Atoms < Bits
What do billion-dollar companies look like?
You’ve probably heard of National Express, the company that operates a vast number of bus and train services in the UK and beyond, turning a profit of £40 million in 2021. You might have heard of Balfour Beatty, which runs large-scale infrastructure projects like upgrading the Piccadilly Line on the London Underground – it made £194 million in profits that year. At the time of writing, both of these are billion-dollar companies, with market capitalisation of £1.2 billion and £1.8 billion respectively. On the other hand, you probably haven’t heard of Pleo – a Danish company selling software that makes it easier for employees to file their expenses claims. Founded as recently as 2015, Pleo has around 400 employees and currently loses money. And yet, its $4.7 billion valuation means it is worth more than National Express and Balfour Beatty combined.
To many people this might seem absurd. How on earth can an obscure start-up have a much higher valuation than these stalwarts of the FTSE 100? There are three reasons. The first is simple: valuations aren’t based on the financial results companies have delivered in the past, but on investors’ expectations about the results they will deliver in the future. This isn’t to say that track record counts for nothing, as strengths built up over time like a loyal base of existing customers or a well-recognised and trusted brand are often ingredients in future success. But it does mean that, when it comes to valuation, achievements matter less than potential. Pleo’s backers clearly believe that £1 invested today in automating expenses claims is likely to produce 2-3x more in profits than £1 invested in transport or infrastructure projects.
The second reason is more complicated, and relates to additional shareholder rights that venture capital (VC) and private equity (PE) investors often benefit from when they agree to finance companies. This means that comparing valuations of companies in the private markets with valuations of companies listed on the stock exchange is illuminating, but not strictly “apples to apples”.
The third reason Pleo is worth so much more than National Express and Balfour Beatty is the most important one. To paraphrase the entrepreneur and investor Peter Thiel, Pleo deals not in “atoms”, but in “bits”. Atoms are the building blocks of the physical world – they make up everything material, from the tyres of a bus, to the steel of a rail, to the fabric of a luminous tabard. Bits, by contrast, are the basic unit of the digital world – the zeros and ones from which all data and computer code are constituted. Compared to bits, atoms are expensive, heavy, and cumbersome. Compared to atoms, bits are costless, weightless, and pliable.
Or, to put it in more prosaic terms: Pleo is more valuable than National Express and Balfour Beatty because it’s a software company. And this is the first key takeaway for anyone nurturing a unicorn dream: build software.
Why “software” and not “tech”? Firstly, because software – being made of bits – is highly scalable. Once you have built it, the cost to produce and additional copies for sale is basically zero. This isn’t the case with computer hardware, where each and every device requires its components to be sourced, assembled, packaged, and physically shipped. Secondly, because developers of software aren’t engaged in inventing something from scratch, with all the uncertainty and risk that that entails. Most software products involve remixing existing code, much of it open-source. My old colleague Gi Fernando memorably called modern developers “Code DJs” – instead of writing their own material, they weave other people’s material into something new. Software developers get to “stand on the shoulders of giants” – something that may not be possible outside the digital world, in fields like biotechnology, nanotechnology and clean tech, where scientific breakthroughs are often required to bring an idea to fruition.
This isn’t to say that other kinds of tech companies can’t achieve billion-dollar valuations. There are 38 hardware unicorns, including warehouse robotics company Exotec ($2.0 billion) and data centre provider Global Switch ($11.1 billion). Meanwhile, mRNA COVID-19 vaccine inventors Moderna ($58 billion market cap), graphene nanotube manufacturer OCSiAl Group , ($2.0 billion) and renewable hydrogen producers Sunfire ($1.8 billion) are inspiring examples in biotech, nanotech, and cleantech respectively. But there are more than 800 software unicorns. So, if you want to build a billion-dollar company, focusing on software specifically will massively increase your chances of success.
Eating the World
It’s these special properties of software that Marc Andreessen was hailing when he wrote his celebrated 2011 op-ed “Why Software is Eating the World” in the Washington Post. By 2011, Andreessen was already a legendary figure in Silicon Valley. He had invented the first web browser and was a highly successful start-up investor and deal-maker, having co-founded the VC firm Andreessen Horowitz (a16z). Shortly before the op-ed was published, a16z had made a profit of $100 million in just 18 months on a single investment in Skype, and its portfolio included stakes in then-unicorns like Facebook, Twitter, Groupon and Foursquare.
