性视界

Thanks for sharing, Natalie! I also thought about writing about Wayfair (surprise!) and found out in the process that they鈥檙e offering an API for third parties to retrieve and use the firm鈥檚 3D furniture models: . That blog post suggests a few potential applications for the API. Do you think they鈥檙e on the right track, or might they be leaving more valuable opportunities on the table?

Thanks for sharing, Dmitry! Related to Cathy鈥檚 question, I鈥檓 curious to know if you think that Google has built a worthwhile value capture model. Given VR鈥檚 slow adoption, I doubt that there is much money to be made selling Cardboard units. And given that the Cardboard SDK appears to be free and available for iOS development, it isn鈥檛 clear that development tools represent a value capture opportunity. That said, might it be that Google simply shouldn鈥檛 be focused on value capture at this early stage?

Thanks for sharing! I鈥檓 curious to know if you think Google鈥檚 selection process for who would be eligible to purchase Glass might have affected the product鈥檚 trajectory. If I recall, early sales took the form of an 鈥渁pplication鈥� process where prospective users had to write about their interest in the product. I鈥檓 not sure how much that data played into allocation decisions, but if it did, I wonder if Google ended up favoring the 鈥渨rong鈥� sort of early adopters (i.e., glassholes) and would have been better off just selling on a first-come-first-serve basis.

More generally, I think your observation about Google鈥檚 model of giving away ad-supported products raises a bigger issue: is the firm really any good at selling consumer products? They have some experience doing so, like with Google Wifi and Pixel. But given that advertising revenues remain a huge share of revenues, do you think Google is really capable of selling non-ad-supported consumer products, and must they make any major organizational or cultural changes to do so?

Thanks, and glad you enjoyed the slides! Loved your observations about cultural challenges of adopting data-driven management. One way of approaching data-driven change that I have seen be successful in the past is to find particular leverage points or “champions” within the company that who most willing to try new approaches, and pursue open-ended projects that generate new questions that they want answered. At The Times I actually found that the executives who oversaw the “old-fashioned” printed newspaper business were incredibly enthusiastic about applying data to their work, and were really happy to finally have better data that they could use to optimize their work.

Thanks, James! Your results question is a good one. I spent a lot of time there building foundational infrastructure and wish I could have stuck around to see more of the benefits that resulted from it, though I’m optimistic that data has paid dividends in some subscriber retention and operational challenges. But the end results of news applications are harder to assess. Quality reporting is expensive, and absent a lucrative advertising business model it is difficult to say how much data-driven reporting drives subscriptions. My feeling is that The Times’s relative success in attracting digital subscriber revenue is partially driven by its data efforts, but I cannot disentangle that effect from others.

As for the barrier between news and business, I would expect it to always exist in some way. As long as the Times has an advertising business, I would expect strong cultural resistance to full dissolution. And advertising aside, I think that fully breaking down that barrier would require reorganizing the company and putting news operations under the control of someone on the business side. I don’t expect that to ever happen, so I would expect that the company simply continues to drill small holes in the wall and develop working relationships with particular business groups (while excluding others).

Your point about the recommendation engine is a great observation. After all, that’s a severe criticism of Facebook’s role in the news landscape鈥攖hat it has turned confirmation bias into an incredibly lucrative business model. I think the fact that editors continue to oversee Times coverage and that the firm doesn’t rely on algorithmic curation lessens risks associated with the recommendation engine. Of course, the fact that people self-select into reading Times content and/or discover it as a result of echo chambers on social media remains a problem.

Thanks for the great post, Yao. I think that your parting thoughts about art and science are spot-on, and that reducing the creative process to a set of metrics and models is incredibly difficult. Comic book films and established franchises aside, outsized movie successes are often novel creative works for which no comparable prior data exists. I鈥檓 curious: given this and the failure of The Great Wall, do you think Legendary鈥檚 analytics capabilities and marketing knowhow can create truly novel hits, or are they best-suited to harvest interest in pre-existing properties like Batman and Jurassic Park?

Thanks for the great insights! One concern I鈥檝e heard about LinkedIn鈥檚 business model is the firm鈥檚 inability to attract regular engagement from users鈥攖hat people visit sporadically and only do so to perform particular tasks, while firms like Facebook have found ways to attract significantly more casual attention. Given that LinkedIn鈥檚 Talent Solutions offering鈥攁 product reliant more on occasional updates than casual visits鈥攄warfs the firm鈥檚 Marketing Solutions and Premium Subscription business lines, I wonder if LinkedIn is leaving money on the table by not leveraging its data to drive user engagement and enhance the latter offerings.

Great post, Lulu! Your analysis of acceptance decisions and diversity brought to mind the SA鈥檚 Airbnb diversity hackathon, and the challenge of combating racial discrimination on the platform. With that in mind, I鈥檓 curious: do you think there are ways that Airbnb might leverage data and analytics to ensure equitable treatment of guests on its platform.

