Interesting! A significant portion of the graduating HBS class gives HBS data on its compensation, but I wonder how many would be willing to give that up to an entity with which they have little to no affiliation. However, I think/hope there’s a general trend towards greater transparency in compensation, hopefully this will help. An example, Buffer follows ‘radical transparency’ and makes all salaries public as well as their equity formula, revenue, fundraising details, and on and on: https://buffer.com/transparency
Great post! It strikes me that to predict a candidate’s success, a company like Pymetrics would need as much data or more on the workplace the candidate will be entering, compared to on the candidate him or herself. It seems like candidate success is highly dependent on fit with workplace environment/culture. Do they factor this in?
I’ve heard it argued that we won’t know how WF and Betterment perform until we see them go through a downturn (they’ve really come into existence following 2008) — however, it seems to me that they would be able to keep data from previous downturns as a factor in their predictions, regardless of whether or not they were around before. Probably more of a customer management challenge than a portfolio challenge.
Choosing a name is really, really difficult. One of the major points of failure, here, especially for NASA, is that they tried to use the crowd for convergence on the final name — I think they would have had much more success using the crowd during the divergence stage, to help surface interesting names that NASA’s team would never have thought up on their own.
Crowds can play a big part in converging on a name, too — however, instead of giving them final say, I think they would be more useful in helping to eliminate poor choices. For instance, it’s impossible for a single person to think up all of the different connotations a name might have across a diverse audience — by vetting a name across the crowd, assuming that it’s representative of the future audience, can help avoid selecting a name with unintended (negative) significance.
Great post. I spoke with a CTO over the summer about how they manage knowledge among their engineers, and one of his comments was that “we’ve largely outgrown StackOverflow-able problems.” By that, he meant that most of the problems his team faced on a day-to-day basis were not of the more general kind for which SO is so useful, but rather problems that emerged from company-specific technologies. For example, internally-built system #1 breaking with internally-built system #2. In the end, though, while SO may not be the go-to source for solving engineering problems today in the company, SO has been a critical ingredient in each engineer’s education, and as long as new talent continues to enter the field and technology continues to change, I think the content on SO will remain rich, current and utilized.
Also, checkout companies like CoderBits, which pull in SO profiles to help create programming resumes of sorts.
This brings to mind an interesting proposal for ‘democracy by statistically valid sample’ — http://www.wired.com/2012/05/st_essay_voting/
The argument goes: where topics are complex and the quality of decisions improves with depth of knowledge rather than sample size, we would be better off with a smaller number of highly-knowledgeable subjects than with a larger number of less-knowledgeable subjects (the ‘crowd’). For the US, as the article points out, we would hit statistical validity with a random sample of about 100,000 voters (out of our population of ~300M).
Would you expect, though, that over time such an increase in demand for rides would result in a corresponding increase in the supply of cars/drivers? I might argue that steep subsidies and lower prices, especially if temporary, are a particularly shrewd way to permanently grow the size of the rider market in China. At lower prices, people may make a habit of taking cars rather than public transport. If DD can subsidize this behavior for long enough for it in fact to become engrained as habit, riders may be reluctant to go back to public transport even as prices increase to more sustainable levels. By this time, hopefully the driver supply has increased as well, so that you aren’t waiting so long!
Here’s a potentially interesting thought — if and when Venmo gets businesses to sign on, they should be able to offer substantially lower processing rates than credit card companies, not because of scale, but because of tiny (near zero) expenses from fraudulent charges. One of the reasons why merchants like restaurants and shops pay such high processing fees is because of the high incidence of fraud. It costs credit card companies much more to process ‘swipe’ transactions because it is very easy for a criminal to steal the physical card and use it at one of these locations. On the other hand, if customers pay with Venmo, it should be much more difficult to use stolen phones for these transactions (especially as TouchID rolls out across Venmo’s user base).
Let’s also not forget that after Venmo was acquired by Braintree, Braintree was in turn acquired by Paypal, which already owns widespread relationships with merchants. Maybe they’re just biding their time, after all…
Do you know if Skype has made any effort to attract 3rd party developers as a platform? I’m trying to think about ways to increase the value of its network, and perhaps this could be one avenue. Other companies, most notably Twilio, offer communication (voice + SMS) APIs for developers who want to integrate communication into their own products (for example, when Uber texts you that your car is arriving). Could Skype have reinforced the power of its network by going this route as well?
Fascinating post! What do you think about an incumbent like Charles Schwab [arguably] self-disrupting and offering its own robo-advisor services? It seems like Betterment and Wealthfront are leagues ahead in any event, though. http://www.nerdwallet.com/blog/investing/investments/charles-schwab-intelligent-portfolios-review/
One of the things I like most about Snapchat is that, among digital modes of communication, it most closely mirrors how people communicate in person. When we speak out loud to one another, our communication is ephemeral in the same way messages sent using Snapchat are. As more and more of our communication happens digitally, communicating entirely through non-emphemeral channels feels as strange as committing to publish every word we ever speak.
As important as I believe ephemeral messaging will become, it feels like it could become a necessary feature across most digital messaging platforms, rather than being confined to a separate application altogether. If we believe this is possible, the content discovery side of Snapchat’s user value proposition that you outline takes on even greater importance.
As a quick anecdote, I wrote this post using Quip’s iPhone app — Had I been stationary, I would have defaulted to the desktop app, but I was on the road this weekend and never had great access to a PC. Quip has designed their mobile apps so that the creation process really is just as good on the phone as it is on a laptop — the only advantage of using my PC is my typing speed, everything else is a comparable experience across devices. Combine that with the fact that I can write / edit wherever and whenever inspiration strikes, and I now probably do 50% of my text editing on a phone.