At the time, many commentators believed tech companies were wildly over-valued, and that the whole sector was heading for a crash – just like in 2000 when the dotcom bubble had burst. For Andreessen, nothing could have been further from the truth. If anything, he wrote, tech companies were under-valued in the public markets. Apple, for example, sold distinctive, premium products and had little competition, meaning its profit margins were much higher than its peers in the Fortune 500 – and yet its stock was trading at the same price/earnings (P/E) ratio as the likes of Exxon Mobil, which operated in a commodity business (oil) that was full of strong competitors (like BP and Shell).
Contrary to popular belief, Andreessen argued, it was in the private markets where tech companies were being appropriately valued. The multi-billion dollar price-tags VC investors placed on pre-IPO companies like Facebook and Spotify were more than justified by the disruptive impact they were destined to have on the media and music industries. The valuations were not driven by the profits they had already made – in fact it was not until 2018 that Spotify would turn a quarterly profit for the first time. Instead, they reflected the superiority of their products, the speed of their user growth, and – perhaps most importantly of all – the dynamics of the markets they were operating in. It wasn’t a question of “if” advertising spend on social media would overtake press and TV, or “if” music streams and downloads would overtake CD sales, but a question of “when”. The two companies were set to dominate their respective markets. Once they had achieved dominance, like Apple, their profit margins were sure to improve and surpass all the companies toiling away in old-fashioned sectors like oil and gas, chemicals, and car manufacturing – so of course they warranted unicorn valuations.
Andreessen’s theory was that “software companies [we]re poised to take over large swathes of the economy”, thanks to the increasing availability of internet access and adoption of digital technology around the world. Amazon had already “eaten” bookselling; Netflix was “eating” home entertainment; LinkedIn was “eating” recruitment; Square was “eating” payments. In some sectors, like grocery retail and logistics, incumbents like Wal-Mart and FedEx were reinventing themselves as software companies. Others, like defence, agriculture and energy, were ripe for technology-driven disruption from the outside.
Following Andreessen’s line of reasoning brings us to the second key takeaway for unicorn dreamers: choose a big market that is dominated by technology laggards.
Billion-Dollar Questions
At the time of writing, the biggest global markets are as follows:
This certainly isn’t an exhaustive list of markets with room to accommodate unicorns, but it’s helpful to consider examples from it as we turn to the next important challenge for founders: what specific opportunity to go after. Depending on the market, there are a few good questions you can pose to generate hypotheses for start-ups with unicorn potential.
For example, if you wanted to build a unicorn in the IT sector, you might ask: “What widespread business activities are still surprisingly manual?”
This is the secret of Pleo, who we met at the beginning of this piece. Founder Jeppe Ringdom recognised that filing expenses claims was nothing but inconvenience for everyone involved. It needed employees to keep paper receipts for every expense they incurred, then go through the tedious process of keying in the amounts, finding the merchant in a drop-down list, classifying the spending and writing a justification for it, and finally submitting the claim to a manager for approval. Personally, it was a task I dreaded for the entire period in which I had a corporate expenses card. Even though I was supposed to file a claim monthly, I often couldn’t face it and sometimes let 3-4 months’ worth of receipts accumulate. I frequently lost receipts, which sometimes meant I ended up funding business travel and client entertainment out of my own pocket. To be fair, it could have been worse – an old colleague’s husband who worked at Lehman Brothers had to write off £7,000-worth of unclaimed expenses when the bank collapsed. Expenses are a pain for employers too – erratic, unpredictable and fraud-prone from the perspective of finance teams, who much prefer real-time visibility of costs. At the same time, filing expenses claims is hardly the work you pay your employees the big bucks to do.
