I left San Francisco to work on my tech startup

For the past 9 years, I've called San Francisco home. For the past 17 years, I've called the USA home. Neither of those things is true anymore.

After an incredible year of travel, I've finally unpacked my bags, moved around the world to Australia, and officially founded a startup. Yup, that's right - I left San Francisco to start a tech company.

Many people have asked whether it's because of a giant orange toddler who shall go unnamed. While I'm mortified and ashamed daily by what I see happening in American politics now, it's not the primary reason. The election results made it easier for me to move away but it's been my home for a long time and leaving was not an easy decision.

While I still consider myself Canadian, I've lived in America since 2000, when I came here for university. I've been fortunate enough to travel to many countries but I've lived my entire adult life in America, working in finance in New York and in tech in San Francisco. If that's not a cliché existence, I don't know what is.

For a long time, I wanted to do something different. I really do believe that innovation and technology can be created anywhere. San Francisco is a special place and has an incredible network and community that I'm proud to call friends. I'll be back regularly (sorry, you can't get rid of me that easily!) but it's time for something new and I'm ready for the challenge.

I'm excited to call Sydney home, and I'm looking forward to showing you what I've been working on. It's a $6+ billion business that hasn't changed in years and has a uniquely Australian bent. From now on, I'll be writing about my journey on the entrepreneurial roller coaster, and my perspective on building a startup in Sydney, after coming from San Francisco.

Being a follower, not a leader

I spent my January 21st watching tennis, and strictly observing a no press policy unless it involved the Australian Open. It was extremely pleasant to stick my head in the sand and simply appreciate some superb athletes demonstrating their best.

But tomorrow is now here and I've spent the morning reading about the protests happening around the world. To be clear, I absolutely admire and respect the idea behind these protests but I doubt whether they actually create change. So, I asked the question on Twitter about what we should do next.

The responses I got were "Protests are effective. Go Google it.", "Run for office" and "Maybe you can start something". While they are valid responses, they weren't what I was looking for and they left me feeling really unsatisfied.

In thinking about it, I realized it's because all of these expect me to be a leader. For most of my life, that's what I've been expected to do and for most of my life, that's exactly what I've wanted to do. So, here's the situation I find myself in:

  1. I want to do something about what I see as a negative turn for the world
  2. I'm working on other things that I find more important on a day-to-day basis
  3. I'm willing to make some time to work on 1, but not so much that it takes away from 2

I'm not someone who has made an effort in the past to be politically engaged. While I'd like to do something now, I cannot make the time now to be a leader in this movement. And it feels like if I'm not a leader, I can't do something to help on an organized level, other than participating in marches. The answer might be that I need to suck it up, make more time and be a leader here. This might be absolutely right but it doesn't feel like the answer for me now.

I definitely think I can still make a personal change by staying informed, finding objective sources of truth and continuing to write and speak about what I see as the truth with proof to back it up. I can and will donate to organizations that are in line with my principles. I will engage and follow more people on social media who represent and lead these causes.

But this time, I'd like to be a follower and find someone who can help me engage within the limited time that I feel I can give.


As an aside, I did spend some time reading about protests and whether they are actually effective. Here are some of the websites that I thought were most useful:

  • https://www.bustle.com/p/do-political-protests-actually-change-anything-29952
  • http://www.nytimes.com/roomfordebate/2012/08/20/what-makes-protest-effective
  • http://www.wupr.org/2016/11/01/want-to-have-an-effective-protest-disrupt-expert-says/

In summary, my opinion is that marches are important as an expression of community but they don't create change on their own. While I think protests can be effective, they still have to be part of a clear plan that focuses on political or corporate change within an existing framework. Protests can spur action, but the results have to be shepherded through. That's the unglamorous work that isn't well reported on by mainstream journalism.

Also, mainstream journalism doesn't do much for protests and protestors. A reasoned and well thought set of demands and follow up actions isn't what tends to be reported. Instead, the news seems to focus on disruptions, drama and irrelevant statistics. They don't focus on the speakers, what the protestors' plan of action is, or what they are actually trying to accomplish.

