Marissa Dearing

  • Student

Activity Feed

On March 5, 2019, Marissa Dearing commented on Gaining Prominence in the Era of Digital Transformation :

I’m a huge fan of AI applied to manufacturing (the fourth industrial revolution!!!) – conversations about the digital economy often give short shrift to the physical side of scaling even ‘pure’ digital companies, much less largely physical companies like logistics and manufacturing. But as you point out, the potential here is enormous for eliminating dead waste in the system (better scheduling, predictive maintenance, process improvements, the whole TOM mother lode). My concern here about PTC is that I wonder if, as platform development and data analysis in general get more plug and play with easily accessible elements like interface templates, Google’s TensorFlow, and more freely available datasets, their competitive advantage in this space will dissolve. I imagine one predictive analytics platform taking in factory equipment data is much like another and I wonder how soon it is until this sort of thing is a commodity, and something a giant like SAP will just absorb as a handy feature (much like Microsoft and IE).

On March 5, 2019, Marissa Dearing commented on Tonal: Your Entire Gym On The Wall :

I’d never heard of Tonal and had never thought through how “AI” could be applied to workout equipment, but it makes a lot of sense. Unlike most of these massive scale, big data, AI plays, however, that >$2K pricetag is eye-watering! We’ll see if they can get enough scale to really apply machine learning to workouts (beyond getting the buzzword funding/stock boost). Also having to install the unit on a wall of your apartment is a big investment that may represent too much of a commitment for many people (especially because it’s really only weight training?). If I’m going to spend $2K and have to dedicate a certain amount of space, why wouldn’t I get a Nautilus or Bowflex and get a wider range of weight training possibilities? You’d have to believe that the big data fueled personalized workouts are real (or the display unit is just a lot more aesthetically pleasing…).

IF they can get the scale, I think the AI-powered personal training and classes are an excellent idea. It would be amazing to know based on your body and what’s healthy how hard you should be pushing, etc. Plus, while some users may really want another person there egging them on, others (like me) would probably prefer to work out without a personal trainer standing over them BUT get the knowledge and calibration personal attention provide. The social sharing aspect is a nice way to motivate yourself through your friend and family network.

As to the future of these AI workout companies, I’ll definitely keep an eye out and my muscles flexed.

On March 5, 2019, Marissa Dearing commented on Bye-Bye Britannica, Hey Siri :

Britannica vs. Wikipedia brings up some pretty scary aspects of the digital economy. One, in this era of network effects businesses (and the other 90% of companies that *think* their market is winner take all and so network effects are critical to their success), companies are willing to temporarily sacrifice profitable pricing in their own or adjacent industries in order to gain share in the hopes of one day making profits. This can be devastating for companies that have less investor capital sitting around or are more focused in this ‘loss leader’ industries. I’m thinking of examples we’ve discussed like Uber and cab companies/drivers, Amazon and paintball gear, and Microsoft and Netscape. These companies decide they can afford to hemorrhage cash while focused companies bleed out. Pretty unnerving for any company in an industry the Tech Giants decide is a nice-to-have platform enhancer. Really goes against sense that the magic of the modern capitalist economy is specialization…

Second, there are tons of benefits around crowdsourcing across industries (big believer in big data), though I’m not sure the benefits are applicable across applications… in the Wikipedia example, I am still quite concerned about the quality aspect… similarly, Yelp and other review platforms may tell you what the average person likes or doesn’t but not necessarily what YOU will like. And that’s if the people who *choose* to post are representative, and in Wikipedia’s case, informed! I’m skeptical.

Also your post conjured up some really nice memories around Britannica and Encarta, thanks for that 🙂

And LOL at your Siri search. Do you even go here?

On February 24, 2019, Marissa Dearing commented on Fitbit Care: Improving Care Delivery in a Digital World :

Two thoughts for you. One, it seems a platform like this desperately needs scale to succeed, and I’m unconvinced Fitbit can pull it off. Given that Fitbit Care is based on data from Fitbit devices, the more Fitbit devices people are wearing, the more data the company has and the more it learns about specific actions patients and/or care provides can take to improve outcomes. But the converse is also true: if Fitbit is eclipsed by Apple Watch, some new Google wearable, Under Amour gear, or any other wearable in this crowded and seemingly very faddy space, the value of the Fitbit Care platform drops precipitously. Also, if Fitbit Care only has access to data from Fitbit wearables but not access to other medical devices like pacemakers or even just wearables by other companies, Fitbit Care’s picture of the patient’s health will be critically limited, and another company that has the credibility and regulatory approval to integrate these different data sources (like Epic, athenahealth, or Cerner) will be far more valuable and likely disintermediate Fitbit Care (if Fitbit Care refuses, integration could be a big selling point for a competitor). Moreover, only these digital health record companies will have access to more sensitive information like patient histories, which make any recommendations far more relevant, tailored, and reliable. Lastly, I bet health systems have some of these same concerns and will be extremely reluctant to share sensitive data with Fitbit, in case it fizzles out as just another fad in the next few years.

