John Franklin's Profile
Interesting insight here, especially on the future of the industry – you really demonstrated an incredible expertise on the industry and a nuanced understanding of the company!
I definitely agree that 3D printing allows for quicker customization of new dispensing mechanisms, which allows a start-up to develop products while limiting capital requirements. However, I question the long term disruption that 3D printing will have in this industry, as there are many other reasons why vending machines are not replaced frequently, besides the high cost to prototype. For example, just the cost of scrapping / replacing an old machine each year and installing a new one could prove prohibitive. Therefore, I wonder if when Centimeo reaches scale, they will be able to innovate as quickly, or be constrained by the finances of their business.
Really interesting topic, thanks for sharing! For a D2C brand, the creative crowd-sourcing, e.g. Lay’s Do Us A Flavor, makes a ton of sense, delivering valuable customer data to the company while allowing consumers to develop a deeper connection with the brand. However, I wonder about the value of crowdsourcing more specific questions, such as “source of protein.” I feel that the issue brands face in discussing these types of specific questions is a biased set of consumers that care enough to answer these questions. The consumers that will take the time to go on innovation websites and give advice to brands are not necessarily indicative of the average consumers – they just might be a “vocal minority.” So, while I think it is create that Lay’s is using this input experimentally, I would be a bit more rigorous when applying this to larger changes in the product line.
Really interesting post! I think that the promise of dating apps (a more efficient way to date) has been marred for a lot of people, as you mention, due to the paradox of choice. It is too easy to see a ton of options, and so people don’t feel the need to commit to one, or even to follow up on one. In that way, this “most compatible” feature that Hinge is launching is super interesting. Given that their is just 1 “most compatible” per person, it really makes it feel like a more serious, deeper connection right off the bat, and their is no other “most compatible” option you could alternatively choose. Therefore, I am not surprised that there is an 8x message rate. However, I wonder about the viability of machine learning to find the “most compatible” option, because I think the data set is inherently flawed. They are using a “click” as a successful interaction, but due to the nature of the app, all of the people that I clicked where people that I did not start a relationship with (or I would no longer be on the app). So, is that really a successful outcome? Therefore, I would think my “most compatible” match would be someone who I was excited to click on, but no more likely to end up having a successful relationship with.
Really interesting post! In regards to your first question (“how can machine learning detect frauds that have never happened before”) I think there are two answers. The first is a simple one – could humans detect frauds that have never happened before? My guess would be that the old system would be no better than the new system in this case, in that both would miss the fraud, so there is no incremental loss in using the machine learning capability. However, I think machine learning actually might be able to detect this new fraud, if the data was still irregular in some way – it might not understand how the data came to be this way, but it could still flag as irregular.
The other question I have is a broader strategic one – how does the use of this tool affect customer satisfaction? And, could it actually become a competitive disadvantage in that so many consumers get flagged as “potential fraud” that they find dealing with Aetna unwieldy, and move their business elsewhere. So, I would be worried about the rate of “false positives” that this method produces.
Really interesting essay – thanks for the good research in putting this together! I think Airbnb is doing a lot of really interesting things with ML, but I would question the level of input data they actually have available to generate quality insights. To point, I would guess that most Airbnb users book 1-3 Airbnb’s per year, if that. How many data points does Airbnb need before they begin to extrapolate the user’s preferences? And, if they extrapolate incorrectly, from an insufficient data set, does that then hurt user retention? So I would just be careful of “over-optimizing” based on a small data set.
However, where this feels very powerful is in the optimization for hosts – there, there is a ton of data to leverage, and by helping their hosts succeed on the site, Airbnb sets themselves up for sustainable success.
Very interesting and well presented summary! I agree wholeheartedly with your point that the Fed should focus on creating the relevant data streams to enable a ML process. This is the main shortfall I see in this potential ML approach – after all, the modern US economy has only existed for ~100 years, and most of the data streams we would have access to today (e.g. consumer credit card purchases) have an even shorter history. So, I wonder if their is enough “training data” to allow ML any confidence interval on predicting what will happen next. However, the Fed should absolutely start collecting as much data as possible today, if only to give more viability to ML in the future.
Really interesting article about a retailer trying to bring innovation to the fashion industry. My main question for the company here would be around their mission to “quantify and articulate what makes fashion cool, and what makes certain designs desirable.” This approach seems to suggest that their is something intrinsic to items of clothing that make people want them. I wonder if this is a backward understanding of fashion – to me, fashion changes are more whimsical than intrinsic. For example, you can have any piece of clothing in the world, and if a celebrity wears it (e.g. Kim Kardashian) people will emulate that. The specific clothing does not have to have any intrinsic value. So, I wonder if using machine learning as a way to jump on trends would be a better use of resources than using machine learning to try to create trends. However, if I am incorrect and ZoZo is correct, this would be a revolutionary understanding of not just the fashion industry, but human psychology too.
Very interesting article – I had not realized Amazon was running their studio this way! I wonder if another reason for the shutdown of this project was the inherent tension of “crowd-sourced” ideas and creativity. This is just my perception, but I would like to see a study done on the sorts of ideas that crowds would come back with in creative spaces, such as TV. I think that creativity is by definition the antithesis of the “popular opinion”, or the widely held opinion, so crowd-sourcing creative content could just end up with a very bland content portfolio. This challenge would primarily relate to examples 2 and 3 above – example 1 is more interesting from my perspective, as they are trying to expand their funnel and bring more creativity into their process.