• Alumni

Activity Feed

On March 21, 2017, kwei commented on Crowdsourcing the Future with Hyperloop :

Think this is an interesting approach to the Hyperloop but ultimately for a project the requires a significant amount of R&D and real capital expenditure to get off the ground. I think you can crowd-source some of the initial design and simulation work as that can be all done digitally and at relatively low cost. But when you get the prototyping stage and more importantly the actual construction stage, who is going to be doing the work and managing the process. I don’t think you can part-time project manage a heavy CapEx project. I like the idea of utilizing a crowd to do the initial ideation, but at some point feels like a real company will have to get involved and take full ownership of the project.

I think this is a novel idea, but I think the decision tree for customers on flights is largely still centered around the cost of the flight. I think the natural process is for customers to consider general buckets of airlines, discount and mainstream, and then compare price across each of those categories. I think most of the frustration around flights happens around delays, which is partially out of the control of the airline; if you happen to have a flight while Boston is hit by a snowstorm, you’re probably out of luck. I think for the rest of the things like legroom, service on the plane, comfort on the seats, I largely associate these with the airline itself, but at the end of the day, is the variation between American legroom and United legroom that big of a deal to make you switch to a more expensive ticket? I think this is a nice function, but unlikely to drive significant switching from platforms like Expedia or Kayak to TripAdvisor for flights.

It is amazing how Wikipedia has created a more accurate encyclopedia than the traditional encyclopedia which existed before. I feel this is the ultimate harnessing of the power of crowds as it would take an immense amount of resources to accumulate the knowledge on Wikipedia with full-time employees as the range of topics and nicheness of topics makes Wikipedia immensely powerful. I think the model works because of people with unique knowledge’s intrinsic desire to share their knowledge and expertise with the world. I actually think if people were compensated for their posts, the model and content might actually be worse as I think people really do take pride in contributing to a free resource for the entire world.

On February 27, 2017, kwei commented on In Airbnb We Trust :

Great post Michal. I think probably the strongest thing that Airbnb has going for its platform, is the fact that travel is many times an international experience (especially considering that Europe is Airbnb’s largest geography); therefore, Airbnb is subject to a global network effect rather than a local one. If we take Uber for example, the reason why Didi can exist and get so large is because most people using Didi probably take 99% of their rides in China; the fact that they can use Uber in Barcelona, Spain or Bogota, Colombia doesn’t really matter as much to them since the large majority of the rides happen locally. Airbnb benefits from the fact that a lot of travel takes place internationally; there are Airbnb “clones” like Tujia in China, which have a much bigger presence domestically in China than Airbnb (although Airbnb is aggressively expanding this year). However, Tujia has very limited international inventory (they are partnered with Homeaway), so if you’re a Chinese citizen trying to go visit San Francisco or Paris, you can’t really use Tujia. Having a big business in Barcelona or Bogota, actually does strengthen your Chinese platform, because travelers are going to want to go there. It would be really inconvenient to have to find a new Airbnb “clone” every time you wanted to visit a different country. On the otherside, to build a competitive platform with Airbnb you’re gonna need to build your supply globally. You have to have scale at hundreds of cities across the world to offer a competitive product to the outbound traveler. This can be done, but would be very hard, and very expensive for somebody to do now that Airbnb has a 7 year head start.

On February 27, 2017, kwei commented on AirBnB moves Beyond Home Booking to Reduce MultiHoming :

Think one of the biggest barriers to multi-homing (or cross-listing as we called it at Airbnb) is the management of the property across multiple listings platforms. If you are renting out an apartment for a few nights a month, it actually is quite annoying to do the calendar management as you probably don’t have a property management system to centralize pricing and calendaring. This means that you have to actually go to Homeaway and block off dates if something gets booked on Airbnb and vice versa. A lot of people don’t really want to deal with this, especially given Airbnbs strong lean towards smaller, urban apartments vs. vacation rentals. This helps a lot with the multi-homing problem, especially because you are likely to pick the platform which gives you the most users, which right now is Airbnb by far. In addition, Homeaway is still undergoing a shift from the classifieds model to the revenue-sharing model of Airbnb; as Daniella mentioned above, if you have good occupancy from Airbnb you probably wouldn’t also pay for a Homeaway subscription.

