So Fove is awesome. Great post. HTC Vive also just announced a $220 attachment that does eye tracking.
My worry about this kind of stuff (higher end hardware, peripheral controllers and sensors, etc.) from a business standpoint though is that it gets commoditized fairly quickly. At some point, being higher resolution than Oculus and HTC Vive means you’re just getting imperceptible improvements.
On going open, I’m not sure that’s always the best idea. For example, Apple (closed) sells fewer phones than Samsung (open), but Apple takes 90% of the smartphone profits. Samsung gets some, but the likes of LG (also open) basically make a penny of profit for each phone.
It seems like being on the open platform works well depending on your role in the ecosystem. For example, Android loves the open platform, but LG doesn’t get a ton out of it, right? It seems like there are certain roles (hardware, operating system, content creator, distributor [app store], etc.), and depending on your role, you might prefer being open or closed. On top of that, early on, when there aren’t a bunch of partners in the ecosystem, you sort of have to do it yourself, right? Being open works once there are critical numbers of partners to join you.
A few thoughts on Blippar: a lot of their “blips” are just information fed in through the Wikipedia API. The issue is, how many people know something / some place is “blippable?” Most of their other blips are just user-generated (like Wikipedia). This crowd-sourcing process is slow, so refining the computer vision algorithms you describe will be critical to scaling. Lastly, the cool thing about a computer-vision-enabled product like Blippar would be that people who are illiterate could suddenly use the internet to search for and learn about things (e.g., if you saw an apple but didn’t know what it’s called or how to spell, you could still learn all about apples).
I’m particularly excited about this HMD, relative to others. I think the enterprise AR focus is spot on for Microsoft’s capabilities, and I think AR fundamentally a better solution for collaborative business problems than VR. Microsoft seems to have found a sweet spot where its capabilities, the technology’s unique abilities, and the business users’ needs are intersecting quite nicely.
I wonder, though, if trying to leverage Windows will become a liability in the long term. At some point, it seems like a prolonged strategy of “leveraging what we built 20 years ago” could become more of a liability than a strength.
This is cool – thanks for the post. It seems like a wide variety of technical trades could benefit from their own industry version of ScopeAR (e.g., anything from surgery to plumbing or fixing cars). In each of these cases, it seems like what’s going on is that AR visualizes and makes available information one would otherwise have to painstakingly learn, memorize, and recall. As a result of that, we don’t need to memorize as much — we just need to be able to follow instructions. All of a sudden, with AR, someone with far less experience or training becomes empowered to do something that was formerly only possible for the most experienced, knowledgeable, and trained tradesmen. AR democratizes many of the technical skillsets that formerly required such high levels of training and memorization.
re: “Going forward, I believe NextVR needs to find a niche and anchor itself to that area.”
I think the answer is NBA fans in China and European soccer fans around the world. One of the powers of VR is that it basically makes “seating capacity” infinite. The problem is, if you’re wanting the experience of being at the stadium, the VR stadium is less compelling for people who could actually get to the stadium and be there in person. On the other hand, selling “seats” to people in China truly taps a new market; they have millions of fans who would like to feel courtside, but rarely if ever get the chance to see their favorite players and teams live. NextVR could uniquely solve that problem for those customers.
Hey Alex, thanks for the questions.
1. Seriously, I have no idea. Something like that, if true, would certainly create issues, especially if the trend didn’t continue. If the trend continued, I presume regressions could enable one to reasonably adjust probabilities for increasingly inclement weather.
2. Definitely yes. There’s definitely an adverse selection problem. And when that winter storm hits, Freebird gets in trouble. I actually just learned more about this today. The company initially offered refunds to people it couldn’t rebook. Then regulators caught wind of it and said that makes it a proper insurance product, and made the company file as an insurer (costly, complicated, blows things up) or cease the refunds. So now, Freebird just has to tell its customers that it’s holding the upfront revenue… uh oh.
Thanks for the questions, Lidiya.
Agreed, G2M / distribution are a big question mark / challenge for them, as noted in the post. Cracking this will be critical.
I would think some sort of distribution agreement with some of the providers you mention would be critical to scaling — otherwise, marketing expenses seem like they could just get crazy. I would guess the airlines would hesitate to partner or just try to eat their lunch, and the rev share just wouldn’t be that enticing, given this seems (to me at least) like it’s not actually a huge market. I don’t know the click-through rates on travel insurance today, but I imagine they’re below 25%, and I think most people feel like they’re already spending plenty by the time they buy a ticket. On top of that, I think a large portion of travel is done in situations where a delay or cancellation is a hassle, but not an absolute game changer (Freebird may be less compelling outside the latter case). All of this just means the # of potential buyers is relatively low, and increases the importance of casting a wide net through distribution agreements to reach them.
