Tag: tech

• Homeschooling is one of the fastest growing trends in education

  Earlier this week, Union Square Ventures announced that it was leading a Series A investment in an online education marketplace targeted at K-12 students. The platform is called Outschool, and you can think of it as a form of homeschooling.

  Today, there about 55 million K-12 students in the US, with around 9% enrolled in private schools. Charter schooling is on the rise (somewhere around 3 million students), but so is homeschooling (similarly around 2.5 million students). Data here.

  Homeschooling, at least in the US, largely started within religious groups. But that is starting to change and it is becoming more widely adopted. USV has made a bet that this trend will continue.

  If you look at Outschool’s model, you’ll see that it shares a lot of similarities with other successful internet marketplaces. It is direct-to-consumer (the internet has a way of getting rid of intermediaries). The courses are significantly cheaper than traditional classroom schooling ($10-15 per course). And the supply-side of the marketplace (the teachers) is far more open and accessible to non-traditional participants.

  USV gives the example of a human rights lawyer who is teaching on the platform and now earning more than $10,000 per month in additional income. I’ve never enjoyed online classes, but now that we have reliable video chat, maybe that starts to change.

  In any event, where my mind goes with all of this is the impact on our built environment. We are heading toward more flexible spaces and we are doing a lot more from home.

  2019.05.30

• Uber Movement introduces new Speeds product

  Since we’re on the topic of large-scale data collection, I thought some of you may be interested in Uber Movement‘s new “Speeds” product.

  First launched in 2017, Uber Movement aggregates anonymized data from their ride-sharing business to create data sets and tools that can help cities make better transportation decisions.

  Below is a (hex cluster) map of Toronto showing average travel times from downtown. I dropped the pin at Toronto City Hall. What is shown is the average for all days of the week during the month of January 2018.

  

  Uber Movement’s new Speeds product looks at how specific streets are performing relative to their “free-flow speed.” Uber defines this as “the average speed of traffic in the absence of congestion or other adverse conditions.” (The 85th percentile of all speed values.)

  As of right now, Speeds is only available in 5 cities: New York City, Seattle, Cincinnati, Nairobi, and London. Here is a snapshot of London during the same time period as above, January 2018:

  

  In comparison to what we were talking about yesterday, I have few concerns with the fact that my Uber rides around town have likely contributed to these mappings. With these use cases, the value really only emerges once you aggregate the data.

  2019.05.16

• San Francisco is the first city in the US to ban facial recognition software

  

  San Francisco recently became the first city in the US to ban the use of facial recognition software by city agencies. (There’s a second vote next week, but it is considered just a formality.) A similar ban is also making its way through the system in Boston.

  I thought the following quote by Aaron Peskin in the New York Times was an interesting one, because it speaks to some of the growing tensions between tech, policy, and city building:

  “I think part of San Francisco being the real and perceived headquarters for all things tech also comes with a responsibility for its local legislators,” Mr. Peskin said. “We have an outsize responsibility to regulate the excesses of technology precisely because they are headquartered here.”

  I can appreciate both sides of this argument.

  For those concerned about crime and safety, facial recognition promises more effective policing. That’s why this technology is already used at many airports, including SFO. (Because it’s under federal jurisdiction, it won’t be impacted by this ban.)

  At the same time, there are legitimate concerns related to the large-scale collection of personally identifiable data. And it is this same concern that is fueling the debates here in Toronto around what Sidewalk Labs is up to along the waterfront.

  I am not an expert on this particular topic (or many topics for that matter). But if you’re a regular reader of this blog, you will know that I believe in innovation and I believe in progress.

  However, I also believe that it is important and healthy for us to be having these debates. Because what I do know is that I wouldn’t want Toronto to become Shenzhen. I wouldn’t want to jaywalk across the street and have facial recognition software automatically send a ticket to my phone and post my photo to a “wall of shame.”

  That doesn’t sound like a very fun city.

  Photo by Chris Leipelt on Unsplash

  2019.05.15

• Landed is helping teachers buy homes

  The average salary of a teacher in the United States was approximately $61,730 last year. This can make homeownership in high cost areas a challenge.

  Here is a chart from Curbed:

  

  Landed is trying to solve this problem by offering downpayment assistance to “essential professionals” — starting first with teachers — so that they can buy homes in and near the communities that they serve.

  The way it works is pretty simple.

  They’ll contribute up to half of a traditional 20% downpayment — so 10% of the value of the home — in exchange for a 25% share in any future gains, or losses.

  Put differently, for every 1% that Landed contributes, it takes 2.5% of any future appreciation (or depreciation). However, on an equity basis, they are actually putting up 50% of the required cash (in the maximum scenario) in order to get 25% of any future gains.

  There’s no monthly payment associated with Landed’s money, but it does need to be repaid at the end of 30 years or when the homeowner exits the agreement, whichever comes first. Homeowners are free to repay Landed at any time should they decide to sell the property or they just want to pay them out.