Why South Park won't help you build a disruptive real estate startup

This post was co-authored with Ruwin Perera (Formerly Softbank, Google, BCG). It's among a series that looks at real-estate startups.

When trying to disrupt a marketplace, most companies use a South Park inspired approach, which consists of three phases:

Phase 1 - Build technology to replace the middleman facilitating the transaction
Phase 2 - ? 
Phase 3 - PROFIT!

In our previous post on the US residential real estate market, we explored the feasibility of this “strategy”. In short, it doesn’t work because agents are so well entrenched that it is irrational for buyers and sellers to eliminate them. While there are companies who continue to approach the marketplace by tackling agent commission, they haven’t seen the exponential growth typical of marketplace disruptors.

Disrupting real estate requires something fundamentally different from merely cutting transaction costs. But how? We have to start with user motivations because the most successful product companies are the ones who meet unmet needs. In this space, there are clear unmet needs relating to creating liquidity for sellers and providing diversification for buyers.

Sellers value a short timeframe and certainty of closing. Having just been through the sale process for our San Francisco properties (a “hot” seller’s market), we can attest to these concerns. The NAR data below shows average days on market, by state:

The time from deciding to sell to finally closing represents very real dollars for a seller. Even in San Francisco, days on market plus closing time will typically exceed two months. In most markets, this time is much longer and carrying costs during sale could easily exceed 2% of a property’s value. 

Even after a house is in contract, there are any number of contingencies that allow the buyer to drop out until almost the last minute, resetting the clock for the seller.  

And of course, that assumes that the home sells at all. Homes don’t always sell, or don’t receive any acceptable offers. In the meantime, a seller is still paying the mortgage, property taxes, and any other related fees on at least one home, if not more. The dollars continue to add up.

Sellers are highly motivated by liquidity. And Opendoor is a great example of a company who builds with this in mind, creating a powerful product and company. 

Their model gives sellers a binding offer on their home within a matter of days, at a price based on comparables. They are fast - a seller could have money in the bank in just a few days. Opendoor then resells these homes using both their own site and traditional channels. In doing so, they eliminate the seller’s agent fee, and could conceivably eliminate the buyer’s fee in the future.

For this service, they charge a 6% fee (which is roughly what a seller would typically pay agents anyway), as well as a 0-6% market risk fee, which they can tweak based on market dynamics. They also finance (and derive the benefit from) staging and light renovations, which can add several percentage points to the value of the home.

The model is capital intensive because they need to purchase the homes outright, and take some market risk. However, they mitigate much of this risk by picking both high-velocity markets (currently in Dallas and Phoenix), and by limiting themselves to mass-market properties that sell in more predictable ways. Assuming they can move the houses within a few months, capital costs can be held to less than a couple of percentage points (in today’s low interest rate environment). 

So, if OpenDoor can charge sellers 6-12%, add a few percentage points of value via renovations and staging, and their costs are a couple of percentage points of interest and a buyer’s agent commission, it follows that an 8%+ gross margin is not inconceivable with this model. 

And it doesn’t end there. On the buyer side, the key unmet need is risk mitigation. Buyers are writing the largest check of their lives, and they are justifiably scared of buying a money pit. In the traditional market, the imperfect solve is a buyers agent and a property inspection. In Opendoor’s case, they provide a “30-day guarantee”, which gives buyers most of their money back if they change their mind. Moreover, they build a brand that stands for quality move-in-ready homes. As they grow in a given market, they could build enough scale that buyers come direct to OpenDoor without agents and their fees.

For investment property owners, there are even greater buyer-side risks. Not only is the investment size significant for most buyers, but it’s also leveraged. A few months of missed rent can make it hard to make ends meet, and a black swan event can bankrupt them. 

What’s more, there’s a huge segment of the broader population that would love to invest in property, but simply don’t have enough cash to make an investment. While some of the more sophisticated ones could invest in REITs, they need to trust the unknown manager’s process  and many investors would prefer to have more control in their property decisions.

