I intend to shift most (not all) of my focus back over to consumer investing in Haystack 3.
In my first fund, it was dominated by consumer. That wasn’t intentional. I was just starting out, I was a dog chasing cars. In my second fund, most of it was consumer, but I started to go a bit deeper into enterprise SaaS and industrial IoT. That exploration led me to start focusing on enterprise and industrial IoT in my third fund, which I’m currently in the middle of. In this current fund, I have focused my investing in two areas so far — enterprise SaaS, security, and infrastructure, and industrial-focused software and robotics. I’m finally able to write more about these so expect some “The Story Behind My Investment In _____…” posts over the next few weeks.
Along this way, in Haystack 3, I have been looking for consumer-focused companies but have ultimately passed on all the opportunities to date. Those were hard decisions, and I’m sure I made some mistakes. So far, in 8-9 months of investing Fund 3, I have only invested in one (1) consumer-facing seed-stage company, and it’s not yet launched and may shift its model to an indirect B2C2B model.
I would like to do more on direct consumer in this fund, and it’s been nagging at me for a while, so I finally wanted to post on it. I do not have a laundry list of categories to hunt down or chase after, and I have a small handful of ideas of where I think interesting consumer behaviors may emerge, but I’d rather see what’s out there and be surprised. Yes, I have thought about areas of consumer spending (insurance, rent/mortgage, self-improvement, health etc.) and consumer attention (VR, AppleTV, apps etc.) and I’ve written about live video and esports and bots… but I would say the pattern I’m looking for is as follows: Something I can experience/test myself or observe others doing; a company which is obsessed about creating and building a direct relationship with a consumer; a team that is obsessive about acquiring customers and users and their CAC numbers; and a team that has a vision for a future on a global scale — doesn’t need to be today, but eventually. A product vision and desired roadmap goes a long way.
I am going to focus on this now more intently and really dig into it over the summer. It’s OK if you’ve already raised a pre-seed or seed or whatever fund. And, I likely won’t make any investment decisions very quickly on this, so am looking to engage more and get to know more people before selecting a few to work with. I’d appreciate it if you could share this with people you like or find interesting in the consumer space, and thank you for reading.
The State of California recently unveiled a plan to increase minimum wage in the state to $15/hour by 2023. Currently, America’s national minimum wage is $7.25/hour, which hasn’t increased since 2009. The concept of minimum wage is complicated, and I don’t want this post to be focused what is fair or isn’t about such an increase. I certainly don’t condone workers having to see most of their take-home pay go to rent and most of their free time eaten up by commuting. That said, I do sense aggressive increases to minimum wages will come with deeper consequences, ones that our society will now wrestle with on a daily basis and for years to come.
Increases in minimum wages will accelerate a massive shift to automation.
I see it on a weekly basis with startup pitches. Name any manual job currently booked at around $15/hour or so, and I’d bet there’s a 50% chance that task will be automated, either with pure software, enabled hardware (in the form of a robot), or some mix of both. Automation via software has been underway for a few years, but when it’s combined with hardware, the results will be astonishing. The components needed for such robots are now readily available and getting cheaper by the year. The hardware/robots can be differentiated by software and the vertically-specific applications they are designed for. The robots become a vehicle by which new technologies and services can be delivered, and of course done so at a fraction of the cost without the additional overhead (like healthcare, etc) that can saddle a balance sheet.
I know this is coming — and fast — because I myself use products on a daily basis that replace traditional human labor with automation and learning.
The way that I describe this trend in general is to imagine your local Starbucks. Say it is about 2,000 square feet total. At a busy time, you may see 8-10 workers in the store. Why? Shouldn’t I just be able to walk up, have a beacon notice my presence, and a robot makes my drink. The machines to do this already exist. It is coming.
We can expect to see the consequences of minimum wage increases (which I acknowledge are extremely complicated) take root in automation, robots, and corresponding services. Some solutions will appear like vending machines, whereas other solutions will mimic human movements and behaviors, just in different shapes and forms. The technology and builders are already here, working on these solutions — the turbulence in the bifurcation of the economy, the force-changes driven by technology, and the slow response to build enough suitable housing and transit may combine to usher in this robot-led era with greater speed. Minimum wage, maximum automation.
