Managing Risks arising from Expeditious Start-Up Innovation: Invisible Hand, Regulation or Self-Governance – A Trilemma

Entrepreneurship and a well-functioning start-up eco-system is an essential part of an innovative and competitive economy, and the society.  They are the forces behind creative destruction that is expected to lead to improved economic and social well-being.

Start-up Eco-System

The early stages in the life cycle of start-ups are already familiar to many of us.  Founders have an idea, perhaps a technological breakthrough, and they start their small firm with savings of their own and any other funding available.  Once they cross a threshold in product or idea development, they seek funding from venture capitalists and private equity firms.  If the product is successful, then the firm grows bigger, and probably at some stage goes for an IPO. Venture capitalists and private equity firms are specialist firms that have the capability to undertake detailed risk assessment, appetite for risk taking, and are expected to provide mentoring, strategic and governance oversight, as they have a skin in the game. There is not much to find faults with, except perhaps on the margins.

Transformation of the Start-up Eco-System and the Risks that Come Along

A confluence of factors has contributed to making entrepreneurship in software or information technology sectors more common and impactful than was possible, say, a decade back. The falling costs of hardware including processors and storage options have brought entrepreneurship within reach of individuals in a way that was not possible earlier. Equally, there are now a range of new technologies, such as machine learning, robotics, virtual reality, artificial intelligence, voice recognition, image reading, and many more, in various stages of development. These technologies, and their various combinations, when combined with different types of data, namely, text, voice and images, offer niche entrepreneurial opportunities that, if executed well, could make a transformative difference to many problems that were previously considered intractable or too expensive to solve particularly in the areas of medicines, healthcare, education and finance.   When the breakthroughs are adequately validated, they have potential to improve life outcomes. If not, they carry significant risks to a common person likely to use them, particularly if innovation is rushed through or unintended consequences are not carefully tracked.

We are seeing seemingly random examples that suggest that the above is no longer a theoretical possibility and there is a need to engage in a discussion before it is too late. Perhaps a well-known and extreme example is provided by the familiar story of Theranos, a firm which claimed to have technology that could revolutionise the blood testing. At various points in time, the firm had George Schultz and Henry Kissinger along with many other prominent public names on its board. The firm at one point had market capitalization of $9 billion. Following the exposure by Wall Street Journal1, it turned out the firm’s claims were not entirely correct. A useful perspective on the firm can be accessed on Gates Notes2 in his review of the book, Bad Blood3, a winner of the Financial Times/McKinsey Business Book of the Year Award 2018.

Similarly, famous brain-training web-site, Lumosity, settled with FTC for $2 million with respect to its claims.4, 5 Cathy O’Neil in her book, Weapons of Maths Destruction6, documented cases of use of algorithms in important domains (such as education, correction facilities, and getting a job), which had important life-outcomes for individuals, but unfortunately were not transparent. Various examples in Cathy’s book suggest that the possibilities (algorithms and use of big data in her case studies) offered by new technologies are such that even established set-ups with resources could struggle in getting it right.  Similarly, there is now a growing concern on the increase in popularity of the DNA testing and claims being made on its’ behalf 7.

The concerns above were not of great significance earlier because start-ups did not have the range of technologies, data choices and the means that exist today to explore complex challenges. The software, therefore, addressed relatively less complex issues that generally did not have a significant impact.  Earlier, the harm from an innovative software could be errors in the telephone or grocery bill, or a refund that did not happen, or poor recommendations on films or songs. Now, the potential harm with rushed innovations could be wrong medical diagnosis, ambiguous improvement in particular health or education outcomes, denial of employment or recruitment opportunities, or perhaps a wrong decision on immigration/visa application or denial of an insurance claim8.And this is perhaps a beginning given the endless possibilities in which new technologies could be used and claims for innovative solutions could be made.

What is the nature of challenge?

It is important to clarify that this note is not about specific apps or web-sites per se.  This note seeks to initiate debate about risks associated with innovative breakthroughs, particularly in the context of start-ups that use hardware and/or software technologies, where the analysis, findings or recommendations could have important, perhaps irreversible, outcomes for an individual and/or family.

In doing so, this note recognizes that many of our complex and intractable problems in the areas of medical and health care, energy, education or economic well-being, could be more effectively and speedily solved through technology driven innovation. This, if anything, adds to the real risks, as both the rushed “supply” of innovation and “demand” (or expectations including a willingness to quickly believe success stories) are likely to go up in a significant way. The competition to stand out could trigger an arms race –and market success could provide incentives for rushing through “innovation” perhaps with compromises. It can be argued that such risks are inherent in any innovation process, and perhaps this is the essence of entrepreneurship. To use Milton Friedman’s famous words from a different context, “there is no free lunch”, though it can also be argued that we may want a reasonably priced lunch and not the one that is too expensive. In addition, the risk is now increasingly on the consumers as well, and risks to consumers and entrepreneurship man not be aligned, particularly in the case of expeditious innovation.