It’s an interesting questions about vacation-rental places, or “volume-hosts” which have multiple properties under management and may often times be a property manager rather than the owner of the vacation rental itself. There the hosts are more business oriented, and are much more likely to list on a lot of different platforms including Airbnb and VRBO. For the large part, because of the dynamics described above, most of the supply has been proprietary to Airbnb, but this is just not going to be an option if you want vacation rentals, or thinking ahead a substantial amount of boutique hotels on the Airbnb platform. On one hand proprietary supply is a very nice sell, on the other hand its better to have non-proprietary supply than no supply at all (Airbnb is admittedly weak in certain vacation rental markets). What I think is important is having proprietary demand, that customers don’t go to other platforms. But what happens as the market evolves and things such as, a metasearch of homesharing starts to be pronounced?

On February 27, 2017, kwei commented on Craigslist: a platform eroded by platforms :

Very interesting post and interesting platform in that it really hasn’t gone out of its way to change the product, provide any services, or really monetize its strong network effects, yet still has such staying power. It’s not even a wikipedia-like model where it leverages crowds to improve the product. Although certain things like dating or vacation rentals have largely been disrupted by things like okcupid or Airbnb, craigslist remains a popular platform to multi-home as there is really no cost for listing or looking on craigslist.

One thing I still can’t believe, is that we don’t have any better solutions for finding apartment rentals than craigslist. Possibly different in other cities, but I am shocked that right now the best way to find an apartment in San Francisco is still craigslist.

On February 3, 2017, kwei commented on Airbnb: Helping the World Belong Anywhere :

Regulatory and political is certainly a big, and perhaps the biggest, challenge to Airbnb’s business. The biggest issue is the effect of Airbnb on the rental market of a city. Let’s use San Francisco as an example (and this is actually going on), many owners can make more money from renting an apartment on Airbnb than to an actual long-term lease, given the lack of inventory and expensive hotels in SF. This reduces inventory from the available stock of apartments in SF, which then increases the price of rents in SF as the demand is the same but supply is reduced. In an already expensive market, Airbnb is making it so more and more people are priced out of living in SF (to give you context, a 1bd in good area probably is $3500 in SF). So what will the government do? – how can it balance the rights of people to offer short-term rentals with the desire to allow rents to be reasonable so more people can live in SF? This is the exact policy happening in SF right now.

On February 3, 2017, kwei commented on Airbnb: Helping the World Belong Anywhere :

Daniel – so I actually worked at Airbnb this summer and this is definitely something the company is thinking of. The main problem on the platform today is curation – there are 90,000 Airbnbs in Paris, how can we show you the ones that are right for you? There are lots of engineers working on this exact problem.

In terms of segmentation, business travel is one of the segments getting a lot attention. There is work being done to make things “business travel ready” such as allowing for 24-hour check-in, having high speed internet, iron, ironing board, and blow dryer. There is another initiative to develop “Lux” which is for high-end properties (and there are some very nice ones on Airbnb!) and building the associated services around this offering such as concierge, private chef services, etc.

But I agree, this is probably the threat – a notable company is onefinestay which was purchased by Accor (owners of Sofitel, etc.) which tries to provide premium hand-selected properties. Homeaway, perhaps Airbnb’s largest competitor, is focused around the vacation rental market.

On February 3, 2017, kwei commented on So Long, Groupon! :

The Groupon business model is troubled not only from the customer perspective but also from the retailer perspective. The value add of Groupon is it is a way to enable local business to acquire new customers, which is reasonable. If you were a small business, how would you acquire customers? – online ads seem hard, flyering is low yield and likely ineffective. Groupon seems like a good option because you only pay for customers that you gain, and don’t pay for anything that you don’t similar to the way CPC pricing works on Google Ads. This sounds good, but do the economics work? On a $100 item you would offer a 50% discount and sell for $50 of which you would split 50/50 with Groupon so Groupon gets $25 and you get $25. Let’s assume your margin on the item is 10% or $10, so the COGS associated with the item is $90. You just sold $90 item for $25 of profit netting a $65 loss which you must recover. As mentioned by you and Yi, that requires a lot of recurring purchase to make up the lost margin making it really tough for businesses to make the math work. Meituan-Dianping is a Chinese group purchasing company which has been able to be successful in the group purchasing market in Chinese – why? – because it has much lower take rates than Groupon, allowing businesses to be more sustainable.