Bots. They are everywhere in startup land. It feels like a gold rush — a bot rush. And, there is a backlash, as well — a botlash. Whether it’s more and more people using Slack and other messaging apps and come across these new bots and features, the trends set by Y Combinator’s Demo Day (where, by my count, four very interesting companies launched in this space), or one of the 101 pitches investors see from founders, bots are everywhere.
In this rush, there is surely opportunity, and also noise. At a high-level, “bots” are attractive because it is a lightweight format by which a startup or messaging app or Amazon Echo can deliver services to users on the simple end of the spectrum, and perhaps one day automate, predict, and conduct work for us as it learns from our interactions and behaviors, as well as learning from other APIs they interact with.
The noise created by today’s frenzy makes it hard to find signal. As a I result, from an investment point of view, I see the bot world in three (3) categories:
1/ Vertical-focused B2B platforms w/ SaaS business model: This is targeting a specific end users (usually businesses) that need to create a bot for news, or commerce, and pay the startup for a suite of tools to do so. 2/ Selling picks & shovels: When you sense a category is going to be big/important, but don’t know what the end-users will use, one can invest in the developer platforms and services that the bot creators will need to make their bots. 3/ Direct to consumer (or prosumer) plays: This can be seductive because messaging apps are huge (easier to get users, distribution), but also dangerous because today’s platform can become tomorrow’s gatekeeper.
How do I think this will all unfold? I have no idea, but I think the three investment strategies above make sense, all for different reasons. I have an investment in 2, looking at 3, and looking for 1. I’ll let you know when I find out.
Separately, on the consumer side, I’ve been thinking about what I want personally in a bot service. I wanted to write that below in case anyone has built it, is building it, or is looking for an idea. Here’s what I want:
1 – I want to create my own personal bot, Semil-bot.
2 – On web or mobile, I go to your page and authenticate with Facebook. This has to be ID-based.
3 – I connect a range of services and accounts, beyond financial like with Digit etc. Definitely Gmail and Google Calendar, like Paribus.
4 – If I give Semil-bot the keys to my digital kingdom, I want it to not only dive in and learn everything about me beyond Facebook, I want it to crawl a network of specific APIs for financial services, insurance, digital media, social networks, etc. and constantly be on the lookout for my best interests, to save me time and money without me having to think about it or execute any actions.
5 – Semil-bot would send me a daily update email with actions that it’s taken an easy way to reverse, pause, or suspend commitments. It would work for me, on my behalf, 24 hours a day. I would pay $10 or even more for this service.
6 – But, this isn’t just a weekend hack project. As the system gathers more information about me and other APIs, it needs to learn and form a real brain. It needs to learn over time and make inferences. It needs to anticipate things. It needs to deliver a service that saves me both time and money, reducing cognitive load and becoming a trusted digital agent to act on my behalf.
I caught up on Lefsetz’s blogs from the past two weeks yesterday. A few stuck out, like always. In particular, he wrote something about his history of Springsteen concerts, which you can read here. His recounting touches on all sorts of nostalgia triggered by following The Boss over the decades, and while I’m not a huge fan of Springsteen, I hold him in high respect — I once saw him perform, without breaks, to a crowd in California for almost 3.5 hours. No stops. It was one of the most authentic musical experiences I’d ever seen. Maybe I’m more of a Springsteen-the-person fan than a fan of his music.
Re-reading Lefsetz’s journal entry, I remembered I had myself written about the concept of nostalgia intersecting with technology. I don’t know why I write about nostalgia, or why I’ve got a few tabs open for an unplanned afternoon of where I should be stack-ranking investment opportunities and focusing on work. It may be that nostalgia, for me, is even more powerful force that interrupts my ability to (attempt to) think rationally.
“…while you can scroll down your Facebook Timeline and travel back in time, a service like Timehop could present older pictures to users in a way that strikes upon a deep emotional chord. It is this element of nostalgia that interests me. It is a product I’d want. I can imagine Timehop simply running in the background on my iPhone, sending me a gentle notification…”
Later in 2012, it occurred to me sharing photos to the right people at the right time can, in fact, trigger nostalgia. In the next post, I invoked the now famous scene from Mad Men, “Carousel,” where Don Draper, tasked with coming up with a pitch for a slide carousel, paints a picture of moving back and forth through time by pictures. This nostalgia, he says, has the power to create a close bond with the consumer. While Timehop can’t compete with Facebook, and while what Timehop pioneered might become a feature of how we all use Facebook in the future, it is fun to look back on how it presents us with nostalgia:
“…it simply gives people what they want in a new form — the place where you can keep your memories. The carousel of old slides, the cigar box of warped pictures, and the Instagrams you’ve taken, now in your pocket, delivered to you in just the right way.”