By creating a system that allows employees to simply take a photo of the receipt at the moment they pay the bill, Pleo did away with all that painful admin for more than 20,000 businesses across Europe. Its unicorn valuation is a reflection of the fact that the problem it solves is universal: you will find it in every business where individual employees spend money on behalf of a company, all over the world. Pleo’s valuation is also a reflection of its scalability. Because it’s primarily a tech (software!) product, with only one physical component (the credit card), realizing the growth opportunity in Asia or the Americas is straightforward: the user interfaces need to be translated into different languages; sales, marketing, and customer support teams with an understanding of the local context need to be set up; and a different set of laws and regulations need to be understood and complied with. It would be wrong to call those tasks “easy” – but they are definitely way easier than expanding a business like National Express or Balfour Beatty, which deal with expensive, weighty things comprised of “atoms” in the physical world.
Alternatively, perhaps you want to build an e-commerce unicorn. In that case you should ask: “What popular products are still bought and sold predominantly ‘offline’”?
New York-based , which IPO-ed at $1.6 billion in 2020, did this with certain prescription medications, including minoxidil and finasteride, which are used to treat hair loss. Having found a wholesale supplier of the drugs and agreed commercial terms, the technical problem they needed to solve was twofold. Firstly, they needed a way of prescribing the drugs without their customers having to attend a physical consultation with a doctor. Second, they had to find a way of getting the medications to their customers on a regular basis, as they need to be taken every day for many months to improve symptoms of hair loss.
Their solutions were brilliantly simple. They hired doctors to help them design a short online questionnaire that would capture enough information from a customer to tailor a hair loss treatment – or determine that the customer could not be treated (you shouldn’t take minoxidil and finasteride if you are trying to get pregnant, for example). The results were then automatically sent to the doctors for approval before payment was taken and the first set of drugs compounded and shipped. To help withdelivery, they designed packaging that would fit neatly through the letterbox – inspired by other e-commerce businesses that had faced the same challenge, like Graze with subscription snacks, and Bloom & Wild with flowers ordered online.
It’s worth noting that the innovation that got Hims & Hers to a unicorn valuation is far from high-tech – but it used software to enable a solution that incumbent hair-loss clinics hadn’t thought of, or weren’t capable of delivering. Not only was prescribing hair-loss drugs online much cheaper and more convenient, it also opened up the market to younger customers who neverwould have considered “offline” outfits like The Belgravia Centre, whose branding and ads seem to be aimed at a Reader’s Digest demographic.
It’s not a coincidence that such a large share of today’s 1,180 unicorn companies are in IT and e-commerce (see the chart below, based on data from tech market intelligence firm CB Insights). But if these sectors sound too mainstream, you could take a different, contrarian angle, asking: “Where can digital tech have a big impact in industries that are very dependent on the material world of ‘atoms’?”. That would point you to markets like property, food, energy – and perhaps even the National Express and Balfour Beatty domains of transport and construction.
A good example of a “bits”-based unicorn that engages with “atoms” is The Climate Corporation. Founded by former Googlers Siraj Khaliq and David Friedberg, The Climate Corporation develops analytics products based on historical weather and soil data, combined with satellite imagery. The software calculates the best time to sow crops, apply fertilizer, harvest, and so on, helping farmers to optimise their productivity. It sold to Monsanto for $1.1 billion in 2014.
Notorious but still instructive is the example of WeWork. The fact that a glorified office rental company was able to command a valuation of $47 billion at its peak is testament to the perceived opportunity for tech disruption in the commercial property industry. It’s a huge market, worth $9.6 trillion dollars – “the size of the prize” helps explain why WeWork was able to attract so much interest from investors. Similarly, the $5.0 trillion size of the food market accounts for the emergence of unicorns which might otherwise seem anomalous, like Deliveroo ($1.6 billion market cap) and Gorillas ($3.1 billion). But I actually want to make a short detour onto their patch to illustrate the type of opportunity I would advise you to steer clear of if you want to build a unicorn.
Food delivery is an example of a “winner-takes-all” market. In major cities around the world, from Beijing to Barcelona, there are “delivery wars” going on to be the monopoly provider. A vast amount of venture capital has been thrown at driving app downloads, recruiting restaurants, and increasing order volume through cash incentives and subsidies. In my opinion, many unicorn valuations in the sector are an outcome of the astronomical amount that has been invested in trying to win these wars, rather than the inherent attractiveness of the opportunity. Clearly tech can play a useful role – apps that make it easy to summon whatever kind of food takes your fancy at any time of the day or night are a big improvement on a kitchen drawer full of takeaway leaflets. Similarly, calculating the optimal sequence and route for a series of deliveries is a hard computational problem, and therefore a potential source of competitive advantage and value that could extend beyond food delivery. But in my opinion, far too much of the food delivery app business depends on “offline” activities that are harder to scale than code, not least recruiting, training, and managing couriers.
