The Hidden Cost of Under-Imagining

Enterprises guiding their technology infrastructure through digital transformation face several barriers, such as technology change management expenses, policies regarding the use of open source software, and policies regarding the protection of confidential data. These have been written about extensively, but there is potentially less familiar barrier to adoption emerging now that hasn’t received enough attention: the risk of long-term commitments or very deep integrations with individual vendors.

The ongoing advancements in artificial intelligence (AI) and quantum computing are creating highly non-linear advancements in software engineering and performance. Advancements in model design and computing hardware are compounding rapidly, leading to a seemingly unpredictable future of how software engineering will evolve, how good AI will get, and which companies will win in this context. Some founders are preparing for Moore’s Law to collapse from a year or so down to weeks in certain areas, like the effect quantum chips could have on the time to decrypt data.

I’ve seen several investors and enterprise leaders conclude that it is best to observe from the sidelines until more firm signals emerge regarding which tech companies are winning. I suggest we observe from a different angle. I propose we accept that engineering and model sophistication is accelerating at an unpredictable rate, and then ask the question, “What other things will get disrupted by this reality?”

Terms and Timing

I’ve seen several market cycles in my prior SaaS and DaaS product roles, including cycles resembling the one we are currently in, where large enterprises were canceling or consolidating vendor contracts as well as negotiating lower pricing. From the vendor’s perspective, my teams did not have any legal obligation to amend contract terms, but we usually wanted to be supportive of our customer’s needs in exchange for reducing our future customer churn and for gaining other non-financial sources of value from the customer. What we really liked was renegotiating to (1) increase the length of the contract term (such as to 5 years) and (2) secure rights to create derivative works from our customer’s proprietary data.

Over a dinner discussion, a friend and I drew a corollary to the cloud infrastructure decisions CTOs faced in the early 2010s between AWS and Rackspace. There were solid reasons at that time why AWS wasn’t ready to meet many large enterprises’ requirements. However, dismissing AWS and overcommitting to Rackspace on a five-year deal in the early 2010s could have been a very costly decision based on AWS’ pace of advancement. This doesn’t disregard the value of waiting to see signs of a leader emerging in the category; committing to a five-year contract with Rackspace in 2007, a year after the release of S3, might have been a better choice than adopting EC2 in 2008. But it does highlight that contract terms can present real risk in a rapidly changing environment.

When an enterprise commits to long contract terms, it and the vendor both gain more budget certainty, but I believe that the enterprise often receives a lower cost benefit, especially when an enterprise is dealing with mature, cash-cow-style vendors. Longer contract terms relax the pressure for the vendor to innovate. They allow the vendor to shift its product focus to other customers and prospects that are at greater risk of churn or that could have a positive effect on the vendor’s near-term earnings.

Opportunity Costs

In a slow-moving innovation environment, this tradeoff doesn’t materially reduce the cost benefit for the enterprise. The enterprise is likely only missing out on incremental innovations, and if it needed to, the enterprise could engage in a services contract with the vendor to increase the prioritization of a product feature that they care about. In an extremely fast-moving innovation environment, though, being locked into a long-term contract with a vendor that is less equipped to embed the innovation can expose the enterprise to significant competitive disadvantages.

Think of how rapidly generative AI has changed in the past 9 months. Now assume the rate of change is accelerating, so that in the next 3 years the advancement may be 20-30x what was observed in the past 9 months. Delaying adoption of forthcoming new technology by years could be extremely impactful to both the top line and the bottom line. It wouldn’t surprise me if many enterprises selectively choose to incur cancellation fees next year to free themselves of contracts with low-innovation vendors, thereby offsetting the expense savings they were seeking.

Protecting Optionality

If you believe in the accelerating pace of innovation that I’ve described, you may want to take the opposite position of what traditional procurement departments seek. You may want to counterintuitively renegotiate to secure cancellation rights and to design workflows to have light vendor integrations. Optionality may be more valuable to the medium-term health of the business than near-term cost savings.

It is difficult to justify paying more for contract optionality because the human brain isn’t wired to accurately calculate how rapidly compounding returns grow. Academically, this is called the exponential growth bias, meaning that we consistently underestimate exponential growth. This bias shows itself in many domains, particularly in financial decision making. I encourage what-if thinking to overcome this barrier. What if the capability of X were to increase by 30x in 3 years? Where and how are we unprepared for this change? How can we create more optionality for our organization to benefit from this change if it does occur at this pace?

Kyle Beatty is a Managing Director with American Family Ventures. This article is for discussion purposes only and is not intended to be, nor should it be construed or used as an offer to sell, or a solicitation of any offer to buy limited partnership interests in a fund managed by AFV.