Thanks for sharing, Meghana! In addition to competition from niche players, I鈥檓 curious to know what you think about Yelp鈥檚 ability to compete with mass-market players like Google (with its rating and discovery features built into Maps) and Facebook (as a marketing platform for local businesses). On Google鈥檚 end, it seems that they have done a good job of collecting review information and exposing it to users, but I don鈥檛 know if many people purposefully use Google Maps for the same local business 鈥渞esearch鈥� tasks that they use Yelp for. (I certainly don鈥檛.)

Thanks for sharing, Jing! I often hear data science described as the intersection of statistics, computer science, and 鈥渄omain knowledge鈥濃�攁 substantive understanding of the field that one is applying the other two tools to. That understanding influences everything from problem formulation to the ultimate applicability of a given machine learning solution (beyond just its accuracy). Both seem to lie outside of the scope of a typical Kaggle competition. I worry that Kaggle might thus inappropriately reduce data science problems to algorithm challenges and prevent companies from fully realizing the field鈥檚 potential.

Have you encountered this concern in your research, and do you think there are ways that Kaggle could better promote involvement of its community in the entire data science process, rather than just predictions?

Great post, Michelle! Related to your and Will鈥檚 observations about the challenges that Reddit faces, I鈥檝e read that the company has struggled to prevent its larger and more 鈥渞ambunctious鈥� subreddits from organizing to comment and vote on posts in others, overwhelming their members and spreading offensive/controversial content beyond its 鈥渉ome鈥� communities. While there are rules against that kind of activity, they鈥檙e enforced haphazardly鈥攅specially in larger subreddits. I鈥檓 curious: do you think that there are restrictions or design features that Reddit could adopt to prevent these incidents, or would they risk compromising the platform鈥檚 sense of democracy and openness by doing so?

Great point about the value of the input data鈥攊t may at least provide insight to Tay鈥檚 designers about how they ought to think about processing it, while bots that are designed to avoid 鈥減roblematic鈥� input may never obtain it, and thus never learn from it.

Also, I think you鈥檙e right that ephemerality could be a promising way to curb abuse of these systems. Further, based on my understanding about how people use platforms like Snapchat, I get the impression that ephemeral messaging tends to produce fairly casual, comfortable conversations. That may be a good way to gather sincere and uninhibited input data from users who might otherwise balk at talking to a machine. And that kind of data may best match Tay鈥檚 intended personality.

Great points, James. The fake news example is a really interesting one that I think a lot of us have been thinking about lately. It seems like such a hard problem to solve, particularly as news veracity is often a pretty broad spectrum: for every blatantly 鈥渇ake鈥� news story, there may be dozens that do things like cite questionable sources, present incomplete information, or apply commentary that some find disagreeable. Identifying fake news often boils down to research exercises and judgment calls, and humans ourselves don鈥檛 always do a great job with those things. And in a similar sense to Libby鈥檚 thoughts on controversy, specific 鈥渇ake鈥� news stories often attract 鈥渓egitimate鈥� meta coverage, which is important for collective truth-seeking.

I鈥檇 like to be optimistic about our chances of meeting these challenges algorithmically; maybe we will eventually. But I suspect that in the near-term we will have to build systems that rely on some explicit rules and human controls, and will have to be humble about what our algorithms can do. Of course, that may introduce yet more bias, leaving us where we started.

This whole thing also brought to mind the recent controversy around Google鈥檚 鈥淔eatured Snippets:鈥� . Even Google, the best company in the world at organizing information and making it accessible and useful, is struggling with this problem.

鈥淭he crowd may be wise but it鈥檚 also a wise-ass.鈥� I鈥檒l buy you coffee if you can fit that into a comment on Wednesday :-). Seriously though, great example of a similar phenomenon that several others seem to have decided to write about. I am glad that the incident gave rise to my favorite open source Google project name, Parsey McParseface: .

That鈥檚 a really good point, Libby. There鈥檚 a LOT of subtlety that goes into communicating about controversial issues, whether we want to discuss them dispassionately or want to express real opinions about them. Humans struggle to converse in person about controversy, and we seem to have even more difficulty doing so in text-only media.

I鈥檓 not too familiar with Zo (having just Googled it now), but starting with a narrower approach sounds appealing. By not trying to learn 鈥済eneral鈥� conversation skills and limiting public interaction, it may avoid the deluge of input that compromised Tay. But it鈥檚 hard to see Zo become a more general-purpose agent if it鈥檚 optimized for such specific interactions and does not solicit more general input (i.e., why would I attempt to converse with Zo about arbitrary topics when I know I won鈥檛 get interesting responses?). I wonder if the best solution is to design bots to learn not just from their own conversations, but from a corpus of others鈥� conversations as well. That data might be expensive to collect and validate if you want to eschew blatantly offensive content, but it may be the best way to avoid constructing agents too broadly or narrowly from the start.

