Read time: 4.4 min.

{{First Name | My friend}},

Every major initiative is built on a set of assumptions. Those assumptions shape the decisions that follow.

Some of those assumptions get surfaced and confirmed. Others don't. They simply govern the decisions underneath the initiative as though they had already been confirmed. There's no big announcement. They don't even register until something forces them into view.

That's what I want to talk about today.

Knight Capital Group was one of the largest market makers in U.S. equities. In early July 2012, the NYSE announced a new Retail Liquidity Program launching August 1. Knight's CEO committed to participating, giving the firm approximately 30 days to be ready.

On the morning of August 1, the firm deployed what appeared to be a routine software update. It looked like the initiative was on track.

Within 45 minutes, Knight Capital had lost $440 million. No, you didn't misread that.

The firm was effectively insolvent by the time trading closed and was acquired within months, ending its existence as an independent company. What looked like a routine initiative under deadline pressure turned out to be the manifestation of something that had been quietly accumulating beneath the surface for years.

Here's just enough of the technical picture to show you the pattern.

There was a piece of legacy code sitting dormant on Knight's servers for nearly a decade. In the push to meet the August 1 deadline, developers repurposed a flag in the system to activate the new trading functionality. What wasn't fully confirmed was what that repurposing would do to the old code still connected to it. The deployment itself was a manual process, with technicians updating eight servers individually.

Well, one server was missed. That server ran the old code. And because the mechanism designed to confirm that orders had been filled was broken, the system kept generating new orders indefinitely, with nothing to tell it to stop.

Each of those decisions was shaped by assumptions that were never made explicit to key stakeholders. I'm not placing a judgment on the assumptions themselves. The nature of them is immaterial. What was very material was that while the assumptions were implicit, their impact was quite explicit.

Most analyses of Knight Capital land on the technology failures and the governance gaps. Those observations are valid. I'd just encourage you to stay with me for one more step, because there's something upstream of it all.

Executive assumptions don't stay at the executive level. They shape the decisions leadership makes, and those decisions become the conditions every other part of the organization operates within. When the assumptions beneath those decisions stay implicit, they govern priorities, authority, and behavior throughout the organization without anyone naming them.

What the SEC found at Knight Capital is a documented illustration of that pattern. The SEC published a cease-and-desist order that found Knight did not have technology governance controls and supervisory procedures sufficient to ensure the orderly deployment of new code. It also found that Knight did not have controls and procedures sufficient to guide employees' responses to significant technological and compliance incidents. Perhaps most tellingly, it found that Knight's 2012 annual CEO certification was defective because it did not certify that the firm's risk management controls and supervisory procedures actually complied with the relevant rule.¹

Those findings are a downstream record of something that started further upstream.

Let's introduce one more wrinkle. On August 1, 2012, Knight Capital did not have a Chief Risk Officer. The role was created seven weeks later, as a direct response to what happened that morning.

Now, the organizational design questions this raises are genuinely interesting. But I'll leave those to the OD professionals. For me, this is all about decision architecture.

The question of what assumptions the executive team was operating from before a major initiative launches is a cross-functional leadership responsibility. It doesn't belong exclusively to any single title. It belongs to the decision architecture that creates the conditions for that question to get asked, across the right functions, before the pressure arrives and provides an uncomfortable answer.

On August 1, that architecture wasn't in place. And the pressure did the work instead.

The 30-day window that Knight's CEO had committed to didn't introduce the assumptions that were governing the deployment. It simply compressed the timeline in a way that ensured execution surfaced them. So, what had been accumulating across functions became undeniable on the morning of August 1.

This is a prime illustration that pressure doesn't create implicit assumptions. It reveals them.

Every major initiative your leadership team is currently driving is built on two categories of assumptions.

The first category: assumptions that have been surfaced, made explicit, and confirmed across the functions carrying them. The second category: assumptions that are shaping decisions right now as though they had been confirmed, even though they never were.

The latter, left unchecked, tend to have an outsized impact.

There are assumptions currently shaping your next major initiative. They may be as implicit as the ones that rocked Knight Capital's reality on August 1, 2012.

The difference is that you still have time to put the decision architecture in place to surface them, make them explicit, and ensure your leadership team is working from a shared understanding that advances your objectives.

If you'd like to talk through what that looks like in practice, I'd welcome the conversation. You can schedule time directly here.

Until Next Sunday,

Shawnette Rochelle, MBA, PCC
Founder, Excellence Unbounded
Executive Decision Systems That Drive Organizational Momentum

If you’re curious to learn more about my work with executive teams, you can find it here.

If you want to have a conversation to learn more, schedule it here.