There are several new players who want to make it easy to invest in property. Firms like Fundrise, and UK-based Property Partner allow ordinary would-be landlords to invest directly in individual properties and projects while maintaining diversification. They have small minimum investments, and allow investors to crowdfund into multiple property investments online. Fundrise focuses on mid-market projects ($5-100M), and Property Partner focuses on individual dwellings. Property Partner properties are also leveraged, so the investment return profile mimics that of a direct residential investment property, but with added diversification. 

The big challenge for these models today is what happens when a buyer wants to liquidate their investment. Getting the property on the market is far more difficult when hundreds of  investors have a say, each with different investment objectives and holding periods. Not every investor will want to liquidate at the same time, especially when the market turns south. 

We haven’t seen an elegant solve to this problem yet, but Opendoor proves that there’s a market for providing liquidity. At the very least, there’s a banker somewhere that would buy a person’s stake in a property, subject to a liquidity discount. Sounds like a problem waiting to be solved.

In house sales, sale price is treated as the only factor but the reality is that price is comprised of many things for buyers and sellers. So while it’s no easy business, there are huge market opportunities for the companies that truly focus on the unmet needs of buyers and sellers.

The real costs of selling your house

This post was co-authored with Ruwin Perera (Formerly Softbank, Google, BCG). It's the first in a series that looks at real-estate startups.

For most people, buying or selling a home is the most expensive transaction they will ever conduct. After recently putting our Russian Hill, San Francisco home on the market, we keep asking ourselves why selling a home is as expensive as it is.

After all, many of the best online marketplaces use the internet to cut out the (usually less efficient human) middleman. Stocks change hands via online brokers; travel arrangements via OTAs; and even labor via Mechanical Turk. So, it’s hard to fathom why the ~$60 billion residential US real estate market [1] hasn’t been disrupted by the internet. 

Let's start by looking at the primary transaction costs of a sale. They include agent fees, legal and recording fees, title insurance and transfer taxes. And regardless of whether the buyers or sellers notionally “pay”, these costs are baked into the sale price, so the sellers always pay in the end. In the US, these costs can amount to as much as 10% of the closing price.

If we assume that US residential property increases in value at a rate of 3-5% per year [2], and the average American stays in their home for 13 years [3], that 10% transaction cost quickly becomes a sizable chunk of real gains. Of course, that’s without taking into account that most people would be hard pressed to see a 5% CAGR outside of strong real estate markets like New York and San Francisco.

What’s more, this 10% rate seems higher than other culturally and economically similar countries:

Source: Global Property Guide, updated 2014.   Note: this is based on typical values, as many costs and taxes are calculated on a sliding scale

Source: Global Property Guide, updated 2014. 

Note: this is based on typical values, as many costs and taxes are calculated on a sliding scale

That’s not to say that each of the individual fees are excessive. Government transfer taxes are a much higher proportion in some countries. This can be a good thing because these costs are often imposed to dampen speculation and volatility in the market. However, it’s hard to apply that same reasoning to agent fees. Now, we believe agents provide real value and should be well compensated for their work. Our only question is why the agent's share is higher in the US. Here’s the comparison of just the typical agent fees for the same group of countries:

Source: Global Property Guide, updated 2014

Source: Global Property Guide, updated 2014

Is there a model where Americans could pay lower agent fees? We evaluated the cost-cutting approaches, and we couldn't find any that would work successfully in the current market structure.

You can’t cut buyer’s commissions (e.g., exclusive listings, FSBO, REX).

Much of the cost differential above is driven by the fact that the UK and Australia don’t use buyer’s agents. But their role in the US seems assured for the foreseeable future. For one, many US buyers tend to value the service they provide, and two, sellers pay both the seller’s and the buyer’s agent commission, so there's little incentive for buyers to forgo agents.

Buyer’s agents are motivated to show clients properties that they get paid on. So if the commission is cut, buyers agents may simply ghost the property. Why waste time going to open houses, doing market research and supporting a client through the process only to not get paid when they finally make an offer?

And while the sellers might not like paying the fee, they know that the more buyers, the better the chance of a high offer. Today, 87% of buyers are represented by agents. Interestingly, penetration is highest amongst those under the age of 35. It is a very brave seller who will risk losing out on 87% of the market. 