Thanks for sharing, Meili鈥攇reat insights all around! I鈥檇 be curious to know how well ridesharing firms like Uber can forecast demand, either immediately in advance of spikes or several days out. I imagine that the ability to do so, combined with the ability to incentivize drivers to plan their availability around those times, would make passenger transportation more like parcel transformation. But it鈥檚 hard to see those forecasts matching the power of knowing which actual packages need to be delivered. In any case, thanks again!

Great piece, Ellen! Having used a lot of really bad enterprise software, I鈥檝e found Slack鈥檚 model of user-centric product development and its 鈥渟ales鈥� process (direct to users/teams) refreshing and fascinating. Though with the launch of Enterprise Grid (), it seems that they are beginning to build out some more 鈥渢raditional鈥� sales competencies to sign up larger customers.

With that and your final paragraph in mind, I鈥檇 be curious to know how you think Slack might try to enter markets current served by its partners. Would they adopt the same 鈥渂ottom-up鈥� product and sales strategy, or might they leverage sales competencies and relationships with Enterprise Grid customers to go after larger accounts? And if the latter, do you think that designing products for traditional sales models might threaten the qualities that earned Slack its 鈥渃ult-like following鈥� in the first place?

Whoops, pardon the poorly formatted links. If you don’t want to copy and paste them:
–
–

Good questions, Cameron. Overall, I am skeptical about individual sellers avoiding Airbnb, or of the possibility that a similarly 鈥渂road鈥� platform will arise to challenge Airbnb. Given that sellers can set their own prices and vet their customers (in contrast to Uber), I suspect that many high-end individual sellers will ultimately find value in listing on Airbnb. We have seen shades of this in boutique hotels doing so (). And in my response to Adam鈥檚 comment, I discussed a few reasons why I鈥檝e come to think that Airbnb might leverage its renter relationships to perhaps build a durable advantage and achieve market dominance.

That said, I think there are opportunities for others to compete in certain spaces. Christy wrote a great post about Hotel Tonight: /platform-digit/submission/the-procrastinator-gets-the-worm-hoteltonight-creates-a-win-win-for-hotels-and-last-minute-travelers/. At first glance, I would expect Airbnb to fare poorly against Hotel Tonight in 鈥渟pontaneous鈥� bookings, as hotels (many of which remain adversarial toward Airbnb) can more easily accommodate those guests and universally recognize the foregone value of a vacant room. Airbnb鈥檚 鈥渢rust鈥� model also relies on contact and disclosure, which are hard to execute last-minute.

Some alternative platforms have arisen that cater to other specific niches, like long-term accommodations for visiting professors or those on sabbaticals. Such a platform might differentiate itself by length of stay (though Airbnb does compete there), and it leverages existing academic networks to attract participants and establish trust. For example, I would be happy to rent to this guy because he is motivated to protect his established reputation as an esteemed academic: . These can be lucrative businesses and Airbnb will likely not be the only game in town, but I would be surprised if a competitor can truly beat them at their own game.

Thanks for the feedback and for sharing your experience, Adam. I was conflicted about how to discuss Airbnb鈥檚 multi-homing potential absent hard data about how buyers and sellers actually use both platforms, and I wonder if I might have been too quick to dismiss the possibility that multi-homing isn鈥檛 so difficult. Still, I get the impression that there is a significant 鈥渕indshare鈥� gap between Airbnb and its competitors. E.g., AppAnnie shows a rankings gap between Airbnb and its competitors, and Google Trends shows wide gaps in search activity: [https://trends.google.com/trends/explore?q=airbnb,homeaway,vrbo,flipkey,couchsurfing]. I am sure that renters vastly outnumber hosts, so this gap suggests that the former might eschew multi-homing, despite my expectation that doing so would be easier for them than for hosts.

Thinking about the product experience, I wonder if the browsing process discourages renters from using multiple platforms. Selecting a rental property on a platform defined by diverse options, and coordinating with hosts, takes non-trivial effort (especially compared to booking a hotel room). A typical traveler may want to use only one platform for tasks like comparing listings (say, by saving options and sharing them with fellow travelers) and communicating with hosts. This is similar to the way that Amazon has captured retail spending from a lot of consumers: I 鈥渃ould鈥� use Google to find and compare buying options, but Amazon offers a frictionless and trustworthy experience, so I just go there first for everything.

