According to Conway’s Law, an organization often looks like its architecture. This can go one of two ways: If an organization is well-designed around its architecture, teams have clarity on their roles and can work from a solid operating model. But if an organization’s architecture looks like spaghetti, their teams and processes also end up tangled. People get confused, and fragmentation accumulates a lot of technical debt.

That’s why CIOs or CTOs need to drive alignment between organizational structure and technical architecture with the CEO, the CFO, and others. That leadership team is not necessarily familiar with the depths of technical operations, so we have to speak in terms they understand — how the critical organizational and people strategy decisions we make will impact their ability to deliver on growth goals. In other words, the engineering strategy needs to be designed around the business mission, product, and existing talent. To that end, we must work backward from the customer.

Aligning people strategy and tech decisions

We best serve the customer by delivering excellent products that disrupt the market, uplevel the company, and fuel a multi-year product strategy. A great tech strategy can stand on its own to some extent, but it must draft off that business plan — meaning it must accelerate growth and shorten the time to market. But to make this happen, you need the right people and the right people strategy.

I recently sat down with John Vlastelica to discuss the idea of talent acquisition and the role it plays in aligning tech strategies. As companies grow and find their product and market fit, tech decisions should increasingly influence both organizational and people decisions. How you organize people and the operation model they follow will determine the decisions that CTOs/CIOs make and how they get communicated among teams, including finding the right balance between engineering generalists and specialists.

The most effective companies reduce the friction to onboard and make building products easier — those companies tend to have more generalists. I’m also a believer in less is more: Our SMEs don’t want to learn yet another programming language. With fewer languages and, in some cases, fewer tools or frameworks, I can onboard more generalists faster and make them more productive. The more focused engineers are on solving business problems, the better the product and the happier the customer. People and culture are everything, regardless of strategy. Eventually, specialists may be required, but that decision to change the engineering mix comes back to the tech strategy and its alignment with the leadership team and our business goals.

Single-threaded ownership

The need to drive organizational design change is more pronounced at companies where product and engineering are kept separate. Functional orgs don’t always operate cleanly. When teams are spread apart, it becomes harder to share the same goals, while dependencies only get more complex and require more coordination across teams — and lack of coordination is the enemy of efficiency and productivity.

The best way to doom a strategic initiative is to make its leadership someone’s part-time job. Yet, far too often, this is exactly what happens. Project managers are not given full authority to own the initiative. The CEO declares it to be critical, but no one is empowered to make it happen end-to-end.

Take an example from my time at Amazon of how we can do it differently. We changed direction during the development of a personalization widget. The initiative had amazing potential, but when our teams were organized by function, there was less clarity, and performance suffered. So, we decided to create a single-threaded owner (STO) who had the power and scope to get us as close as possible to a 100% conversion rate.

The STO owned the stack, data, algorithm, success metrics, everything. Once we made personalization its own mini company (and the single-threaded leader its CEO), it took off and is now one of those great products that seems to have lasted forever. The STO model turns strategy into real results with support from leadership, but integrating this approach should not be rushed.

Invest wisely to make reorgs work

When companies begin to recognize the inefficiencies of their org design, they often get too excited and say: “I want cross-functional teams. I want to do what Amazon did.” They then revert to what they know best — reorganize — but without making the necessary investments. In their haste to enjoy the speed and focus of cross-functional teams, they fail to properly understand how to build people, architecture, and the operating model into the strategy so that teams have the autonomy to be successful.

It is critical to get this strategy right and align it across the organization. It takes time. This is an investment for the long term because the decisions that CIOs and CTOs make to bring engineering, product, and business into alignment are aimed at transforming the company into an outcome- and product-driven organization. Then, when organizations minimize the need for coordination across teams because autonomy and ownership have been made part of the architecture, it’s safe to say they are getting on the “right” side of Conway’s Law for the future.