Second (less related to platforms specifically), although I’m not someone especially paranoid about privacy, the dangers here seem real, especially in the hands of a private company like Fitbit (yes, you could easily reply, “But Google!” and my same worries would apply). I’m not sure consumers want their bosses to know how healthy they are, what their blood pressure is, when they’re sleeping, and potentially, where they are at all times… opens up a lot of doors to uncomfortable involvement by an employer and potentially discrimination against less healthy folks on criteria entirely unrelated to their job performance. We need more regulation around data ownership before I want my employer giving me a $100 bonus for eating right.

On February 24, 2019, Marissa Dearing commented on Li & Fung Digital Platform: Transforming Global Supply Chain :

The most amazing thing to me about platforms is their ability to aggregate supply and demand so that it’s a win for every party. It’s like diversification – a magical, essentially free benefit to both sides. Scale seems to be absolutely critical to Li & Fung’s success here: by aggregating thousands of manufacturers (suppliers) and thousands of retailers (buyers) on the same platform, LF is able to lower search costs as well as substantially lower transaction costs by, for example, providing volume discounts, enabling more efficient production runs, optimally allocating production across multiple factories, and providing access to technology and digital infrastructure too expensive for a single player, all savings and benefits smaller players would have little access to without a platform like LF. Your comment about LF eliminating the need for retailers and manufacturers to visit tradeshows and physically meet reminds me of how Alibaba dramatically lowers search (and transaction) costs by solving the online purchase trust void. The efficiency gains here are sort of amazing to think about: oversimplifying in this case, instead of each supplier meeting each retailer (100 suppliers x 100 retailers = 10,000 meetings), i.e., what would be required to recreate the network without LF in the middle, just LF meets each supplier and retailer (100 suppliers + 100 retailers = 200 LF meetings), a dramatically smaller number! Plus, from your description, it seems LF has taken advantage of the amount of data on production it’s seen over the years to optimize allocation and planning algorithms, an efficiency gain over and above what the full network of retailers and suppliers could achieve even if those 10,000 meetings did take place. That sort of accumulated expertise benefit extends beyond the planning element to the entire supply chain (e.g., virtual design, as you noted). Truly amazing platform with immense value creation opportunities.

On February 24, 2019, Marissa Dearing commented on Running out of Steam? :

Interesting take on chinks in Steam’s armor. You do a great job outlining Steam’s strengths in addition to its large user base (built up reviews, tailored recommendations, developer tools), but I agree that the threat to Steam is very real. If another platform is able to attract larger games with lower transaction fees, then I could see the smaller games also listing their games on that platform as well, given that the big games could draw enough of a user base to start the network effect flywheel again, especially as the costs of multi-homing seem to be incredibly low for users (downloading an additional desktop application for free, for example). Steam should be very worried about a flight of small developers following the big names… Furthermore, the reviews and recommendations you mentioned as potential switching costs for users seem relatively weak given the ease of looking online for trusted reviews and finding recommendations on blogs or review sites based on games you’ve enjoyed.

I also wonder if there are negative network effects here as Steam has gained such scale – do you think having 781 million games on the platform is a deterrent to game developers? How can a new developer possibly get noticed given the amount of competition? I guess you could say the granular recommendation algorithm might be able to suggest a new game to users based on their past interests and encourage discovery that way, but joining a newer platform and being one of a smaller assortment of games might be attractive to smaller developers.

This kind of platform seems to me very vulnerable to disintermediation by nature… given internet distribution, it’s fairly easy to set up a website and download link and use online marketing to attract beta users and bloggers, and so I’m pretty shocked that Steam is able to charge 30% for a service the developer might do themselves relatively cheaply. I think the existential threat here is a race to the bottom in terms of transaction fees across multiple game distribution platforms as each element of Steam’s value prop is broken up into more specialized pieces (especially for the big money making names): review and recommendation sites, separate game development packages, and separate download sites. I’m actually not sure what Steam could do in response except hold users’ purchases to-date hostage, though that could really spike user rage…