Nostalgia even led me to join a startup company, Swell. A few years ago, as someone who grew up as a radio and audio junkie, I got swept up by the old memories of listening to transistor radios and studied how radio and radio imagery influenced the brands, lyrics, and sounds of some of my favorite musicians. In writing about Swell’s product:
Radio fills the dead time in my life, when I am free to be more at ease, more relaxed, and as a result, my brain seems to expand a bit more to let in more information. Yes, we are visual and textual creatures, and images are central to how we process information, but audio is equally important for me, and when it comes to knowledge, the ambient awareness provided by radio is perhaps the most powerful.
And all of this went through my mind watching the Boss at the Sports Arena. My life slid by. People my age are thinking of retirement… But once upon a time we were the youth, we were the cutting edge, there were no social networks, cell phones were a “Star Trek” fantasy, we had to leave the house to connect, to feel alive, and where I felt the most comfortable was at the show… It was completely different. No one stood, except for maybe the encores. There were seats. You didn’t go to be seen, you went to communicate with the music, bond with the gods. And it was like that Thursday night. And it won’t ever be that way again. It can’t be. Mystery is history. You can see it all online. And scarcity is a thing of the past.
To connect, to feel alive, and places where we feel most comfortable — I’ve been thinking about that passage in particular. Today, Oculus Rift virtual reality headsets are shipping. eSports, where fans worldwide watch other people play video games. Legions of EDM or Taylor Swift concert-goers are recording their experiences through Snapchat Stories. The web and mobile devices are empowering people who may have once felt lonely now connect with likeminded people across social classes, political borders, and beyond.
And now that everything is recorded and documented in real-time, accessible by search, searchable on billions of devices worldwide, it is both empowering and unsettling from the point of view of nostalgia. Growing up in the 80s, most of my memories are locked in a few videos, many still photographs (that haven’t been put online), and in my mind — but what about my daughter, who is almost three years old, who is the adoring subject of thousands of photos? She will be able to see so much of her younger self in digital form, but where will her nostalgia reside? In digital form? In the corners of her mind?
I’m not sure I have a great conclusion here. Perhaps there isn’t one. I’ll end with one of Andy Weissman‘s old posts, back from 2013, where he writes this passage about the idea of putting old videos from his youth on YouTube:
When I tell people this story, they mostly have the same reaction. “You need to put the shows on Youtube!” The video tapes – cartons of them – are spread out. Maybe in California. Maybe at my mom’s place. Some in Woodstock. Maybe they are gone. These requests usually set off a flurry of internal emails amongst ourselves: should we do this? Have you watched them? Which one should we digitize? This year we will really get around to it, yes this year we will, right And then when I think about it, I realize we probably shouldn’t, and most likely won’t, digitize them and put them on Youtube or Vimeo or wherever. It would ruin the memories.
The last big wave which helped generate huge returns for technology investors was largely driven by phones. Building apps and services on top of phone sensors, connected to the network, gave us new media apps like Instagram and Snapchat, new communication tools like Whatsapp and Messenger, and new ways to travel like Airbnb and Uber (thank you GPS sensor and Google Maps API!). The mobile wave was/is big enough to create opportunities for others, as well, though these are the biggest outcomes.
In early-stage investing today, a bunch of friends and peers who invest at seed always wonder — what’s next?
Lately, when I’ve been asked this question, here’s the analogy I use to the answer the question:
Imagine we are all surfers in a surf competition. During the day, we each are allowed to pick a set number of waves to ride and are scored on them. There are lots of surfers doggy-paddling in the open ocean, and the goal is to identify, pick, and line up to catch the biggest wave of the day. The problem, of course, is that in the moment, from your vantage point with your head bobbing on the surface of the water, it’s hard to identify and commit to the wave at the right time. If you move to soon, you may pick the wrong wave; ff you wait too long for the wave to take shape, you may not have enough time to catch it properly.
