My first Demo Day in London: Entrepreneur First (#ef6)
Yesterday I had the opportunity of attending Entrepreneur First’s Demo Day here in London. It was my first exposure to such an event here. I found it to be very different from Demo Days in the US. First and foremost, Entrepreneur First is not your typical accelerator. They have a unique model — I haven’t seen anything like it in the US — where they start from talent, not from a team/startup.
Highly skilled software engineers (and other talent) join Entrepreneur First program without knowing what startup they will build and who will be on their team. The magic that happens inside of that program is unknown to me, but the results are what I was exposed to yesterday.
My first impression was: Wow!
Not entrepreneurs, but applied scientists.
Most teams there didn’t qualify as one would expect from an entrepreneur/hacker trying to build a business. Most of them were hard scientists in software and mathematics, applying their Ph.Ds., research, and knowledge into a product idea. What they lacked in business acumen they compensated in brain power and domain expertise.
The companies were raising between £350K to £750K and being honest, as an investor, if you blindly bought 100% of the round of all of them you’ll have an excellent return on your investment. At least half the teams will make for very high-priced acquihires and a few could easily make up for the whole investment.
The good stuff
So many startups were pitching themselves as AI companies, I was expecting to see the type of if-then-else “AI” that’s so prevalent in the US right now. It turns out; they were quite above that crowd. Some of them not only had the exceptional technology (looking at the surface) but found a great application with high monetization potential.
The pitches were well delivered even when they had to speak about complex topics (Quantum cryptography, Neural Network server optimization, and genomics data compression).
The neutral
- Rounds were quite small, at least comparatively speaking to the seed rounds in Seattle (which are even lower than rounds in the Bay Area).
- There was a significant amount of emphasis on the founders pedigree. What startups usually spend 30 seconds or less during Demo Day in the US, here they would spend 2 minutes boasting about their published papers, research work, and patents.
- At least two startups said things no US-based startup would ever say. One of them about using a city’s CCTV to identify license plates of cars. It’s not that the idea wouldn’t work in the US, but the nonchalant views of individual’s privacy felt strange to me.
The not so great
- Roughly half the startups presented unreasonable Total Addressable Market numbers — not so different from US startups during Demo Day, it’s just a pet peeve of mine when startups don’t know how to quantify TAM (one even had a math mistake in their deck that would make their “TAM” be actually 10x bigger, but that was probably too ridiculously high, so they just removed one zero to make it more plausible).
- Most of the CEOs will have a long road ahead to learn the skills to be a great CEO, even at small scale teams. I was convinced by their abilities to crack complex computational problems, not so much about their capacity to recruit and manage talent, or market and sell their product.
- Some fluff and meaningless slides, even more so when they say things like “… and we are in talks with the top X (industry) corporations…” (either you have an LOI, a pilot or a paying customer, or you have nothing). Or, when they state platitudes irrelevant to their pitch (see #2 on my post about great pitches).
- About two-thirds of the presentations had no mention or very light mention of anything resembling a distribution, go-to-market, marketing or sales strategy. (It would make Dave McClure go into a “fuck” rage).
- A few were excessively UK-focused. Not that they pitched as a UK-business, but they probably could not replicate what they are doing without a whole new technology piece (not necessarily bad if you are pitching as your entry market).
- I have a very good imagination, and it’s easy for me to imagine the pivots (or new markets) a startup can take to make itself into a billion dollar business. I didn’t get the sense most of these entrepreneurs were aiming for this kind of goal, nor that they could reach that kind of goal. As much as it’s eye-rolling to see US entrepreneurs pitching their $100M ARR in year 5, these companies were presenting something an order of magnitude smaller.
Here is my review of each startup and why they fit into my top picks, missed the mark picks, or the unclear category.
My top picks:
Brolly: Selling and managing insurance via Mobile.
Thoughts: Probably the most “boring” business of all presenting
companies, but the CEO was very knowledgeable about the industry, they
had real partnerships, and I can see how a whole generation of
millennials would pay a premium to not have to deal with an insurance
agent.
Calipsa: Traffic monitoring.
Thoughts: Despite my concerns about privacy, I think this is a huge
problem for cities who spend gazillion dollars in an inefficient process
to collect even simple data points about roads, cars, traffic & more.
I’ve seen too many times DOT employees sitting on the side of road
counting cars.
eBlur: e-sports AI coach.
Thoughts: I’ve very familiar with eSports Coaching having built a
marketplace prototype matching coaches and players. It’s a big industry
and what they did is cool. They will be able to make the offering cheap
enough that becomes accessible to millions of amateur players. People
already spend hundreds of dollars a year on videogames, spending a
couple of dollars a month on a service to help you take to the next
level will be a drop in the bucket for consumers (this is if eBlur
figure out how to reach consumers).
