Xamazoners: A question of quantity vs. quality?
Tuesday, Hunter Walk tweeted about the Xoogler Demo Day, an event put on by Xooglerpreneurs, a group bringing “together Google alumni and current Googlers interested in helping ex-Google entrepreneurs.” This site should have a list of presenting companies but… doesn’t yet.
While some may question ‘Why so few successful startups have come out of Google?’, the company has developed a robust alumni network. Ex-Googlers have built notable start-ups (Instagram, Cloudera, Asana, FriendFeed and Factual), become well-respected venture investors (Chris Sacca/Lowercase and Hunter Walk, above, among others), and permeated the startup landscape (per Angel List, there are 5,100 “teams” with former Google employees).
Hold onto that.
Two weeks ago, Amazon released its annual Letter to Shareholders. It led with:
This year, Amazon became the fastest company ever to reach $100 billion in annual sales [and] Amazon Web Services is reaching $10 billion in annual sales … doing so at a pace even faster than Amazon achieved that milestone.
What’s going on here, Jeff Bezos asked. His view: an entrepreneurial culture that allows “tiny seeds” to be planted and then grow quickly into “meaningful and large businesses.” In the letter, Bezos played up culture as one of the principal drivers of the company’s success — broadly — as well as in that of some of its recent products and services. (In doing so, he also responded indirectly to this August 2015 expose in the NY Times, in which the Company’s culture was described as “Darwinian” and “punishing.”)
What’s struck me when reading this was not that Amazon has an entrepreneurial culture: there clearly must be some entrepreneurial magic responsible for this: a $285 billion equity market capitalization, a share price up 11x since January 2009, and the archetype of Marc Andreessen’s observation that “software is eating the world.” As Wal*Mart was to the local/mom & pop businesses of the past, Amazon is to every retail business of the present. Layer in disruptive products and services like AWS, Echo, Kindle, and PrimeNow and the company is an innovative behemoth.
And yet…
Despite similarly high levels of success and innovation, Amazon doesn’t have the same reputation for developing founders or a deep alumni network like peers Google, Apple and Facebook.
Is this observation correct — or fair? (Please let me know if you think I’m wrong.) And if so, why?
Alumni network: What are we measuring?
Well firstly, does Amazon have as deep an alumni network as its peers? Here, there are two considerations: quantity and quality. Quantity = the size of the network as measured in [founders, startups, people in the ecosystem]. Quality, on the other hand = the success of the network as measure by [exits, IPOs, market cap, and/or funds raised]. The former is much easier to gauge and we’ll start there.
Quantity
There’s a lot of data out there. Angel List, Crunchbase, Datafox, Mattermark, Pitchbook, and VentureSource are a few of the many private market business intelligence tools that track the startup landscape. Without access to an API (yet), I manually dug through these in search of answers.
There are ~830,000 companies in the Angel List dataset. What it may lack in historical depth, it likely makes up in contemporary depth: it is as good a snapshot of today’s startup ecosystem (in the US) as any. The dataset can be filtered by ‘team’, which means “one of the team members previously worked at [the company].” Per above, there are ~5,100 former Google employees working at startups, while there are ~6,200 former Microsoft and ~1,700 former Facebook employees working at startups. Amazon? Surprisingly, only 705 companies within Angel List include former Amazon employees on their team. Compared to peers (and even Yahoo) this is really low.
By this metric, Amazon’s alumni network is considerably weaker than peers. I’m also interested in how these companies compare on ‘founders,’ as ‘team’ is just a proxy for ‘founder.’ However, I didn’t find an easy way to aggregate data on ‘founder employment’ using these datasets and rely on the supposition that ‘team’ should be directionally accurate. I’ll revisit when I have access to a more robust dataset and some data engineers.
Could the Amazon data be correct? I looked at start-up data by city. If the data for Seattle was better than that suggested by the Amazon data, perhaps the data was off. Here, I used both Angel List and VentureSource (whose data goes back much further). In both datasets, Seattle significantly under-indexed as the location for startups. As you can see below, in the Angel List dataset, ~1,700 startups were based in Seattle (vs the average of 9,591),but more importantly, Seattle had far fewer startups than other metro areas in the Angel List dataset and fewer than even San Diego (!).
On the basis of quantity, Amazon’s alumni network in startups is considerably shallower than peers. Should this be?
Quantity: how does it arise?
How do alumni networks develop? A company is founded, founders and early employees play a role in finding product-market fit, they are compensated in stock, the company is successful, there is a monetization event, and these people now have significant liquidity which enables two things: they can leverage their own experience finding product-market fit and some financial flexibility into new ventures and/or they can invest some new-found liquidity into their peers’ new ventures.
Should Amazon have developed a robust alumni network?
- Product-market fit
In answering ‘why so few successful startups have come out of Google?’, David King (now at Blippy) says:
I think there is another way to phrase this which is primarily a more general observation of the same thing. Most Googlers joined the company long after the product/market fit was achieved and Google’s end-user PMF didn’t significantly change over time as the revenue model was discovered.
And related, Keith Rabois says:
In addition, after Google became successful, the type of candidate who applied and was hired shifted from the entrepreneurial to the smart yet homogeneous type.
