## How Enterprises Should use Size as an Unfair Advantage to Innovate like Startups
The average age of companies in the S&P 500 is under 20 years–down from the 60-year average of the 1950s. An implication of software eating the world is that, just to survive, large organizations need truly breakthrough innovations. Many companies delegate this critical task of realizing startup-like breakthroughs to their corporate venture arms or isolated “innovation” groups. Yet if survival is what is at stake, for most companies the __core__ business *should be* transformative innovation.
My convictions about the importance of making multiple bets on enterprise transformation comes from over 1,000 projects I have seen over the last three years. In my experience, companies that run multiple, customer-centric experiments end up with much better outcomes. The best venture capitalists have honed a successful model for engineering breakthrough innovation and provide an instructive model for the CEO, CTO, CIO and key executives to repeatably engineer breakthrough innovation.
## Experimentation implies a minimum cost to innovation
It’s natural for large companies to worry about wasteful investments and to agonize over failed projects. This concern, however, is often misplaced and leads companies to under-invest in the right kind of innovation. In fact, to innovate at a transformative level, it’s key to identify the minimum investment required to achieve breakthrough results.
The root reason for this minimum investment level is that innovation requires experimentation. Experiments, by definition, frequently fail. If they were guaranteed to succeed they would neither be experiments nor innovative. The randomness of experimental outcomes implies that innovation is irreducibly costly because we have to place multiple bets to be assured of a successful outcome.
## Venture Capital as an instructive model
Venture capital (VC) investing offers an instructive parallel. The entire goal of a VC is to fund multiple experiments with the clear objective of finding unicorns, or industry-defining companies. Looking closer at VC investing models, we see a good benchmark for understanding how much it should cost to run an experiment and how successful the experiments can be.
1. __The minimum viable cost of innovation__ —If each experiment costs *c* on average and the likelihood of failure is *f*, then we can picture the Minimum Transformation Investment as *c/(1-f)*. Some interesting questions for organizations that arise: what is the average cost of an experiment that yields true innovation? What are typical failure rates?
2. __Most experiments will fail and each experiment is costly__ —Unsurprisingly many startups fail. However it’s actually the case that *most* startups fail. According to Shikhar Ghosh at Harvard, 75% of startups fail to return investors’ money. Since the goal of funding a startup is not merely to get a 1:1 return on investment, the failure rate climbs over 90% if we only look at startups that produce significant returns or go public. In addition, the average cost of the initial experiment (the seed round) is about $1.6M. Thus not only is there a minimum cost to innovating, the minimum cost is high–a VC is willing to commit about $1.6M per bet and is willing to make 4 -10+ bets in order to see just one pay off.
3. __The minimum viable cost grows over time but the failure rates go down over time__ —Although an experiment should start out small, as it becomes more promising, an increasing amount of investment is required to capitalize on the full opportunity. To get a sense of this, the sequence of experiments run in VC and the average investment is detailed below:
– Seed: early product market fit ($1.6MM)
– Series A: scaling product market fit ($10MM)
– Series B: scaling the business ($25MM)
– Series C onwards: more capital to scale ($50MM)
Overall, the median amount raised on the way to an IPO is $100MM and the biggest venture capitalists have $1B fund sizes. At a 90% failure rate it costs ~$1B to fund one successful breakthrough all the way to IPO-like status–a level at which one can claim to have transformed the parent organization.
Reliable innovation is not cheap, and since for most companies the alternative is extinction, it’s also not optional.
4. __Good News: The top winners pay for all of the failures many times over__ —When experimentation yields a successful breakthrough technology product, one great success pays many times for all of the other failed experiments. It is well known in VC that returns are power law distributed–all of the returns are concentrated in a very tiny number of investments. The top few performing investments are worth more than all the other investments in the portfolio combined. We can take the real example of Uber–where a $9M investment by Benchmark (~ 2% of the overall $400M fund) will return $6.7B or 740X the original investment and *17X the entire fund!*
Thus, to achieve breakthroughs, attempting to avoid failure is misguided. Failure is inherent to large-scale innovation but the resulting successes can be extreme and more than compensate for the many losers.
## How to Win: Adapting Lessons from VC
While enterprises may have venture capital arms, most are not set up to use these principles as part of their __*core innovation process*__. But here is why and how you should:
1. __Place Many Simultaneous Bets__ —Because the Minimum Transformation Investment is high and the cost also grows over time, an enterprise company with deep pockets has a significant advantage. Why? Large companies can place multiple bets concurrently and run several experiments in parallel, resulting in a faster path to breakthrough innovation. Startups typically can only run them quasi-serially–not nearly as efficient or effective. In working with F500 customers and startups to deliver over 1000 projects, I’ve seen these ideas in practice with a few recent F500 clients attempting to improve their customer experience. The companies that run through multiple experiments with the customer at the center of it end up with much better outcomes and learn faster. Although many of these experiments are early, we are able to set up strong tests and quickly shut down paths that are less promising and pursue ones that track better towards the ultimate business objectives.
2. __Make sure the experiments are relatively diverse__ —Pick many small but different projects that teach you a lot about your customer and your market. Testing similar ideas leads to less learning overall and is wasteful. To get this diversity of experiments, collaborate with known internal contrarian teams as well as external entrepreneurial organizations.
3. __Set aside a large enough dollar commitment to finding breakthroughs__ — If a company gets serious about a bet–it should be willing to commit $100MM over 5-10 years in order to see it bear fruit. It should also be willing to fund ~10 such bets in parallel in order to find a break out. Of the top 100 retailers, there are only two in the top 10 for IT spend–one of which is Amazon. Given the current levels of spend of top software natives, many incumbents across all industries (and especially in retail) could easily be spending up to ~5X more dollars on breakout innovation. With the understanding that most of your bets will fail, commit to fund promising ones adequately to succeed through their phases of maturity. Over time, you will get better at deciding what bets to fund.
4. __Run each experiment like a startup__ —Alphabet, the parent company of Google, is an extreme example of this concept with each experiment literally run as an autonomous startup. Amazon is perhaps the example closer to what most organizations can and should aspire to–a company that re-invests almost all of its profits back into multiple experiments run in parallel. These risks have a high chance of failing and only pay off many years in the future but have yielded company-defining successes like Kindle, Alexa and AWS and have transformed an online bookseller into a device manufacturer and leader in cloud computing.
5. __Partner for speed__ —Lean on tried and true methods from how startups operate to execute optimally on each individual experiment. Or better yet, partner with entrepreneurial organizations that can bring the DNA of learning fast and failing fast so the winning ideas worthy of serious investment can be revealed quickly.
While companies need to start small within each initiative, the reality of their environments is that they need to make big long-term commitments to change. When under-investing is the same as extinction, the game is only worth playing if companies understand the magnitude of investment that is required. Established technology giants and startups armed with billions of dollars in funding as well as a mindset that embraces system-level failure and experimentation are waging war across all industries. For everybody else, the only way to win is to join the fray and fight fire with fire.