The goal, of course, is to pick the right wave and time it perfectly. Picking the right wave will be scored well and rewarded by the judges, will give the surfer an unforgettable ride, and will pack enough kinetic, nautical energy that will propel the surfer to reach new speeds. That is the goal.
In reality, right now in the early stage, everyone’s wading in the open ocean, surveying the horizon for promising waves to form. There are waves we anticipate in tech but we don’t know when those waves will reach a point where we can ride them — waves around Artificial Intelligence and Machine Learning; or vertical marketplaces; or SaaS network; or Virtual and/or Augmented Reality; or bots and agents; or autonomous robotics; or…name any other big category. We all just don’t know what that wave will be, where it will come from, and what it will look like.
In the face of this uncertainty, some elect to follow their peers who are known to have a good nose for spotting big waves. Some have studied up on how to pick out a big wave looking at data or other physical properties. Some are happy to pick out a portfolio of a few waves and hedge their bets (that’s what I’m trying to do). Every strategy comes with its benefits and risks. Entrepreneurs, of course, see these waves before most investors do, but both founder and investor can pick the wrong wave or move too late when the big wave is forming. It will be just a matter of time before this next wave emerges — because no one knows when it’s coming — and it will be exciting to see what it looks like and where it comes from.
Where is the early-stage financing market for startups today, mid-March 2016? In my opinion, and depending on where specifically you sit, right back where it left off.
After a few bumps over the last 9 months, with little shocks in July 2015, a bigger shock in August 2015, and a ruthless slashing of very visible and great public technology companies in January and February 2016, many people (and yours truly) commented that the entire environment has “chilled.” Now, as March is unfolding, that is changing a bit. Working backwards, here’s a brief snapshot of what I’m seeing in the SF / Bay Area market — just one guy’s opinion:
Series B (“the second board member”): These are known as the Rounds of Death. As the private late-stage dries up, Series B and C investors, who structurally need billion dollar plus exits to make their models work, know that IPO windows are jagged, brittle, and chilly, and since M&A over a billion is quite rare, many have either held off or suggested flat/down rounds to make new investments. (Yes, I know, some companies still get funded.)
The Series A Rounds (“the first board member”): These are happening, albeit slower, but the caveat is that there are now new kinds of Series A rounds which can distort where a company is. It used to be that a classic Series A is when an institutional VC joins as the first real board member of a company and makes a life-cycle commitment to the company and founders. Now, there are more funds and more people doing rounds the sizes of the old-school A, say $5m give or take, plus or minus. This could be insiders protecting an investment, or LPs coming in to buy more direct ownership — or, it could be one of the many new funds being formed almost weekly to put dollars to work in tech and get ownership.
All The Seed Notes Before The A: This includes pre-seed, seed, second seeds, seed prime, seed extensions….the never-ending parade of convertible notes that we’ve all come to love. It is here, early before the A, that I see little to no change in the market relative to how Series B investors are extremely cautious or Series A investors that are taking their time (yes, I know, and still doing deals). There is not only so much money to be deployed across the seed market, I have learned from being on the fundraising circuit myself that there is more money just sitting on the sidelines, and we will see this in the form of new funds (new ones every week), and they have to deploy those funds, and seed is where most of us play because there is no barrier to entry — there are some barriers to entry to doing Series As because founders want those to be branded and institutional, but at seed, it’s loose, fast, convertible debt that matures in two years or, if you’re lucky, less. Given these factors, founders in the right spaces with good teams raising seeds now can get higher prices, and that’s especially true in/around some of the most successful accelerators and incubators.
This is where I see the market today, mid-March 2016. Seeds are a plenty, and growing well; Series As are happening, but a bit slower; and Series Bs are when founders feel the most friction. All in all, it is still a great time to start a company, but we may find out over the next few years that we’ll have startups turning over quicker and perhaps more replenishment at seed as the larger institutions wait for conditions to become “just right” before making a life-cycle commitment to a new company.
This post is not about taking a side on a political issue or for a particular candidate, or drawing a line between what’s right or wrong. Before this could be twisted by others, let me say I do not condone the messages being used in the campaigns in question below. Yet, as a longtime observer of presidential elections (and as someone who worked both for and in/around government, as well as for a startup that tried to play into the 2012 election), I find this season’s chapter of national politics truly remarkable in that I believe we are seeing unprecedented change.