While there is already a debate on data privacy and its harmful use, it can be argued that the debate started a little late. It is important therefore that a similar debate be initiated in real time with respect to “innovations” and use of technology in the areas of healthcare, education, finance and other important sectors. It is perhaps not practical, or even needed, to decide whether the problems from such unproven innovations could prove to be a bigger challenge than say data privacy and its misuse. That said, it is possible to consider additional angles as we think more about creating the right outcomes that work for everyone.

  1. Expeditious innovations and products (including algorithms), if they are not fully tested and validated, and not transparent, are more likely to adversely impact individuals, who perhaps are most vulnerable or lack institutional support to protect their interests. This was one of the messages from the case-studies documented by Cathy in her book. This could over time lead to more inequity and differentiation in society. One such example relates to “recidivism risk scores”, where “the scores themselves are calculated in problematic ways”9
  2. The institutional framework and support may not be robust enough to act as guardrails in case of any deviant behaviour from the innovative start-ups. Theranos was operating in an area (blood testing), which had regulation. However, the complexity of regulation did not help detect the problem sooner.  Another limitation of relying on regulatory framework is that it varies across regions and countries.  This indicates a possibility of regulatory arbitrage, and at the very least the risks of “innovative” technologies being tested in regions with weaker frameworks;
  3. Such new products and services have a data angle to them, and this data could be sensitive and important. As the Facebook and perhaps other less familiar examples show it is not easy to build guardrails around them, and what is adequate privacy and safety for one service or one technology at a point in time may turn out to be inadequate later;
  4. In digital world, the best does not always become the industry standard or a norm, as examples from early days of technology show. “A de facto standard emerges when, in a growing market, one way of doing something gets a slight advantage over competitive ways of doing that thing”.10 (pp 50)  

This implies that an expeditious innovation, even when not fully tested for its’ unintended consequences, could under certain condition may become industry norm, and the best may simply lag behind.

We are now slowly beginning to develop some appreciation of market forces in the context of intangible economy.  Jonathan Haskel and Stian Westlake in their book, Capitalism Without Capital11, identified four S’s of Intangible Investment, namely, Scalability, Sunkenness, Spillover and Synergies.  These four characteristics, inherent in an intangible economy, could result in development of like-minded eco-system of small and big firms, where it is not easy to ensure and monitor the alignment of interests among various parties in a robust way.  This eco-system, when it is dysfunctional, is a more serious challenge than deviant behaviour by a single firm, as the recent example with Facebook has shown.

The nature of challenge, therefore, is:

How to harness the potential of various new and emerging technologies to the fullest to address the most complex problems, but in a way that their potential harmful impacts in domains, which include important individual life outcomes (as in health, finance, education and others) are eliminated or at least addressed on a pro-active basis?

What are the options?

Given the significant role that new technologies and start-ups could play in addressing some of the most complex challenges in different domains of economic activities, what are various options to proactively mark out “good” innovation from the “bad” innovations?

In general, the three options are:

  • Market forces, the famous invisible hand of markets,
  • Regulation, and
  • Self-governance or Self-awareness.

Invisible hand of market forces is currently the main mechanism to separate the good companies from the bad ones, and, based on the examples available, it is not very clear how effective it is likely to be in uncovering the issues and providing a voice to the risks. The efficacy of market intervention also crucially depends on the timing of intervention, and if it is late, it may make it even more difficult to address the risks. Also, as noted in the previous section, we perhaps need better understanding of how the invisible hand might work in an intangible economy.

In a market-based system, consumer preferences are considered to be a more effective way of disciplining suppliers.  However, the effectiveness is function of the extent to which the individuals have a choice, which may not always be the case. Innovations sold for reasons of lowering costs (particularly in health care) are likely to adversely impact (or benefit, if innovation is right) the more vulnerable sections, who may not have other choices.

Regulation has enabled progress in the areas of data privacy and right to be forgotten in the EU context. That said, as a corrective mechanism, regulation brings many challenges with itself. It is usually expensive (which makes it even harder for start-ups with limited resources), takes time, and creates its own unintended consequences. This is somewhat further compounded by the fact that entrepreneurship is hard as it is.  Since public policy generally seeks to encourage it, it probably will not want to make entrepreneurship any more onerous or difficult than what it already is.

Self-governance is another approach that is available for dealing with the risks arising from expeditious innovation. While self-governance is ideal, it is perhaps more difficult particularly because of the first mover disadvantages, arms-race to get to market first, and the celebrated management wisdom in the digital work of failing fast to be able to improvise fast. This last practice, while good for consumer software and products, is debatable when used in products or software which influences important decisions, such as medicine and finance choices. In addition, there is an emerging organizational tendency, particularly in the resource starved (cost cutting) context of both public and corporate sectors to seek /prefer technology driven innovations that cut costs as opposed to, or in addition to, innovations that solve difficult problems or create superior experiences.  When the principal organizational motivation behind adopting an “innovative” technology is to cut costs, it has the potential to change innovation process and priorities as well as the definition of “success” used by the start-ups.

A Possible Framework for Resolution of Trilemma

An effective solution would need to be open and flexible to reflect the emerging practices and the context. It also needs to combine elements of all three approaches identified above.  The specific framework would vary by the sector, the nature of innovation, consumers involved, and the regions. With these caveats, following is a broad framework, which could be used as a start point.