What else made Groupon compelling? – well it was an expensive form of getting working capital and financing. You get paid by the customer upfront and get cash and don’t have to deliver the goods until later (or potentially never). However, I feel like many retailers misjudged the real cost of this financing as the profit hit from using Groupon is oftentimes too much to overcome.

On February 3, 2017, kwei commented on Could the Rise of Digital Media Spell Doom for the NFL? :


Not sure if you’ve seen it in your research, but Bloomberg Businessweek published a great cover article on this exact topic.

Sports content is so valuable to content providers because it is one of the few remaining entertainment mediums which has a differential value when watched lived vs. watched recorded. Games are just much less exciting when you already know the outcome of the game, and the ease of access of highlight summaries from various sports networks as well as the NFL via its app and YouTube really lowers the value of watching a recorded game. Since SVOD can be consumed on demand while sports still mostly must be consumed live, the competition for viewership I would think would still tilt towards the NFL.

One culprit I do believe is having an impact is the growth of the NFL Redzone, unless my home team is playing in a game, I often find myself watching the Redzone channel instead of actual games; I find similar watching patterns in many of my friends who are casual football fans. It is a great consumer product; however, the expansion of Redzone’s appeal beyond the target market of the hardcore fantasy player has really hurt the overall ability of the league to generate revenue since any gains from increased subscriptions on Redzone are will be canceled out by lost advertising revenue from declining viewership from full games on major networks.

I also agree with Sijia’s point on the lack of compelling storylines explaining some of the NFL’s decline; many games lacked a compelling storyline or were notably one-sided – who really care about seeing the winless Cleveland Browns play a very mediocre Baltimore Ravens team?

I think one proving point for the value of a compelling storyline and good games is the 2016 season of the Dallas Cowboys. 5 of the 6 most viewed games of the 2016 NFL season involved the Dallas Cowboys; the Division Playoff game between the Dallas Cowboys and Aaron Rodgers’ Green Bay Packers was the most watched early-round NFL playoff game of all time with 48.5mm viewers. What made the Dallas Cowboys, who went 4-12 in 2015, to such high viewership numbers this season? – a compelling storyline. First, Dallas is a big market team which is known as “America’s Team” from its previous success. It’s starting quarterback, Tony Romo, went down during the preseason, but it was able to rely on two rookies Dak Prescott and Zeke Elliott to take them to a highly unexpected leading 13-3 record; hate them or love them, everybody wanted to watch to see if these rookies and this team would succeed or finally crash week after week. The Dallas-Green Bay game was full of exciting storylines – would Dallas finally overcome its troubles in playoffs, how would the rookies fair under the bright lights of their first NFL playoff game, would the “Bad Man” Aaron Rodgers, who led to an 8 game win streak after starting 4-6, be able to pull out another win? All these things led to ratings, ratings, ratings. The NFL is still relevant, despite its recent decline and threats from competing programming, but as long as it is able to find compelling storylines, keep relative parity in the league (i.e. see NBA’s Golden State Warriors), and make the most of its rising young stars (Cam Newton, Dak & Zeke, OBJ, etc.) I believe it will continue to be a highly valued and highly watched form of entertainment.

On February 3, 2017, kwei commented on Goldman Sachs: A Bull in the Digital Age :


Very interesting point of view. Certainly the financial regulations have significantly impacted the ability to banks, especially Goldman Sachs, from gaining the attractive ROEs it once did as capital requirements and the reduction of proprietary trading have hindered the ability to take risk within the business. It used to be the case that investment banking revenues could be funneled towards the proprietary trading desks at Goldman, turning an already profitable business into a highly profitable business. Lower capital requirements meant that traders (and sales & trading make no mistake still is what powers GS) could do more with the amount of capital which they were allocated. Goldman is one of the last remaining banks to hang on to some principal investing businesses (most notably, it’s private equity fund Goldman Sachs Capital Partners) but as regulatory pressures continue, these more attractive, but more risky lines of business may also be forced to be divested. These regulatory developments will really put pressure on Goldman Sachs to earn the attractive returns it did in the 1.0, pre-crisis era; luckily, its peer set is also facing the same pressures and time and time again GS has proven its ability to more quickly and intelligently navigate a shifting landscape.

I think the other interesting development at Goldman Sachs is the the introduction of the retail banking unit, Marcus, which will attempt to access a new capital base to bolster its balance sheet and lower its cost of capital.