“The Rise of Trump” seems to be such a black swan event, many people are trying to make sense of it. Scott Adams, the outspoken creator of the Dilbert comics, has been writing about the Trump phenomenon on his blog for many months now; earlier this year, well-respected media critic Clay Shirky wrote this tweetstorm about how social media has helped reshape what a political party looks like in America; Stratechery’s Ben Thompson wrote a detailed analysis sharply titled “The Voters Decide“; and one of the sharpest minds in the world of startups, Naval Ravikant, took to his blog in a piece title “American Spring” to link the rise of bipartisan populism and the rise of both Trump (on the right) and Bernie Sanders (on the left) to social media’s disintermediation and re-intermediation of our political reality.
For me, this only came together recently as I was listening to Keith Rabois talk about the rise of Trump using language of how we talk about startups that become black swans. I talked to Keith today and asked if it was OK for me to quickly transcribe his talk with @Jason. As usual, Keith has a truly unique way to explain things and tie them back to how startups form, operate, and scale. To be clear below, I quickly transcribed this tonight, and Keith explicitly says he doesn’t want Trump to win (original source video, skip to 32-minute mark to hear this part of their conversation).
@Jason Calcanis: What are the chances Trump could get the nomination?
Rabois: Higher than I would like…it’s not the case Republicans fielded a bad set of candidates, it’s just that voters aren’t voting for them. If you look at the resumes of the 13 people who ran for the Republican nomination, they’re incredibly impressive people…you’re talking about the governors or former governors of Florida, Texas, Wisconsin, New York, New Jersey, Virginia, etc….but none of the voters [want to vote for them]. It’s like shipping a good product on paper with a really good executive team that has a great LinkedIn profile, but none of the consumers (the users) are downloading the app, that’s basically the problem… It’s not necessarily obvious what to do when the people are just rejecting the traditional candidates, it’s a little bit like building a startup. Basically, someone is coming out of left field with a completely different model and, in some ways, is defying all the rules that people took for granted and is having success. That’s what most of us do for a living… I don’t want [Trump] to [win]. I don’t think it’s impossible — for him to win either the nomination or the Presidency. I think the percentage chance [of a Trump Presidency] is higher than people think because he’s shown this disruptive ability — translating this back to entrepreneurial pursuits — he’s clearly defying all of the rules in a way that’s resonating with real people, and when you see that, and you see that starting to scale, ignore that with a lot of peril, just like incumbents ignore these startups all the time.
The media narrative and intra-party opposition to Trump has placed focus on his dangerous use of language, but we haven’t yet heard of his rise in the language of startups — in an era where traditional lines of morality continually erode, we are now seeing a disruptive force in national politics (like we did eight years ago), defying and rewriting many of the rules political insiders, parties, and voters have taken for granted for decades. And with each passing milestone, each passing attempt to curb its rise, it seems to get stronger as it scales. In the language we all understand, we are trying form and/or fund the best teams with the right backgrounds and abilities to unlock markets, but even with all the planning and know-how in the world, the next thing usually is explicable only in hindsight.
Late last night, around 10:30pm (“late” for parents), my brother (who is visiting from the NYC area) and I were standing in my kitchen as he was getting an UberX to head back to the city. There’s never really surge at that time. He started with 1.7x. Waited a bit, 1.6x. In a few minutes, it was 1.4x and he locked it in. Note that he could expense this ride for work, but it was his own credit card. At the same time, standing next to him, I tried to hail an Uber. We both did not input his destination address (SF Union Square area). I went from 1.6x to 1.7x to 2.4x!
We may have initially though of Uber’s “surge pricing” as being based solely on the available supply versus demand for a ride in that moment. On January 1st of 2012, I wrote a post about the future of Uber’s surge pricing that was widely distributed (relative to most of my posts). This was the day after Uber’s New Year’s Eve surge pricing became a sort of Twitter tradition, of people complaining about price hikes and sharing screenshots of their surge multiple. In that post, I hinted at a future in which, via mobile devices, we generate and share so much data that, one day, in exchange for all the convenience mobile apps paired with offline services deliver to us, we may pay the price for that through more levels of dynamic pricing.
And, with Uber, it feels like that day has arrived. I am not sure if this is actually happening now, I don’t have inside information or could prove it’s true, but it’s not out of the question and many people I’ve been tracking across in the country in one of my filtered Tweetdeck columns seem to lament that they’re getting “surged” more frequently and at times when they don’t expect to.