Of course, in practice, entrepreneurs often start companies because they have a special insight into a particular problem, or care passionately about changing a specific industry. Alex Dalyac had a billion-dollar idea during his computer science masters degree at Imperial College London, while working on an unpromising-sounding project with a company that inspected plastic pipe welds. He had a technical specialism in a sub-field of artificial intelligence (AI) called deep learning, and realised that applying it to automate visual inspections could be very valuable. The trouble was that plastic pipes were too niche, with only 300,000 welds taking place in the UK each year. To create a unicorn, he would need to find a much bigger market that would also benefit from the tech. Six years later, the company he co-founded, Tractable , provides software used by major insurers to inspect damage to vehicles and buildings, and was valued at $1.0 billion following its Series C financing. In an interview with the start-up incubator Entrepreneur First, his advice to founders was to “speak to academics. Find out where, across all areas of science and engineering, the breakthroughs are happening…and be the first one to work out which industrial use cases are going to be enabled”.
Similarly, working at Deliveroo was what gave Sten Saar and Harry Franks the idea for Zego ($1.1 billion). The company offers car and motorbike insurance for drivers and couriers on gig economy platforms like Uber, Just Eats, and Gophr – which in my opinion is a much better business to be in than taxi services or food delivery itself. If you’re starting with a specific idea, the key is to understand how it relates to big markets and the technological deficiencies of the incumbents. You want to be able to say something like this to potential investors:
Zego aims to capitalise on an ever-growing market currently underserved by the insurance sector… Zego is able to price policies based not only on traditional factors, but also driver behaviour data and working habits data. In fact, overall, the information Zego can collect amounts to five times more data per vehicle than competitors, or 50 data points per second. This means that we have a much more comprehensive understanding of risk than competitors, enabling us to provide best-value insurance coverage, from policies ranging from one hour to one year.
That was Saar speaking to Techcrunch in 2021. Notice how he juxtaposes the market opportunity –the high-growth delivery sector being poorly served – and Zego’s tech-powered competitive differentiation from the incumbents – 5x more data to make better risk and pricing decisions.
Storytelling > Innovation
In fact, the story a start-up tells about its tech is much more important than the technical reality. There are still plenty of industries where simplyselling a product or service through a website and using digital marketing channels to recruit customers is enough to count as “tech”.
In its first incarnation as boughtbymany.com, this was more or less what ManyPets did. When we launched in 2012, the insurance industry was still heavily reliant on face-to-face or phone-based brokers to distribute products. Even companies like Direct Line who had cut out the middle man and were selling directly to customers had websites which weren’t designed to be used on mobile devices. Nor had they figured out how to use social media: many insurers didn’t even have a Facebook page, let alone expertise in running social media ads. We stepped into that vacant space, and within two years had a bigger Facebook following than any other UK insurance company – a ridiculous situation, when you consider that they had established brands and hundreds of thousands of existing customers, and we were 8 people sitting round a single table in a serviced office.
In our pitch deck for potential investors, we laid this story on with a trowel. We chose the UK’s oldest broker, immaculately turned out with pocket square and blue silk tie ay 81 years old, as the face of an industry that was long overdue disruption.
Recommended by LinkedIn
Combined with our early traction on Facebook, a plausible story about the disruptive impact mobile and social media was set to have on the sleepy and complacent insurance industry was enough to get ManyPets several rounds of angel funding and a Series A led by a major VC. Although we were experimenting with emerging technologies (like a machine-learning model that predicted people’s insurance needs using their Facebook likes) and would later undertake serious developments like re-architecting the core systems insurance companies use to administer policies and claims, our “tech” in our early years was pretty straightforward. We were just much better at putting it to work than the insurance industry’s slow-footed incumbents. The technological bar would be a bit higher for an insurtech start-up beginning its journey today, but insurance is still a classic example of a big market where it’s easy to show how incumbents are “behind the curve”.
