I think before even considering failure rates, it's worth first answering another question first. What precisely is a startup anyway?
In its broadest sense, it is a new business in its earliest stages of development.
This definition is too general, however, and as a result - misleading. A new restaurant outlet is also a new business in its early developmental stages, but most people in the startup community would tell you a hairdresser salon isn’t a startup.
A startup usually has two important characteristics:
Innovation: A startup is testing assumptions that haven’t been tried before – sufficiently new technologies, products & services, or markets.
Growth: A startup has the potential to grow exponentially rather than linearly. It is scalable. This usually happens because technology provides leverage.
So, a startup is in essence, a business experiment with potential. This means that real startups are prone to failure by definition.
They are testing assumptions, and it’s very likely these assumptions are wrong. The more innovative the startup, the riskier the assumptions and the more likely it is to fail.
When you put this new kind of risk on top of the traditional risks of starting a business (finance/cash flow risks, operational risks, team risks, marketing risks, etc.), it’s no surprise most startups fail.
The Usual Suspects
As an Early-Stage Startup: You are searching for a product-market fit. The goal is to validate your assumptions as quickly and cheaply as possible and to give yourself time to pivot if necessary. Getting a good grasp of the meaning of MVP, validation experiments, and validated learning is crucial. Learn to prioritize and change your priorities based on customer feedback.
Startups need 2-3 months longer to validate their market than most founders expect. (The implication here is that cashflow/availability problems can kill the project before you can properly test the waters.)
Founders overestimate the value of the intellectual property before product-market fit by 255%.
Startups that pivot 1-2 times have 3.6x better user growth and raise 2.5x more money. Startups that pivot 0 times or more than 2 times do considerably worse. (The implication is that it is prudent to secure sufficient time and resources to attempt up to two pivots.)
As a Late-Stage Startup: One of the biggest traps is premature scaling. Startups that scale prematurely are inconsistent. A startup can be broken into four stages: Discovery, Validation, Efficiency, and Scale.
Inconsistent startups write 3.4x more code in their Discovery phase and 2.25x more code in the Efficiency phase.
Inconsistent startups raise 3 times more capital in the Efficiency stage and 18 times less capital in the Scale phase.
The self-reported valuation of inconsistent startups before reaching the Scale phase is $10 mil. Consistent startups report $800k. Inconsistent startups have 75% more paid users in the Discovery and Validation phases. Consistent startups have 50% more in the Scale stage
Market interpretations
There's a lot of scope for failure when it comes to Market variables, but as an obvious general rule, companies can fail if there is no market for their product. This can happen if the value proposition is not compelling enough, or if the timing is wrong. The product may be ahead of the market by a few years, or the market may not be ready for the solution. It's important to find buyers that have a pressing need for the product, and to be aware of whether the product is a "nice to have" or a "must have".
Failure to find a profitable Growth Model
A lot of founders are too optimistic about how easy it will be to acquire customers. They assume that because they will build an interesting website, product, or service, customers will dig a path to their door. That may happen with the first few customers, but after that, it rapidly becomes an expensive task to attract and win customers, and in many cases, the cost of acquiring the customer (CAC) is actually higher than the lifetime value of that customer (LTV).
The observation that you have to be able to acquire your customers for less money than they will generate in value over the lifetime of your relationship with them is stunningly obvious. Yet despite that, I see the vast majority of entrepreneurs failing to pay adequate attention to figuring out a realistic cost of customer acquisition.
The VC Problem
There is one defining characteristic amongst VC firms that are not equally distributed or broadly practiced, and that is Investment Discipline. In the venture capital industry, investment discipline can be described (at the fund level) as maintaining optimal fund size, investing in a reasonable quantity of companies per fund, investing at a certain stage (or stages) of a company’s life cycle, maintaining objectives for portfolio diversification, setting concentration limits, maintaining adequate coverage for team members, avoiding trends and not investing outside of expertise, setting valuation guidelines, setting investment reserve minimums, knowing when to focus on the winners and knowing when to let go of your losers, and finally, having a defined process for how investment decisions are made.
But it seems to be that there is an abundance of a lack of discipline and competence amongst VCs, leading to inevitable but very preventable flooding of capital into markets and the funding of lackluster startups. Moreover, an abundance of such kinds of poor allocation decisions leads to skewness in the general populace’s perception of what it takes for startups to truly succeed, and at the same time, an Investor’s ability to make risk-adjusted bets.
In most walks of life, when things are going well, people are less likely to notice the mistakes they are making and often fail to examine how they could do things better. Success tends to mask strategic errors, especially at the organizational level. Great venture capital firms are those with the conviction to stick to their investment strategy and are careful not to let success lead them down paths that could alter their identity, and more importantly, their investment performance.