Wow. This is super interesting.
I know some businesses (e.g., VidAngel) that have developed algorithms for editing edit movies that basically tag / catalogue every moment of a movie so consumers can edit out words / scenes etc. they don’t want. I wouldn’t be surprised if movie-making could eventually get to the same level of granularity, with predictions of box office outcomes based on how any moment, scene, or word from a script will affect sales. Imagine how formulaic the creative process could become! Data scientists could be the next directors…
This is a fascinating post. The opacity of medical expenses is such a problem.
Any idea how Amino gets that data? Is it something they pay for? You mentioned it has been created each year for a long time — sounds like some organization is aggregating all these insurance claims? I had no idea this was going on. Fascinating.
Also, you mention monetizing much later. Are they doing anything now? e.g., referral fees?
Super cool post, Yezi.
When you mentioned that few parties are really making use of the data, it made me think about how powerful it could be to have cities investing in IoT / AoT infrastructure to provide a data platform for private sector companies to build off of. Imagine if a city collected all that data and just let people innovate off of it. Then we’d be leveraging the power of data AND crowds (boom).
Well said, Ali. I’m with you on this one.
I love your point about how Nate Silver built a moat around his media business by doing analysis others would not be able or willing to do. So interesting.
The idea that you can be wrong in a public fashion seems like it shouldn’t be such a big deal, though. Is the problem the way Nate Silver didn’t share data, or is it just that we are misunderstanding the way the statistics work? There must be so few models that are actually perfectly predictive, right? Do we really expect anyone to get it exactly right 100% of the time?
Thanks for the comment, Brandon.
I think, at least for now, massive volumes of images may be the critical factor here and in identifying other industry applications. As long as the data is already in numerical form, computers seem like they’ll just dominate. The same increasingly holds true with language, as natural-language processing improves. Of course, “computer vision” is improving, as well, and is going to get lots of investment as part of the VR / AR innovation wave. That may make humans playing games to identify or map images less useful. But for now, perhaps industries or services that have huge image libraries could benefit.
For example, I could imagine a game that allows users to help spot shoplifters in store security footage. Computers would struggle to do this, but we pay guards, and they simply can’t watch all the screens / cameras.
Maybe that’s a silly / not valuable example, but I think it illustrates what might be the underlying pattern that determines where gamified data analysis works.
Thanks for the comment, Ophelia. So they’re actually quite up-front about the fact that users are helping with cancer research — that’s one of the main draws to play. Some of the games, as mentioned, are also very explicitly scientific. For example, as an alternative to blasting asteroids, you can play other games that require you to draw lines or dots to identify certain parts of cells, and you’re just looking at an image of a real sample cell. People just get the “high” off of pattern recognition (e.g., learning to quickly spot the cancer cells vs. non-cancer cells).
Thanks for the question, Sonali, and glad you liked the post. So they currently pay to use traditional marketing tactics to get game players, and the money comes from donated funds (obviously not the ideal use of funds for the foundation).
In terms of monetization, the challenge is that you need cash to grow and analyze more data. If CRUK really tried to scale as they operate today (develop and market their own games through these crowds / hackathons), they’d run out of donated funds (due to marketing expenses, to your previous point). I read some more today, and CRUK is actually going down the path of basically giving away access to its scientific data so other developers can build and market the games, using the data and general positive cancer-focused PR to boost user adoption.
Ancestry is awesome. Anybody who hasn’t tried genealogy should definitely check it out — it’s amazing. I was able to find photos and stories about my grandfather, whom I never got to meet, and it was absolutely incredible.
I delete or unsubscribe from most services’ “hints” and other e-mails, but with Ancestry’s, I feel like I HAVE to look. “You mean you found someone I’m RELATED TO?!” It’s crazy.
I also think your thoughts on the potential for DNA are really interesting — e.g., filling in gaps and connecting people where genealogical records have been destroyed. So far, my friends who’ve tried AncestryDNA haven’t been thrilled with it, but the potential seems undeniable.
Super cool post, great company with an awesome mission, and powerful example of truly connecting the crowd.
re: “I believe that Quirky fell into a dangerous ‘crowdsourcing trap’ by assuming a positive response from the crowd was an accurate indication of market demand”
This was a theme in the Frito Lays post, too, as the flavors people voted for ended up being commercial flops after launch. You’re definitely on to something.