Another important aspect of storytelling is combining a rational argument about a lucrative opportunity with an appeal to the emotions – or, in other words, speaking to the heart as well as the head. You will often hear unicorns articulate their goals with phrases like “We’re on a mission to…” or “We’re passionate about…”. They love to talk about how their chosen industry “sucks” or “is broken”, and how they are going to “disrupt”, “shake up”, “fix”, “transform”, and “reinvent” it, or even “turn it on its head”. The purpose of this talk is to provoke anger about the mediocrity of the status quo, and get investors, journalists and customers motivated about trying to change it.
Finally, great storytelling means connecting your start-up to the tech trends of the moment – even if your technology is not exactly cutting-edge. Early in the 2010s, when I was researching merger and acquisition targets for Experian, I saw database marketing companies that had been in business for twenty years successfully reposition their stolid products as “adtech” or “martech”, increasing their valuations by as much as 10x. Then came a wave of hype around AI, which presented a fantastic opportunity to businesses which had previously been known for practicing the more humble-sounding “data analytics”.
In 2018, an old colleague who co-owned a data analytics business invited me out to lunch. Although they were profitable, he was thinking about raising growth funding and wanted to pick my brain about the world of VC and PE. My advice to him was to reposition the company as “AI”, because I knew the label would make it more attractive to potential investors, who would already be thinking ahead to their “exit” (the deal in which they would eventually sell their shares on, preferably for a significant profit). He visibly winced: his business partner thought all the talk of AI in the media was hot air, and had been saying so in conference presentations and meetings with their clients. Leaning into the AI hype would be embarrassing at best: at worst it would look hypocritical. I tried my best to explain the upside: as well as exciting investors, it would increase their chances of winning press coverage, and ultimately maximise the valuation they would achieve when they were ready to exit. And unlike many of their competitors, they even had actual AI capabilities already, in the form of machine-learning algorithms which recommended products and content to their clients’ customers based on purchase histories, web browsing data, and demographic information. I couldn’t convince him – and I do respect his integrity for refusing to join the AI bandwagon. But if your goal is to build a billion-dollar company, that kind of integrity won’t always serve you.
Fast-forward to today, and the hottest tech trends are web3 (a new way of describing blockchain) and the metaverse (an immersive and all-encompassing virtual world of the type depicted in The Matrix and Ready Player One). Now, I roll my eyes as much as the next person when I hear video games like Animal Crossing and virtual reality (VR) apps like Beat Saber suddenly being described as “the metaverse” by journalists and management consultants. But VCs – including Marc Andreessen’s a16z – are pouring capital into it, and if I was running a start-up in video gaming or VR, I would absolutely reposition it as a metaverse company (while keeping as straight a face as I could manage).
Recurring Revenue Rocks
Now you have a good idea for a product, and a compelling story that connects it to a specific market opportunity and one or more prevailing tech trends of the day. The next thing to consider is your business model.
There are a number of options. If your start-up is business-to-consumer (B2C), you could follow a well-worn path and have no business model at all to begin with. If you can attract millions of users with a free product, you will almost certainly be able to keep raising money for as long as you can sustain user growth. That way, you can postpone a decision about how you are going to generate revenue – perhaps even until you get acquired, like WhatsApp.
But personally, I think the downsides of this strategy outweigh the benefits. Only products with in-built network effects grow quickly and predictably enough for a “no business model” approach to be viable. Even if your growth is driven by network effects, with no revenue you inevitably burn cash faster, meaning you need to raise more capital and give away more equity to investors. At the same time, for as long as you are giving everything away for free, you are learning nothing about what your users are worth – either in terms of their willingness to pay for what you are providing, or other businesses’ willingness to pay you for access to them.
Nor am I a big fan of advertising as a business model. The reason is that you have to achieve massive scale before you can make any real money. When I worked at Experian, I looked into acquiring a leading player in classified ads. With more than seven million visits each month, it was one of the UK’s most popular websites. However, it relied mainly on selling advertising space to generate revenue, and the rewards were modest – only a few pence per user per month. That was fine as far as the owners were concerned, because they were running it as a lifestyle business – but turning it into a unicorn would have involved expanding into fifty countries while significantly improving the monetisation of the ad inventory. Because two of the world’s largest tech companies by market capitalisation – Google and Facebook – sell advertising space, many people assume it must be a good opportunity. But these giant incumbents are very much the exception: I can’t think of any examples of current unicorns with advertising-based business models.