I also wonder how Airbnb鈥檚 recent moves鈥攅xpansion into experiences and (maybe) flights, and purchasing Tilt鈥攎ight affect multi-homing in the future. If Airbnb鈥檚 鈥渏ob鈥� grows from booking unique accommodations to 鈥渂ooking travel鈥� more generally, they may be in a better position to capture exclusive relationships with renters. Sure, a user 鈥渃an鈥� book discrete parts of travel on different services, but I suspect many won鈥檛. And to your observation about 鈥渟ocializing鈥� Airbnb鈥檚 experience, the firm bought Tilt ostensibly to further its group travel initiatives ([https://techcrunch.com/2017/02/22/airbnb-finalizes-deal-to-buy-social-payments-startups-tilt/]). If Airbnb can cultivate direct network effects though these efforts, users might settle on the most popular platform simply because it鈥檚 the one that all of their friends use.

Great post, Meili! I had never heard of Flex, and this is a great summary of how it works.

I think your observation that Flex competes with Uber (and others, like Postmates and Instacart) for drivers is an interesting observation about how 鈥渙n demand鈥� delivery markets work and might evolve in the future. Uber鈥檚 expansion to deliveries (via UberRush) suggests that drivers might be indifferent (drunk riders aside) to particular delivery tasks that they perform鈥攖hat delivering passengers, parcels, prepared food, or groceries are fundamentally the same 鈥渏ob.鈥� With that in mind, I鈥檓 curious: do you think that the end state of this market may be a dominant platform for virtually all 鈥渓ast mile鈥� deliveries? And if so, will that firm control the consumer experience, or is there still room for other firms to 鈥渟it atop鈥� a dominant delivery platform and manage the purchasing process? To put it another way: might we rely on a firm like Amazon or Uber to both order and deliver groceries, or might we place our grocery orders through Instacart, and receive them via Flex or UberRush?

Great analysis, James! I found your iTunes and App Store comparisons interesting considering the ways that digital media and mobile app sales have evolved. 鈥溍� la carte鈥� media sales models like iTunes look like they鈥檙e on their way to replacement by subscription streaming bundles like Spotify, Netflix, and Amazon Prime (not to mention Apple鈥檚 own Apple Music). And mobile app economics and purchasing experiences have led mobile game developers to often give games away and monetize them with advertisements or in-app purchases. Do you think that any of these trends (broadly: cross-developer bundling, all-you-can-eat subscriptions, in-game advertisements, or in-game purchases) represent possible evolutionary opportunities for Steam or the PC gaming industry more generally? (I suppose mods and expansions are sort of similar to in-game purchases.)

Great analysis, Robert! (Though I鈥檓 admittedly biased, having worked for the Times.)

I鈥檝e been impressed with the Times鈥� willingness to focus on the subscription model, and with the latest increases in digital subscriptions and revenue. Though I was struck by how little the 2020 report discussed social media or search engines, which offer significant inbound traffic but have effectively 鈥渃ommoditized鈥� many other newspapers. I鈥檓 curious to know if you have any thoughts on how the Times or other legacy outlets (e.g., the Wall Street Journal) ought to approach their presences on platforms like Facebook and Google in light of the Times鈥� success. They have often made accommodations for those visitors in the past, like lifting paywalls for them. Is there opportunity in reigning in those accommodations, or might news organizations simply end up turning away valuable traffic that they might eventually convert to paid subscriptions.

Loved your thoughts on HBO鈥檚 digital transition鈥擨鈥檓 really impressed by how they鈥檝e managed to adapt in ways that other 鈥渙ld television鈥� firms have not, even if their strategy still has unanswered questions (like Nupur鈥檚). Thinking back to our Samsung discussion, it was striking how fragmented the digital television consumption and content landscape has become. With that in mind, it seems that firms like Netflix and Amazon are doing a great job of leveraging fine-grained viewership data both for content development and user recommendations, enabling them to find content niches to target and to improve their user experiences. I鈥檓 curious to know if you think this approach is a threat to HBO鈥檚 content model or HBO Now鈥檚 viability. Do they need to build out data expertise in order to compete over the long term, or will their existing content development capabilities and 鈥渙ne size fits all鈥� user experience remain competitive?

Great perspective on what has made Amazon so successful over the past 20+ years!

One thing I have found fascinating about their diversification efforts is just how interconnected they are, despite not always appearing so. For example, it isn鈥檛 immediately obvious that a retailer is well-suited to run a cloud computing service. But my understanding is that AWS started as an internal project to standardize and better manage Amazon鈥檚 private server infrastructure. Along the way they realized how compelling it would be to sell access to the system, and so we got AWS.

Relatedly, I have read that Amazon is also very stringent about using APIs for internal communication. Because the e-commerce operations must use APIs to exchange data with Alexa, the Alexa team has a strong incentive to develop and document a robust API. And that API, now open to a proliferation of third-party 鈥渟kills,鈥� has enabled new functionality that strengthens the ecosystem and Alexa鈥檚 value proposition. Following this pattern, Amazon has effectively set up several of its business units as major 鈥渃ustomers鈥� of other units; in so doing, the firm鈥檚 customer obsession has benefited Amazon鈥檚 partners and customers.