Clearly votes, upvotes, likes, and potentially even crowdfunding aren’t super reliable predictors of success and demand. The latter may be the most predictive of the short list here (no research or data on hand here to support or debunk that assumption), but, in general, how do you think we get around this trap? You mention “more thoroughly assessing the market.” Does that mean something different than paying someone to go talk to / interview the crowd in focus groups? Is it anthropological research (e.g., following and observing customers to try to see what they don’t realize about their own behavior)?
I just read this one after the Frito Lays one. It seems like Frito Lays has people vote on products but many of the winners have ended up being flops upon launch. Lego seems to have achieved different outcomes. What do you think is the secret here? Is it the review process?
I also like the 1% of sales economics. Most crowdsourcing efforts I’ve seen are all about competitions and winning prizes, instead of setting it up like a venture that the crowd actually gets to participate in. It obviously may cost more, but I could see people being more excited and incentivized, given the revenue sharing.
Garlic Cheesy Bread and Chicken & Waffles were totally delicious.
Your point about voting not being indicative of purchase intent is an important one that I’ll remember. It reminds me from our class discussions of why we should look for and talk to superusers — they demonstrate real commitment and even innovate their own solutions without our help. That visible, already-expended effort is a real demonstration of commitment to the idea and the magnitude of the unsolved pain point.
The broader idea you address here — is it really to develop products or just a marketing ploy — is also an interesting one. I wonder if participating in contests is more a symptom or driver of brand loyalty. It likely has both effects, but I wonder which dominates.
Oh man. Totally forgot about this. What an epic and hilarious failure.
The point about vetting is a really interesting one. Do you really want ideas from “the crowd,” or do you just want ideas from a relevant audience that’s external, new, broader than you’re used to? It seems like the critical question on vetting is what criteria or dimensions you use to vet participants and ideas. It’s not hard to imagine coming up with criteria that basically cut out the value of crowdsourcing in the first place.
I also wonder if there are optimal sizes / limits for crowds. For example, is 50 ideas as good or better than 100? Do you get all the value you need from about 20? Or how does the optimal size of the group vary based on what you’re trying to do?
AirBnb is a super interesting one, obviously. Really liked your thoughts on the quality / reputation management bit. I mentioned this on Alice’s post as well (see UpWork): the idea that building a platform is simply a race for numbers is an oversimplification, and I think you’ve articulated nicely the hazard of taking any user or complementary provider you can get.
Sort of related to Cam’s question above, but AirBnb to me seems like the perfect candidate for high multi-homing on both the supplier and the user side. For example, when I go to book a vacation, at least for a group (thinking of a specific recent experience), I see the same properties on AirBnb, VRBO, Homeaway (obvious connection of ownership here with VRBO), and sometimes even OneFineStay. Where AirBnb gets differentiated listings seems to be on the single-room sharing.
I also feel like there are bound to be interesting opportunities to socialize AirBnb — e.g., when I’m going to Chicago, AirBnb should notify me if my friends or friends of friends are hosts, or it may notify friends that I’m looking, etc. That could build even more trust in the interactions.
Great post, Alice. Thanks for this.
Really liked the insight about the hazards / risks of not vetting who you let onto the platform. We’ve talked at length in class about how important it is to get people on both sides of the platform fast, but the quality / type of participant on the platform is, as you’ve shown, another critical consideration.
The principle applies more broadly, too — e.g., Uber and Lyft need driver AND rider ratings for things to work and people to be happy. Same with AirBnb. Same with Amazon seller ratings. The list goes on…
Really interesting post, and obviously a headline-making story of platform growth. Thanks for sharing this.
I hadn’t thought much about the risks of being a platform partner, in this case. The stories of early SaaS integrations that I had heard and thought about were all more along the lines of the positive ones you mentioned, but the Bluejeans story is a really powerful example of how dangerous it is to be a tool or feature plugging into your customers’ a platform.
The $80M developer fund is also an eye-opening example of what it might take to achieve Slack-like growth. That is a ton of money to offer developers when you’re a 1-2 year-old company. My takeaway is similar to what I took away from our class simulation: if you want to win the platform game, be ready to raise and burn a nice chunk of change.
Super interesting post. Thanks for supplying news on the Priceline write-downs, as well.