Adjacent to advertising and slightly (but only slightly) more promising as a business model is lead-generation. It involves sourcing customers for other companies, and being paid a commission on each sale in return. Price comparison websites are a classic example of “lead-gen”: they attract visitors with quirky TVs ads and promises of money savings, then refer them on to banks, insurers, utility companies and so on, getting a kickback when the referral results in a sale. Commercial agreements between lead-gen businesses and their partners usually create incentives to introduce the right sort of customers, meaning it offers a better financial return than selling advertising space.
However, there are still two problems with lead-gen. Firstly, it’s really hard work, because you have to earn every single penny of your revenue every single month. If your website falls over, you stop earning revenue until you can get it back online. If there is an external economic shock, like the subprime mortgage crisis in 2008, or the war in Ukraine, you feel the revenue impact immediately, and there is nothing to cushion it. I have run two lead-gen businesses: a comparison site called LowerMyBills when I was at Experian; and “v1.0” of ManyPets, the website boughtbymany.com. Both required the majority of team members to focus on short-term revenue, meaning it was hard to make time to work on improvements to the product that would increase the company’s valuation in the longer-term.
The second problem is that partners almost never pay you the true value of the leads you bring them. Their websites typically have design flaws, meaning many people abandon the process without transacting, for banal reasons like their address not appearing in a drop-down list, or a “continue” button vanishing off the bottom of the screen on a mobile device. At the same time, their tracking software will usually award a sale to whichever marketing channel delivered the “last click” – so if a customer you introduced gets distracted and comes back to the partner later via Google search, you probably won’t get paid when they complete their purchase. Worse, none of this is within your power to change. In short, lead-generation revenues are annoyingly flaky. That said, in addition to billion-dollar public companies like Moneysupermarket Group (£1.1 billion market cap), there are a handful of lead-gen unicorns including India’s insurance-focused Policybazaar.com ($2.4 billion).
There is less flakiness to other transactional, commission-based business models. In fact, there are multiple examples of marketplace unicorns, including Paris-based Back Market , which enables re-furbishers to sell gadgets ($5.7 billion), Lithuania’s Vinted ($4.5 billion) – the eBay of second-hand clothes – and Singapore’s Moglix ($2.6 billion), a sort of Amazon Marketplace for tools and office equipment. Unlike lead-gen companies, marketplaces typically “host” the transaction between buyer and seller, meaning the whole end-to-end process is within their control. At the same time, many marketplaces serve markets like fashion and apparel, home improvement, and business supplies, where customers make multiple purchases every year. By building a base of users who can be continually re-engaged at no cost through email, social media, and push notifications, they mitigate the lead-gen issues of having to source brand new customers every single month.
A variation on the marketplace model involves reinventing well-established businesses as online concerns. This is what Cazoo , which IPO-ed at $7.1 billion in 2021, and CARS24 ($3.3 billion) have done with second-hand car dealing in the UK and India respectively. Brazil’s Loft ($2.9 billion) has done the same thing for residential property sales – it’s like Rightmove without the estate agents.
The challenge with the marketplace model is its two-sidedness. To build a thriving marketplace, you need a critical mass of both supply and demand. Without an audience of potential customers, suppliers won’t want to go to the effort of listing their products. But without a wide range of products to choose from, those potential customers won’t have any reason to register with you in the first place. Building a marketplace unicorn is dependent on having a strategy for overcoming this “chicken and egg” problem.
This brings us to what I consider to be the champion business model for unicorn companies: monthly subscriptions. For B2C businesses, it is far better to be able to bill your customers on an ongoing basis than to rely on them regularly transacting on their own initiative. Get someone to subscribe to your service, and the default position is that they keep paying – forever. Pause for a moment and think about your own subscriptions to digital products – as a minimum you probably have Amazon Prime ($25 billion in annual revenue), Netflix ($102 billion market cap), and Spotify ($22 billion market cap). A Dropbox ($8.5 billion) subscription has likely replaced external hard drives and USB sticks as the place you archive digital photos and other large files. Perhaps you support a few creators on Patreon ($4.0 billion) or newsletter writers on Substack (~$1.0 billion). Maybe you have some atoms-based subscriptions too – meal kits from Gousto ($1.7 billion), or dog treats from BARKBOX, INC. , which IPO-ed at $1.6 billion in 2021.
