0 --> 1
04 FEB, 2024
New Business Ideas
Fail in Whopping Numbers
4-MINUTE READ
Launching a new business idea is a high-risk bet on its success. While various stats try to back this notion with stratospheric failure rates, it's impossible to dig into the solid evidence for those numbers.
I wanted to know if there is any basis to treat our miserable odds with all the seriousness. Well, you can't pin your hopes on what I've found.
The hard truth is you have only a 14% chance to succeed. In other words, every one out of seven ventures will work out. It isn't encouraging, but these are the facts.
On average, 86% of new business ideas fail. However, some aspects of a given endeavor may shift the chances for better or worse:
Location matters: You are better off when you start in The United States (83%) rather than in any other corner of the world (90%). Yet, it is naive to think it is only about picking a headquarters location. Instead, I would seek an explanation in the rare blend of cultural values that has shaped American entrepreneurism, with encouragement for risk-taking included.
Even though the seven percentage points difference for the US is a minor deviation in relative terms, it's spectacular in the outcomes. Just think about all the breakthrough concepts spearheaded by Apple, Tesla, Netflix, Google, Amazon, and Microsoft, to name a few. Without question, the best part of cutting-edge ideas still sprouts on the fertile ground of America.
The failure rate of new business ideas according to geography:
Global
85.9%
Europe
88.3%
The United States
82.6%
Asia
90.5%
The Rest
92.4%
89.7%
0%
avg.
The method behind this analysis is described in the appendix.
The market dictates the difficulty: No matter the field of play, more ideas fail than succeed. But starting in an insatiable market like semiconductors is a safer bet (64%) than trying to innovate in the overcrowded arena of web design solutions (98%).
The failure rate of new business ideas according to the market:
Web Design Solutions
98.3%
Language Learning
94.6%
Fashion
93.7%
Renewable Energy
91.6%
Medical Devices
90.5%
Payments
82.1%
Food Delivery
80.1%
Podcast Solutions
79.5%
Pharmaceuticals
72.0%
Semiconductors
63.6%
0%
avg.
The method behind this analysis is described in the appendix.
New technologies bring above-average risk: Besides big data (80%), other technologies only bring additional struggles. The more unexplored a technology is, the lesser its odds for success. No wonder ventures powered by the latest buzzwords of the blockchain (93%) or virtual reality (90%) fall apart in the sky-scraping numbers.
The failure rate of new business ideas according to a type of applied technology:
Blockchain
92.7%
Virtual Reality
90.5%
Natural Language Processing
88.2%
3D Printing
87.2%
Speech Recognition
87.1%
Augmented Reality
86.4%
Image Recognition
86.2%
Artificial Intelligence
86.0%
RFID
85.7%
Big Data
80.5%
0%
avg.
The method behind this analysis is described in the appendix.
But, but, but: There are signs that the low likelihood of hitting a home run isn't the only fate. There are dozens of serial outliers who succeed not every seventh time but every second or even less. To find them, I have switched the perspective a bit: from examining those who launch new business ideas to those who invest in them. And the latter group made me think.
The aggregated failure rate of new business ideas funded in 1990-2015 according to a backing venture capital fund:
number of bets
70%
100%
For details, check the raw data used in this chart.
The bottom line: Some have cracked the code of improving their chances, or at least they know what makes a good prediction for an endeavor to succeed. What secret formula lets them hit the jackpot with unprecedented regularity?
I don't know it yet. However, the research shows that the scientific approach to entrepreneurial decision-making works well for the de-risked navigation through uncertainties of new ventures.
by Gustaw Jot
APPENDIX: methodology of this study
Context: Measuring such a phenomenon as the success rate of new ideas is no mean feat. The official indicator is missing, and big companies keep such information out of sight. At the same time, small risk-takers have other worries than providing data for curious minds. So, I had to find a proxy for the analysis.
The data produced by venture capital funds has become a good trail. Why? These funds exist to invest money in promising ideas pitched by startups and hope for a return on their bets. The nature of their activity brings a strict definition of success: to build a new company to sell it for more money than invested. They have a fancy name for it: an exit.
Database selection: The choice came from the most comprehensive and up-to-date comparison of available sources. According to the research paper by Andre Retterath and Reiner Braun from the Center for Entrepreneurial and Financial Studies at Technical University in Munich, the most accurate ones are VentureSource, Pitchbook, and Crunchbase. Since Dow Jones withdrew VentureSource, I stuck to Crunchbase as the most dimension-rich database, with high-level results double-checking with Pitchbook.
Metrics definition: Venture capital funds classify the maturity level of backed startups. What is more, they have a specific category for early-stage ventures. It's called seed funding.
After reviewing the detailed criteria behind the classification process, I've decided to use the seed funding category as it aligns with the purpose of this investigation.
To analyze failure, I had to clarify what counts for success. In the venture capital world, it's straightforward: an exit in the form of an Initial Public Offering (IPO) or Merger & Acquisition (M&A).
The last part was to establish a timeframe for the analysis. That was crucial as it takes time to launch a startup, build its market position, and grow its worth to make an exit. The cohort analysis of the average time-to-exit gave me a reasonable cutoff date: December 31, 2015.
Average time-to-exit [in years] for startups funded in a given year:
2000
2005
2010
2015
Considering the downward trend, a 7-year buffer (2016-2022) seems quite a safe exclusion in my analysis. That gives a base of n = 53 845.
To sum up: The study was conducted in early 2023 and is based on startups funded between 1990 and 2015, backed by seed funding, and covered in the Crunchbase.
Pssst! If you spot any inconsistency in my logic or see an opportunity to make this study even more reliable, please reach out via Linkedin.
V.2.1
WARSAW, EUROPE