So, it seems like there are real indirect network effects here — more users would get more restaurants to want to sign up, and more restaurants would get more users. Why do you think OpenTable wasn’t able to effectively capitalize on those? Was it that they were too geographically diffuse in their customer / restaurant acquisition, so they never got critical density in key markets? Or perhaps as you may be suggesting they got to focused on building out other features to try to compete head-to-head with Yelp / Zagat?
A takeaway for me is that the acquisition in the UK was probably a smart idea, but charging the restaurants early on left room for OpenTable to be undercut.
Perhaps in the hands of Priceline, OpenTable could — like its competitors — be subsidized by other parts of the business and grow more rapidly.
Really interesting. This seems to me like a space where there will only be room for a few winners. Network effects, as you describe them, seem high, especially as the business is in earlier stages of growth (e.g., there may be diminishing returns to incremental users as the coaching AI gets really great). There are, however, some strong reasons for multi-homing for both job seekers and recruiters. Together, these forces would seem to imply a limited-but-plural set of winners.
Given that, I wonder if the future of HireVue will extend the platform into other parts of the recruiting and onboarding processes, or if it will remain more of a candidate-acquisition-focused tool that plugs into other HR platforms. For example, today, it seems like recruiters may use HireVue in conjunction with tools like Namely or Jazz to post jobs and get applications, and then of course use HireVue for the initial assessments, as you describe. Subsequently, recruiters then then transition to other software such as Gusto, BambooHR, Workday, ADP etc. depending on the size of the business, to do things like collect onboarding data, select benefits packages, execute payroll, manage time off, and act as a core system of record.
Within this context, I could see HireVue’s decisions about API integrations and product extensions being especially critical to their ability to really achieve the benefits of being a core HR platform.
Also, on your point about value creation, I think much of what you wrote could basically be turned into “the guide to building a customer value proposition as an API business.” It’s all about ease of use, reliability, security, and making the developers’ jobs easy. What Stripe has done to continue creating more value — e.g., building APIs for the front-end — is super interesting. API entrepreneurs will need to think creatively about how deeply and broadly they can build for their customers.
Stripe is an awesome company. I was lucky to have the chance to meet the founder last year, and I was super, super impressed by his vision.
More than anything, he impressed upon me the value and power of third-party APIs like Stripe’s. These API businesses are truly building the new infrastructure that is going to change the way digital businesses are built. Already, we can see that the components required to build a business online are increasingly available off the shelf — e.g., Algolia for site search, Stripe for payments, Shippo for shipping, Twilio for telephony, etc. Companies like these are contributing in a HUGE way to the digital transformation we’ve discussed in class. They are one of the reasons why it’s so much cheaper and faster to test an idea / startup in the modern economy. And with every new API, it’s just getting easier.
Still laughing about “Johnny’s growth spurt.” Nice touch.
In all seriousness, this is really interesting. It’s a perfect, under-the-radar example of how something as simple as a mobile device and cloud-based order forms (I’m assuming these are not “high” on the software sophistication scale, but maybe I’m wrong) can change the structure of the industry.
It’s particularly interesting to me that they buy out the local distributors, instead of just selling them the software platform. Vertically integrating the field sales, distribution, and software businesses seems like it could drive operational efficiency, but it seems like the ecosystem could also work well with a massive software player serving the highly fragmented local sales and distribution businesses. Given the software you’ve described, and the “flash store” e-commerce tools they offer for teams to sell their merchandise, why wouldn’t BSN try to be a pure software player that just sells digital services? Are margins on the merchandise itself really healthy?
In any case, given your endnote, it sounds like a fantastic investment.
Really like the post, Kyla, and very much agree with your diagnosis. I think your point about how value creation and capture have had to change in NYT 2.0 is spot on. In some respects, the news business hasn’t changed — it’s still all about getting people to pay and read so there’s an audience for advertisers. BUT, like you say, the way we as consumers want to get the news, and what we do while we read the news is very different than it used to be.
I could actually see a world where the ongoing digital transformation of news means it is simply no longer delivered by large organizations. Instead, as you say, bloggers / individual contributors will rule the world (an overstatement of what you said). Companies (maybe NYT 2.0 / 3.0) will just be masters of UI / UX and they’ll fight with each other to partner with individuals who publish content and build followings, or they’ll compete to provide the best tools to those individual publishers. Perhaps news’s next step could be a transition into business models like a Getty Images (which competes to partner exclusively with talented individual photographers and offer their content with [really unfair] revenue-sharing agreements).