The key from an experience perspective is helping develop a product that finds market. Amazon was founded in 1994, and you’d think product-market fit was found long ago and employees today are just keeping the wheels on the bus. But the way Bezos describes how new products and services are developed, suggests there are quite a few people engaged in finding product-market fit in 2016. For example, Bezos description of the inception of PrimeNow sheds light on an entrepreneurial process:
Prime Now offers members one-hour delivery on an important subset of selection, and was launched only 111 days after it was dreamed up. In that time, a small team built a customer-facing app, secured a location for an urban warehouse, determined which 25,000 items to sell, got those items stocked, recruited and on-boarded new staff, tested, iterated, designed new software for internal use — both a warehouse management system and a driver-facing app — and launched in time for the holidays. Today, just 15 months after that first city launch, Prime Now is serving members in more than 30 cities around the world.
And a customer development example from Amazon Taktal:
Last year we ran a program called Amazon Chai Cart where we deployed three-wheeled mobile carts to navigate in a city’s business districts, serve tea, water and lemon juice to small business owners and teach them about selling online. In a period of four months, the team traveled 15,280 km across 31 cities, served 37,200 cups of tea and engaged with over 10,000 sellers. Through this program and other conversations with sellers, we found out there was a lot of interest in selling online, but that sellers struggled with the belief that the process was time-consuming, tedious and complex. So, we invented Amazon Tatkal, which enables small businesses to get online in less than 60 minutes.
These are just two examples. More: Fire phone (failure), AWS (massive), Echo (I have one), Kindle (a success), Auctions (didn’t work), third party stores (crushed eBay).
Check. There should be a large number of people at Amazon who were there when product-market fit was found for various products and services.
2. Compensated in stock.
As it relates to stock compensation, simply, there are two levers: you get compensated in stock and it goes up in value. Here’s a back of the envelop comparison between Amazon, Google and Facebook:
Back of the envelope conclusion:
- A similar stock comp profile to Google
- If the average Amazon corporate employee is getting $90K in stock comp, using the “law” of compensation averages (people in position of authority get multiples of the average), there should be Amazon employees with liquidity to to support new ventures, theirs or peers.
Check. People are getting equity at Amazon.
3. The company is successful
See above for share performance since 2012 (+152%) and since 2009 (+1000%).
Check. The company (as judged by its stock) is a success.
4. A monetization event
With Amazon, subject to vesting schedules, given a liquid market for the stock, there is a daily monetization event. Check.
Conclusion: While the data suggests it doesn’t exist, the pieces are in place for a robust alumni network as measured by quantity.
Quality
So far we’ve dealt with quantity. When it comes to the quality of startups coming out of Amazon and Seattle, it’s not as cut and dried. Yesterday, Sandi MacPherson published a piece on Medium in which she noted the ascendancy of NYC as a startup hub (based on rankings in this Compass report) and “Seattle’s decline as a world-class city for tech and startups.” Though, I’m from Seattle (though I live in NYC), I don’t bring this up to quibble or criticize, but because her post elucidated a passionate response from Tren Griffin (who’s blog is fantastic). He asked:
“Have you ever actually been to Seattle? Oracle, Alibaba, Dropbox, HP, Expedia, Tableau, Zillow, GOOG, Facebook, AMZN, MSFT, Adobe”
and
“MSFT, AMZN, Starbucks, F5, Getty, Avvo, Seattle Genetics, Zymogenetics, Redfin, Moz, Real, Classmates, etc all founded in Seattle.”
and
“What exactly are the NY success stories? Story is a bunch of hand waving. Flailing? WTF. Tableau,, Zillow, Zulily, Concur, Juno”
Good points. There are a lot of Seattle startup success stories. The story of modern Seattle is basically a story of how startups (Microsoft, Starbucks, biotechnology) and their offspring can turn a timber and aerospace backwater (which I loved) into one of the country’s most vibrant metro areas.
Vis-a-vis Amazon, there are at least 3 very successful startups founded by Amazon alumni:
- Instacart, which has raised $275MM. Founder Apoorva Mehta worked on on Amazon’s Fulfillment Engine for 2 years before applying to Y Combinator and launching Instacart. He also has a BASc, Electrical Engineering from the University of Waterloo, which itself has developed a reputation for developing smart technologists and startups (Kik, Pebble, Chamath Palihapitiya, etc).
- Flipkart, which has raised $3.2BN, founded by the Bansal brothers, Sachin and Binny, who previously were a Senior Software Engineer at Amazon Web Services and a Software Engineer at Amazon, respectively.
- Twilio, which has raised $234MM, founded by former Amazon Technical Manager (for a year and a half), Jeff Lawson.
When compared the Google successes (Instagram, Cloudera, Factual, etc), these three companies stack up quite favorably (although worth noting only Instagram has a monetization event). For Seattle and Amazon, there’s a disconnect between the quantity data in Angel List and VentureSource and the quality examples. Does the city and the company lose on quantity but win on quality? It’s more nuanced.
Quality vs quantity?
When Tren asks “What exactly are the NY success stories? Story is a bunch of hand waving” he underscores that when it comes to monetization events, NYC success stories are fewer and far between. That Etsy is NYC’s biggest startup exit ever isn’t really something to hang a hat on.
I don’t think it’s as simple as quality > quantity. There are two separate arguments: one is what happened in the past and one is what is expected to happen in the future. Most startups fail. A robust ecosystem should increase the chances of future success (best practices are passed down, fundraising is more available, etc.). While Amazon and Seattle both have historical success, the considerably lower level of current startup activity isn’t suggestive of similar outperformance in the future. For people investing in startups — or going to work at one — this is an important consideration.
But why?
Based on the data from Angel List and Venture Source, Amazon doesn’t seem to have as deep an alumni network as that of peers like Google, Apple and Facebook, despite having many of the pieces in place for one.
Why is this? I’m out of room, and will consider this at another time.