  • There is perhaps a need to identify sectors, which will require higher scrutiny. These sectors could be where the impact of “innovation” is irreversible and significant for the individual. They include health care, finance, judicial processes, education and data ownership. These sectors could depend on the public policy goals of a given region or area, and this list could evolve. Innovations in these sectors could be made more open and transparent, subject to IP protection, on testing done, data used, matrices and methodology used to validate the outcomes, complaints process and transparency obligations, and risk factors. Some of these, such as complains process and transparency obligations are being used in the financial sector.
  • In these “high risk” start-ups, the management team and board equivalent in the firms should have obligations, which extend to the public impact of the products and services made by the firms. Usually, the skin in the game for management team and board equivalent are mainly around the financial success, though other obligations exist as well.  The key idea is that the risks to the management body or the board equivalent should be proportionate (and if possible not less) to the risks that the consumer will bear out of erroneous performance of the product, if it is a result of expeditious innovation;
  • Perhaps individuals, both as consumers and employees, have an important role in encouraging the right kinds of innovations. It is difficult for employees to influence the right processes beyond a certain threshold, particularly if the leadership team has other priorities or not ready yet. It is perhaps still important to ask right questions, and recognize that, everyone (employees, managers, or leaders) could be at the receiving end of some start-up innovation at some point in the economy, because of similar decisions made and thought processes used in the other firms. An effective option is therefore to raise the bar in one’s own limited context, and hope others in similar firms are doing the same to make it better for the economy as a whole.

As a consumer, an individual could also start asking questions about the methods, processes and comparative data, particularly when they appear out of range or counter intuitive, or simply “too good to be true”.  The bar for the “right” innovations will go up if individuals make choices, to the extent they have choices available, on how well and transparent the firm is in sharing the processes and explaining the methods of innovation, how well its’ staff knows and able to explain the workings of “innovation”, how impatient or inattentive staff becomes when asked probing questions, and whether or not they are empowered to escalate or trigger the need for manual over-ride.

Most of the examples that we know of come from the US context. It is not clear whether the lack of similar practices from other geographies is because of lack of information, or presence of right practices already, or simply that the pace of innovation in other regions has not picked up to generate enough data as yet.  It will be useful for each geography to understand the context and be more informed and deliberate in using technologies, particularly when they drive or facilitate the innovative service or product.


  1. John Carreyrou, “Hot Startup Theranos Has Struggled With Its Blood-Test Technology”. The Wall Street Journal, Oct 16, 2015, accessed 21 April 2019. https://www.wsj.com/articles/theranos-has-struggled-with-blood-tests-1444881901
  2. Bill Gates, “I couldn’t put down this thriller with a tragic ending”. Gates notes, December 3, 2018.  accessed 21 April 2019.  https://www.gatesnotes.com/Books/Bad-Blood
  3. John Carreyrou, Bad Blood: Secrets and Lies in a Silicon Valley Startup. (London: Picador, 2018).
  4. Press Release, “Lumosity to Pay $2 Million to Settle FTC Deceptive Advertising Charges for Its “Brain Training” Program”, Federal Trade Commission, January 5, 2016. accessed 21 April 2019. https://www.ftc.gov/news-events/press-releases/2016/01/lumosity-pay-2-million-settle-ftc-deceptive-advertising-charges
  5. Lumosity Blog, “A letter to our members on the FTC settlement”, Lumosity. 11 Jan 2016. accessed 21 April 2019. https://www.lumosity.com/en/blog/a-letter-to-our-members-on-the-ftc-settlement
  6. Cathy O’Neill, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy, (Penguin, 2017).
  7. Internet Health Report 2019, “23 reasons not to reveal your DNA”, Moz://a, April 2019. Accessed 27 April 2019. https://internethealthreport.org/2019/23-reasons-not-to-reveal-your-dna/
  1. “Machines and Morals I The Why Factor” BBC Podcast. 23 minutes. 26 March 2018. Accessed 27 April 2019.  https://www.bbc.co.uk/programmes/w3cswrjp

Note: There is a short discussion on robot ethics and transparency in AI driven decisions (3 to 4th Minute of podcast) and then on temporarily denial of insurance for Virginia Eubanks (4th and 6th Minute of podcast). Virginia Eubanks is also author of book titled “Automating Inequity”.

  1. Read Highlights, “Cathy O’Neil on Weapons of Math Destruction”. The Library of Economics and Liberty. 3 Oct 2016. Accessed 2 May 2019. http://www.econtalk.org/cathy-oneil-on-weapons-of-math-destruction/#audio-highlights
  1. Bill Gates, The Road Ahead, (London: Penguin Books, 1996). Pp 50.

Note: A short, but informative analysis on de-jure and de facto standards, and how VHS format won over Betamax format in the video cassette recorder battle of late 1970s is provided on pages 50-51 of the book.

  1. Jonathan Haskel and Stian Westlake, Capitalism without Capital. The Rise of the Intangible Economy”. (Oxfordshire: Princeton University Press, 2018).




Leave a Reply