The companies who provide these services won’t like to admit it, but inertia is a huge factor in their success. If a month or two goes buy without you even logging on to their website or app, do they proactively contact you to ask if you want to keep subscribing? Absolutely not. Consequently, for you to stop paying them will require two things: remembering that you even have the subscription in the first place; and summoning the will to go through the process of cancel it.
Meanwhile, many B2B software companies owe their unicorn status to changes brought about by the rise of cloud computing. In the old days, software used to be installed locally on servers the client owned – now it is provided over the web, as a service. Software-as-a-Service (SaaS) has lots of practical advantages over on-premise installations. It is much easier for clients to benefit from upgrades, bug fixes, and security patches. But from the perspective of the would-be unicorn founder, its key benefit is in how it is typically paid for – that is, by a regular subscription which keeps generating revenue until the client makes an active decision to cancel it. Germany’s Celonis ($11.0 billion), which provides software for optimising business processes in sectors like Fast Moving Consumer Goods (FMCG), and Amsterdam-based Mambu ($5.5 billion), which offers digital banks a platform for customizing savings and lending products, are just two examples of literally hundreds of SaaS unicorns.
The implications of a subscription-based business model for a company’s financial performance are profound. In businesses based on regular recurring payments, over time an increasing proportion of revenue comes from existing customers, rather than new ones. This significantly lowers the stakes on new business performance month in, month out, since even if no new sales were made the company would still get the vast majority of its revenue from the existing subscribers. It’s the inverse of the flakiness problem with lead gen businesses: income from subscriptions is both “sticky” and easy to forecast.
In summary: if you want to build a unicorn, you’ll maximise your chances of succeeding if you choose subscription as your business model.
Location, Location, Location
So far we’ve been talking about what sort of company to create, but does it matter where you build it?
There are different schools of thought. On the one hand, today’s unicorns seem highly concentrated in terms of their location. Some 633 (54 per cent of the total) are in the United States, with another 175 (15 per cent) in China and 69 (6 per cent) in India. Within these countries, unicorns tend to be clustered in particular metropolitan areas. The most famous example is San Francisco, which is home to far more unicorns than any other city in the world (167) – but cities like New York (110), Beijing (62), Shanghai (46) and Bangalore (31) are also unicorn hotbeds.
Economists explain this in terms of what they call “agglomeration effects”. Similar businesses benefit from being in close proximity to each other because it creates a large, specialised talent pool which they can all draw on. At the same time, the movement of employees between companies or to found start-ups of their own; client, supplier, and partner relationships; and continual mixing at conferences and social events feeds innovation, as knowledge spills over from one firm to the next. With all this in mind, there are many people who think that if you are serious about building a unicorn, you should move to San Francisco – or at least the biggest cluster you can realistically reach. That will give you the best possible chance of benefitting from agglomeration effects and making your start-up a success.
On the other hand, there are plenty of examples of people building unicorn companies in very different places. With a population of just 52,000, Altrincham in Cheshire has a unicorn – data management platform Matillion , valued at $1.5 billion. So does Kilkenny in Ireland (population 26,000), with business payments infrastructure provider TransferMate Global Payments ($1.0 billion). Duderstadt in central Germany (population: 20,000) is home to the world’s market leader in medical prosthetics, Ottobock ($3.5 billion): Vodnjan on the Croatian coast (population 6,000) to cloud communications platform Infobip ($1.0 billion). With remote working now firmly established following the COVID-19 pandemic, perhaps we will see unicorns appear in other unexpected places.
But if you have a choice over where to build your company, my personal recommendation would be to look for a happy medium. Drawbacks of the major global clusters include eye-wateringly expensive real estate and intense competition for talent, meaning your costs will be higher and you will need to raise more capital. Conversely, if you base yourself too far off the grid, you will find it harder to attract and retain the right people, and to have the kind of serendipitous encounters at events or in cafes and bars that lead to new ideas, sales opportunities and offers of funding. From personal experience, I would say London, which has 36 unicorns, is an excellent place to build a billion-dollar company; Copenhagen, where Pleo is the sole unicorn, less so.