  Landed pitches the service as another version of “the bank of mom and dad.” And for many prospective homeowners, I am sure that it makes all the difference in the world.

  At first glance, it would seem that each homeowner also benefits from a kind of positive leverage. They only put up 50% of the required equity, but they get to enjoy 75% of the potential gains. However, each homeowner is also responsible for 100% of the carrying costs.

  I ran a couple of quick return scenarios, assuming a $500,000 purchase price and a 10 year hold, in order to test whether Landed or the homeowner would receive a higher IRR once the property gets sold.

  I didn’t carry any transaction costs, but I did factor in principal recapture, as well as utilities, insurance, and maintenance.

  My rough numbers suggest that it depends on the annual rate of appreciation. If appreciation stays close to the rate of inflation, it could tip in favor of Landed because they don’t put out any money after t = 0.

  But at higher rates of appreciation, the homeowner starts to benefit from the favorable 75/25 split at the end of the hold period.

  Either way, Landed is providing a service to people who may not otherwise be able to afford to buy a home. That has value. Here’s some more information on how it works, in case you’re interested.

  2019.05.11

• Uber’s seed investors made this much money

  $UBER went public on Friday. Notwithstanding the initial stumble, Uber will go down in history as one of the most lucrative venture capital investments of all time.

  The stock is down from its IPO price of $45 per share, but at that price, the initial seed investment of $510,000 that First Round Capital made back in 2010 was worth about $2.5 billion on Friday.

  Here is a list of some of the other notable investors from Uber’s seed round and what their initial investments grew to over the course of 9 years (chart from the WSJ):

  

  Of course, for every Uber, there are many more failed companies. And for every investor who turns $5,000 into nearly $25 million, there are many more who decided to pass on the opportunity.

  In the case of Uber, many early investors couldn’t see how the product could go mainstream. It initially started upmarket with limousines, which was actually a clever way to hack the chicken-and-egg problem that plagues marketplaces.

  Many also wondered how many metro areas outside of San Francisco had the kind of urban density and supply and demand drivers to support this kind of a service.

  Today, some nine years later and many billionaires later, lots of people — including myself — are still wondering: Will Uber turn out to be a great (i.e. profitable) business? Hindsight is always 20/20.

  2019.05.10

• Redfin is rolling out an online purchase option for homes

  There’s a lot of money at work right now trying to reinvent the way that homes are bought and sold. Perhaps the most popular trend is “instant buying” or algorithmic home buying. I have been writing about this for years, mostly because of Opendoor. But now there are lots of companies competing in this space. With this model, home sellers get the benefit of an almost immediate sale, though usually it’s at a slightly lower price.

  Redfin, on the other hand, is returning to something that it first tried out back in 2006: a buy now button on its online listings. It failed back then. But maybe it was simply too early. The feature allows unrepresented buyers — that is, buyers without an agent — to make online offers. Naturally, it’s far from a single click process. But when accepted, the seller ends up paying about half the amount of commission.

  According to the New York Times, the company started testing the feature in late March in the Boston area. Of the 120 homes listed on Redfin with a “start an offer” button, 5 ended up being purchased via an online bid. That’s more than I would have expected. But Redfin positions these offers as being the stronger option because they save sellers money. There’s also an option to tour the home on your own.

  

  Given this initial response, the company is now working to roll out this feature nationally, market by market. Is this the future of home buying?

  2019.05.08

• One great big exit (not the Brexit kind)

  

  Wired’s oral history of how the London startup scene came to be is a good reminder that, typically, a city needs some great big exits (acquisition or IPO) to really kickstart an ecosystem. In the case of Silicon Valley, you could perhaps trace things back to Fairchild Semiconductor (1950s). But a more recent example of this phenomenon would be the PayPal Mafia, whose members have gone on to found Tesla, LinkedIn, YouTube, and other companies that you may have heard of.

  Put simply: success begets success. When a startup does really well and the founders and employees of that company get rich, it is likely that many will go on to found/fund other successful companies in that same city. In the case of London, that catalytic startup was arguably Skype (at least according to Wired). Microsoft acquired the company in 2011 for $8.5 billion, giving birth to the Skype Mafia. Of course, that wasn’t the only ingredient, but it sure helped (excerpt from Wired):

  Since 2008, according to data compiled by Dealroom.co, the UK has created 60 unicorns (tech companies valued at $1bn or more) – 35 per cent of the 169 created across Europe and Israel. In the past three years, the UK has created more unicorns (25) than France, Germany, the Netherlands and Sweden combined (19). And London has produced 23 unicorns with a combined value of $132bn, compared with Berlin’s eight, worth $32bn.

  The world has changed since Skype was founded. It’s now cool to be doing a startup. But given that every city seems to be trying to establish a thriving startup scene, I think it’s valuable to point out just how important a single big exit can be, not just for the people within the company, but for the broader city. Easier said than done, right?