If we step back and think about the inputs needed for a company like Uber to more precisely discover our willingness to pay for a ride at a particular time, consider all the variables now that Uber can mine and extrapolate from:
1/ Likely Destination: You’ve probably used Uber enough by now for them to have confidence about where your intended destination is for a ride. Without explicitly revealing that to the drivers in your area, they could already have a range of which type of surge will actually make the ride transaction go through, even after a driver has accepted a pick-up.
2/ Point of Origin: Uber can take the ZIP code you’re ordering from and make instant inferences about one’s willingness to pay, perhaps even down to the level of an address. I don’t know if they do this, but imagine leveraging a Zillow-like API to inform their system that someone is ordering from 800 Jackson St (recently sold for $5m) vs 826 Jackson St, which is a set of apartment buildings.
3/ Previous Surge History: You’ve likely accepted surge before. How often? Hit the button a few times over the months and Uber can start to build a profile of when you’re willing to submit to the surge.
4/ Type of Credit Card: I hinted at this above. Now that Uber has a line of sight into corporate travel and can be expensed by many for work, they could juice surge for those customers knowing that it will just get expensed anyway, thereby making the customer price insensitive.
5/ Behavioral Data: Maybe Uber can make inferences from the type of Spotify music I’ve integrated with and play in the car, or has a data arrangement with Facebook in its partnership with Messenger, or can access more motion data from a user’s M8 chip in iOS. I’m surprised by now they haven’t asked for calendar access on the phone to do more predictive pushing. There’s probably a host of things they could look at which can further strengthen their data and algorithms.
“If” this is true and happening (again, I can’t prove it), it could theoretically improve Uber’s margins and, one could assume as it makes more money, drivers will make more, too. (I know self-driving cars is another looming issue, but while the technology there is advancing at incredible rates, I still contend it will be a while before various transportation authorities authorize this across the United States.0
I thought I’d end with an edited portion from the piece on Uber’s surge I wrote over four years ago:
This reality is the other side of the daily deal market, one not driven by discounts and demand, but rather premiums for things that are scarce. Which brings it all neatly back to Uber. Some riders last night wanted the combination of a guaranteed ride at a time of their choosing, but also at a price that they deemed “reasonable.” Unfortunately, since everyone else also wanted rides around 10pm and 2am last night, the demand so far outstripped the supply that what seem to be gross surcharges were actually automatically generated to make sure a consumer’s willingness to pay matched the good offered…However, this is also a wakeup call for consumers, those who use Uber and in general. As devices and ecosystems enable us to share more and more data about our location and what we truly want at any given time, time-based pricing is simply a natural extension of this grand bargain…”
Third installment of our new podcast, “While You Were Away,” a sort of Rap Genius for tech Twitter conversations. Joel Andren (who hosts this podcast) does a terrific job stalking Twitter to see conversations unfold and learning from them.
1/ Startup in SF – Worth It?, relating early adopters in the Bay Area and challenges in finding and retaining gig labor. Conversation includes Adora Chung, Keith Rabois, Adam Besvinick. (Twitter thread)
2/ Complaining About Twitter, relating to how startups should prepare for the 2016 fundraising environment. Kicked off by Danielle Morrill, joined by many others. (Twitter thread)
3/ How To Become A Pro VC, where we break down the great conversation started by Kanyi Maquela, including tweets from Keith Rabois, Josh Elman, and many others. (Twitter thread)
Second installment of the new podcast, “While You Were Away,” a sort of Rap Genius for tech Twitter conversations. Joel Andren (who hosts this podcast) does a terrific job stalking Twitter to see conversations unfold and learning from them.
1/ Startups vs Regulation, relating to changes at Zenefits. Conversation includes Modest Proposal, Keith Rabois, and Ben Thompson. (Twitter thread)
2/ Complaining About Twitter, relating to changes to Twitter product. Kicked off by Paul Kedrosky, joined by Sean Garrett and many others. (Twitter thread)
3/ How To Be Taken Seriously By Professional VCs, where I give guidance on how to get a real VC to pay attention. Includes Bubba Murarka, Josh Felser, Keith Rabois, Josh Elman, and many more. (Twitter thread)