For the whole ten years of its existence, ManyPets’ headquarters has been in Farringdon in Central London. Even though the largest concentrations of employees are now elsewhere (specifically in Birmingham, Stockholm, and Atlanta), moving the HQ has never been on the agenda. The geography of the local area helps explain why not. From the front steps of the office on the corner of Summers Street and Back Hill, you can see two pubs – the miniscule Gunmakers and the more gentrified Coach. On weekday evenings from 6pm to 9pm, they are chockfull with drinkers, often spilling out onto the cobbled streets. Most are in their twenties and thirties, wearing jeans, t-shirts, and hoodies – and almost all of them work in tech. Across the road from the ManyPets office is the HQ of Moonpig, the online greetings card business founded by Dragon’s Den’s Nick Jenkins, which IPO-ed at £1.2 billion in 2021. Next door is Passion Capital, the early-stage VC behind payments company GoCardless ($2.1 billion) and digital motor insurer Marshmallow ($1.3 billion), where Eileen Burbidge, chair of the UK tech sector’s trade association, is a partner, A 5-minute walk through Leather Lane’s bustling street food market takes you past the HQ of rail booking app Trainline (£1.7 billion market cap) to the offices of Octopus Ventures, who led ManyPets’ Series A round, and previously backed the likes of property website Zoopla (sold to the PE firm Silver Lake for $2.9 billion) and fashion marketplace Depop (sold to Etsy for $1.6 billion). Turn left there and a 20-minute stroll will bring you to Old Street’s Silicon Roundabout, the spiritual home of London tech, passing the HQs of business transport platform Gett ($1.5 billion), neo-bank Monzo Bank ($4.5 billion), and online event service Hopin ($7.8 billion) en route. Being immersed in the tech scene doesn’t even require you to hang out at one of the countless start-up events or co-working spaces: whichever pub you drink in, whichever food stall you queue at, whichever café you go into for a flat white, you are constantly surrounded by people who are building and financing tech companies with unicorn aspirations.
Contrast Copenhagen – with the exception of the two-day festival TechBBQ, which takes place in September each year, there is no identifiable physical hub for the tech sector. Pleo is based in the hip Nørrebro district, while the main co-working spaces are across town in picturesque Christianshavn, and VC investors like Northzone are dotted around the genteel northern suburbs. Danish-founded unicorn Tradeshift ($2.7 billion), which develops e-invoicing software, has an office in the city centre but has moved its HQ to San Francisco – likewise content management systems provider Sitecore ($1.1 billion) and wine app Vivino (~$1.0 billion). In fact, Danish tech companies often seem to relocate once they are in the scaling phase – Just Eat ($3.6 billion market cap) moved to London after five years, while online reviews platform Trustpilot chose the London stock exchange for its £1.1 billion IPO in 2021. So, while there are many other things to love about Copenhagen, in my opinion its tech scene is not mature enough for it to be a great place to build a unicorn – at least not yet.
Which other cities else might offer a London-like happy medium? Based on the CB Insights data, in Europe and the Middle East, I would say Berlin (20 unicorns), Paris (19), Tel Aviv (12), Stockholm (7), Amsterdam (7), and Dublin (6); in Asia, Shenzhen (20), Hangzhou (16), Seoul (13), Singapore (13), Mumbai (9), and Jakarta (6). In North America, you are spoilt for choice, with 20 cities besides San Francisco boasting 5 or more unicorns, including Austin (10), Miami (6), and Vancouver (5). In South America, Sao Paulo has 12 unicorns and Mexico City 5. What matters is being based in a city where you can become part of a unicorn tech scene without even trying.
Summary
If you want to build a billion-dollar company you should:
Onward!
Author's note: this chapter was written August 2022 as part of the proposal for a business book called Unicorn Dreams: How to build a billion-dollar company – and why you might not want to. All data was correct at that date.
For various reasons, I decided not to go ahead with the project. But if you enjoyed it, you'll love my books Good Data: An Optimist's Guide to Our Digital Future and Generative AI: An Executive Guide 😇