Flexciton: Energy efficiency for rotating equipment.
Thoughts: It’s a software problem playing in the hardware world. It
would cost these companies millions of dollars to build anything close
to what these folks did, so their opportunities to be acquired or
signing many multi-million dollar contracts is very real.
Kheiron: Machine learning applied to radiology.
Thoughts: They have trained models and working tech. I’ve seen three
companies over the last few months doing something similar, and Kheiron
is by far the best. It will be an easy acquisition by Philips or another
MedTech company.
Neurofenix: Device & service for stroke rehabilitation.
Thoughts: Not a lot of innovation in that space and if this works, it
would be huge. Then, it can be expanded to so many other types of
rehabilitation.
Petagene: Pied-pipper for Genomics. Sorry, I meant to say
Compression algorithms for Genomics data.
Thoughts: If they play their cards right, they can build the Dropbox
for Genomics. If they don’t, they will become pkzip.
Neo AI: Neural network CPU/Memory optimization.
Thoughts: They will be acquired by Google/Microsoft/NVidia in 3, 2, 1…
Drafter: AI for salespeople.
Thoughts: It’s probably the best one I’ve seen of the bunch. They
figure out how to apply NLP/AI in an incredibly elegant way (integrating
with email), in an industry with a very real problem. I was sold on the
pitch. They just need to productize it correctly to cash in.
Sanctum Technology: Noise cancellation for home.
Thoughts: Their application to eliminating airplane noises in the UK
is very clear cut. But there are many other applications to what they’ve
built, including manufacturing, car noise reduction, even inside of
airplanes itself. They can go far.
Xihelm: Street survey/asset management.
Thoughts: Feels right-up-there with Google’s mission of organizing the
world’s information — in this case, road information. Once they have all
the data, new applications will become available, and they will have a
2–3 year moat to the next competitor.
I buy the problem; I don’t buy the solution
Accurx: Reduce inappropriate antibiotics.
Thoughts: Too many moving parts in this space (doctor, pharma,
patients, data, privacy, status quo) and little monetary value being
created for it to work as a business. I get the “destruction” part; I
didn’t get the “create destruction” element.
Crypto Quantique: IoT security.
Thoughts: Too many competing solutions out there. There’s a long road
ahead of them.
Cyra: AI for recruiting.
Thoughts: Yes! We need much better ways to match candidates to jobs
(and vice-versa). The technology is impressive, but the angle they
decided to take I believe to be wrong. Heck, maybe I’m wrong.
Intelligent Robots: Basically,
Kiva.
Thoughts: It seemed like they had nothing more than a prototype. No
on-going pilot. No pending pilots. Just a cool self-driving robot
prototype.
Quotable: Procurement for Small Business.
Thoughts: Really like the problem, but the pitch was very
unconvincing. A basic marketplace where I was expecting some real
innovation.
I’m not convinced (or, I’m confused)
Bloomsbury AI: Enable anyone to create and train AI models
online.
Thoughts: Probably the best presentation of the day, but light in the
business model and real application examples. I wish they would have
gone deeper into a single vertical to be more concrete about the
capabilities and how they were going to monetize it.
Keypla: Build 360 View for real estate sales.
Thoughts: This has been tried so many times in the US! At best, they
are getting a tiny slice of a small slice of a transaction. No way to
build a big business.
Loop Perfect: Just-in-time compiler for client applications.
Thoughts: Yes, it’s a problem. Yes, this solution kicks ass. At best,
I see an acquihire here. Developers are finicky about paying for
software, so unless they are monetizing through some other (magic)
means, I can’t see how they can make oodles of money (Psst.
@scottgu, you should buy them).
Suits Me: Personalized lifestyle assistant for fashion.
Thoughts: I just can’t see it being a big business. I’ve seen startups
with more sophisticated technology fail in this space. The problem is
small, AI as creative agent sucks (still), and retailers won’t care
about a small lift in sales. Now, if you package that into a competitor
to Shopify focused on fashion, then it could be interesting.
Alpha I: It’s a hedge fund bot.
Thoughts: Once every fund is automated and optimized, the edge is
gone.
Even the startups and products I wasn’t a big fan of had impressive teams. The average IQ of these founders is scary high. Hopefully, they can manage the challenges of going from a feature to a product to a business, and even the ones that weren’t my favorites left me with a feeling they were just a pivot or positioning away from hitting the mark and being successful companies.
All that said, I take my hat off to Entrepreneur First. So far, the best organization incubating startups I have ever seen. I played with the scenario of a free-agent entrepreneur organization that never became concrete, and EF is the one that comes the closest to it.