Source:  National Association of Realtors, Accessed 2016

Source: National Association of Realtors, Accessed 2016

You can’t cut seller’s commissions (e.g., Redfin and other discount brokers)

This is one of the biggest transactions in a seller's life, so there is a huge perceived monetary risk in choosing a "bad" agent. For this reason, even discount brokers need to hire the best agents. We think the best agents are those who can provably move the most dollars of real estate. Of course, "best" can be defined in many other ways but a history of high sale prices is extremely compelling to a seller.

Now, there are a lot of agents in the US; the average UK agent closes 40 deals per year but the average US agent closes only 7 deals [4]. However, this is somewhat misleading because the US has many part-time agents who only close 1-2 deals per year. For example, in 2011, part timers made up 43% of all agents [5]. So there are “agents” and then there are agents. For the purposes of this post, we only care about the top tier of the latter group.

Those agents are already well served, as the system today rewards the best. The top 5% producing agents can keep more than 80% of the fees they generate, with less than 20% going to the firm. Put another way, they pocket nearly 2% of the purchase price of the property.

Given these economics, it’s hard for us to see how a salary + bonus + cost reimbursement model (e.g., Redfin) can both make money and attract top agents when their proceeds from the sale are only 1.5%. Redfin passes through the typical buyers commission. 

While we haven’t factored in the costs of being an independent agent, which will decrease net pay, these costs are a smaller percentage of revenue for a top agent vs an average agent. If you believe that agent quality is the key determinant for success, it follows that discount brokerages will have a hard time hiring and keeping the best agents. And if they can't get the best agents, they can't convince the sellers.

So, if you can't just cut commissions on either side, what else might you do? That’s the next post in this series. Stay tuned.


[1] Wall Street Journal | U.S. Existing-Home Sales Rebound in December derived from 2015 units sold and median sale price, assuming 5% commission

[2] Federal Housing Financial Agency Statistics

[3] Eye on Housing | National Association of Home Builders

[4] The Economist | The Great Realtor Rip-off

[5] Redfin | How Much Does a Redfin Agent Earn?

The best feedback is collected through the entire experience

Most marketplace companies, if not all, have some sort of two-sided feedback mechanism in place today. After both sides complete the transaction, they provide a rating for each other.

Normally, this makes sense. After all, it's not exactly fair to rate something without seeing it from start to finish. But, we do this constantly in our daily lives whenever the transaction experience is long. In a two hour movie, most people aren't waiting for the credits to roll before they decide their thumbs up/down rating of the film. Similarly, a book with twenty chapters either grips you in the first few chapters or it doesn't. While you might still go through all the way to the end, your mind has been made up. And by the time you get to the end, it's very possible that you've already forgotten your pointed criticisms or examples of praise. With the constant distractions around us, it's hard to remember those specific details without a conscious effort to track it somewhere.

Moreover, in marketplaces, especially on-demand marketplaces, the rating feature mostly appears to consumers right before they're about to make their next transaction. At this point, motivation is focused on access to the next service, not rating the previous service. If there are any notes, they're generally not top of mind.

In companies like Uber and Lyft, this would mean that a star rating could come up during the trip. To avoid snap judgments or biases, it might require some level of text input or multiple choice selection about the problems or positives. The Kindle already does some related things by allowing you to highlight favorite passages or annotate the book (like margin notes but for feedback).

Another point here is that the feedback aren't necessarily limited to apps or even software. For example, what if a movie theater provided each viewer with a simple device that would allow them to choose their three favorite and last favorite moments through the movie? Or even if this was just done for focus groups as they were in the process of making the films? It could be a new model for iterative film development! Even restaurants with long degustations could allow users to mark their single favorite or least favorite dish, either to turn that into an a la carte option or to remove from the menu.

Ultimately, when it comes to feedback for these extended transactions or services, I'd suggest that we need to think about new mechanisms on the following fronts.

  • deliver feedback on important moments for praise or criticism as they happen
  • direct the user's memory to retrieve critical information
  • present opportunities for feedback without blocking future transactions

What other ideas do you have?