Technology, (April 26, 2018)

[4] ClickZ, “5 businesses using AI to predict the future and profit,” https://www.clickz.com/5-businesses-using-ai-to-predict-the-future-and-profit/112336/, accessed November 2018

[5] Under Armour, “About Under Armour,” https://www.underarmour.co.uk/en-gb/about-under-armour.html, accessed November 2018

[6] Mitch Kummetz, “UAA: Initiating Coverage with HOLD Rating,” Pivotal Research Group, October 5, 2017, p. 2, accessed November 2018.

[7] Under Armour, “Wake Up and Hustle,” https://www.underarmour.com/en-us/tb12, accessed November 2018.

[8] Men’s Health, “CES 2018: Under Armour Debuts the Smartest Shoe You’ll Ever Wear,” https://www.menshealth.com/fitness/a19546348/under-armour-smart-shoe, accessed November 2018.

[9] “The Gym Experience of the Future: The Convergence of Innovation Will Change the Member Experience.” Williams Emmett, MYZONE, Club Industry; Overland Park (Feb 23, 2018).


Crowd-sourcing the Secret of Life: 23andMe and Open Innovation


CircleUp – Identifying the next “Big Thing” using Machine Learning

Student comments on What if you showed up at the gym and your phone told you exactly what workout to do to achieve optimal gains?

  1. Thank you for this great essay on UnderArmour. I agree with your concern that UA needs to take fitness to the next level, and by focusing on incremental changes they could drop the ball on their retail performance which could hurt their ability to innovate further. While I agree that they should move in the direction to leverage the digital age, I am concerned that Nike will get there first, since they have a much larger company with a strong innovation culture. Nike’s recent article on Medium (https://medium.com/nikeengineering/serving-athletes-with-personalized-workout-recommendations-285491eabc3d) shows that they may be ahead of the UA, albeit currently geared towards athletes – though one could also foresee an easy transition of the technology to regular customers.

  2. Thanks for the interesting read. I guess it’s hard to say what’s cooking with the Cognitive Coaching System at present. For the better of humanity (and my physique), I hope under armor figures out a way to help us all work out more efficiently. I do have some doubts over the quality of data the company can collect from its users. Perhaps they can use caloric data from My Fitness Pal, step counters/heart rate monitors on smartphones and watches, rep/set count apps etc. But some of these are notoriously inaccurate (eg. myfitnesspal) because they are self-reported. In any case, maybe some guidance from the machines is better than none.

  3. Personalization is a key trend in consumer/retail industry. Machine learning enables brands to give a more customized experience to consumers. As tech and athleisure getting bigger, UM is well positioned to enter the broader fitness category. Others that are silimarly well suited are Lululemon, Alo Yoga, etc, whom already did offline fitness classes offerings for years to their consumers. Alo Yoga started to offer online classes already. It’s a matter of when they take things digitally.

  4. Really insightful piece! This is an example of machine learning that makes one ask how long an organization should fund potentially groundbreaking technology rather than recognize the R&D as a sunk cost and move on. Unfortunately, it seems as though UA ventured outside of its core competency and began investing capital in a space that it lacked knowledge, underestimating the task at hand. On the other hand, can you blame an athlete first company for trying to push the boundaries of health and fitness through the use of innovative technology that would significantly help it capture market share from its largest competitors?

Leave a comment