  Photo by Benjamin Davies on Unsplash

  2019.04.30

• The WeWork of vacation rentals

  

  The word on the street is that Sonder — the marketplace for vacation rentals and competitor to Airbnb — is close to finalizing a $200 million investment round that would value the company at $1 billion.

  I first wrote about Sonder back in 2016 after I met someone from their business development team here in Toronto. I have yet to stay in a Sonder, but I’ve looked at their rentals a few times.

  One of the main differences between Sonder and Airbnb is that the former head leases their rental supply. And they do this by trying to go higher up on the food chain and partner with developers and real estate operators.

  In this regard, they are similar to WeWork. And it allows them to sit somewhere in between Airbnb and a conventional hotel. The supply is distributed, but the service offering is more consistent.

  Of course, this arguably makes their business model slower (they have to negotiate leases) and more costly (they’re committing to fixed costs). So it becomes a question of: How valuable is that consistent service offering?

  Lately when I travel, I’ve been trending more toward hotels, as opposed to Airbnb-like rentals. I like the experiences that many hotels are now focused on creating and I like knowing that if my flight arrives late (in a place like Brazil), I’ll be able to get into my room.

  I guess consistency does matter.

  Photo by Spencer Watson on Unsplash

  2019.04.28

• The artificial intelligence bias

  

  Machine learning is one of the most important trends in tech right now. But like anything new, it naturally raises a number of important questions and concerns. Benedict Evan’s most recent blog post provides a good explanation of what he refers to as the artificial intelligence bias. Here are a couple of excerpts that I found interesting.

  What machine learning does:

  With machine learning, we don’t use hand-written rules to recognise X or Y. Instead, we take a thousand examples of X and a thousand examples of Y, and we get the computer to build a model based on statistical analysis of those examples. Then we can give that model a new data point and it says, with a given degree of accuracy, whether it fits example set X or example set Y. Machine learning uses data to generate a model, rather than a human being writing the model. This produces startlingly good results, particularly for recognition or pattern-finding problems, and this is the reason why the whole tech industry is being remade around machine learning.

  The rub:

  However, there’s a catch. In the real world, your thousand (or hundred thousand, or million) examples of X and Y also contain A, B, J, L, O, R, and P. Those may not be evenly distributed, and they may be prominent enough that the system pays more attention to L and R than it does to X.

  What AI isn’t:

  I often think that the term ‘artificial intelligence’ is deeply unhelpful in conversations like this. It creates the largely false impression that we have actually created, well, intelligence – that we are somehow on a path to HAL 9000 or Skynet – towards something that actually understands. We aren’t.

  The conclusion:

  Hence, it is completely false to say that ‘AI is maths, so it cannot be biased’. But it is equally false to say that ML is ‘inherently biased’. ML finds patterns in data – what patterns depends on the data, and the data is up to us, and what we do with it is up to us. Machine learning is much better at doing certain things than people, just as a dog is much better at finding drugs than people, but you wouldn’t convict someone on a dog’s evidence. And dogs are much more intelligent than any machine learning.

  Photo by Ales Nesetril on Unsplash

  2019.04.24

• Liquidity network effect

  Uber filed its S-1 last week in anticipation of going public in May. The WSJ reported on it, here. These are always interesting documents because you get access to previously private information. Here we can see that Uber’s ride-hailing market share in the US is down to 67% (as of February 2019) from 78% two years earlier. Revenue from this business line — which is the company’s biggest — also seems to have levelled off (chart from the WSJ):

  

  The ride-hailing business today has become a commodity. A lot of people, myself included, simply check to see which service is the cheapest (usually it’s Uber vs. Lyft). So this space feels to me like a giant race to build the biggest network and get to something new, whether that be autonomous vehicles or delivery drones. Uber calls this creating a “liquidity network effect.” Here’s an excerpt from the S-1:

  We have a massive, efficient, and intelligent network consisting of tens of millions of Drivers, consumers, restaurants, shippers, carriers, and dockless e-bikes and e-scooters, as well as underlying data, technology, and shared infrastructure. Our network becomes smarter with every trip. In over 700 cities around the world, our network powers movement at the touch of a button for millions, and we hope eventually billions, of people. We have massive network scale and liquidity, with 1.5 billion Trips and an average wait time of five minutes for a rider to be picked up by a Driver in the quarter ended December 31, 2018. Every node we add to our network increases liquidity, and we intend to continue to add more Drivers, consumers, restaurants, shippers, carriers, and dockless e-bikes and e-scooters. We also hope to add autonomous vehicles, delivery drones, and vertical takeoff and landing vehicles to our network, along with other future innovations. Our strategy is to create the largest network in each market so that we can have the greatest liquidity network effect, which we believe leads to a margin advantage.

  If you’d like to download a full copy of their filing, click here.

  2019.04.14