A primer to Organisational Longevity

9 min readJust now

–

Imagine you’re in a company experiencing rapid market change, challenged by new competitors with innovative products and business models. You’re at risk of becoming disrupted, unless you innovate yourself.

You have a choice between two AI projects. Both cost €1 million to implement.

• Project A will give you a return of €1,1 million
• Project B will give you a return of €50 million

Obviously, Project B is the best choice. But what Project will your organisation choose?

The answer is not as straightforward as you’d think. In this post I’ll show why the struggle to adopt AI in most organisations doesn’t revolve around AI. It revolves around what projects we choose to implement. I’ll show why so many organisations choose the wrong projects, the deeper thinking that’s at play, and what we can do change our ways.

Markets that don’t exist cannot be analysed

The term Disruptive Innovation was made famous in The Innovator’s Dilemma published in 1997 by the late Clayton Christensen (Harvard Business School). He analysed why incumbent companies so often fail in competition from new and innovative products, and how the pattern repeats itself over industries and decades, from steel mills, to hydraulic lifters, to harddisks to Gasoline Cars — from the early 1900s up until today.

Christensen writes he initially thought the companies that got disrupted just had bad management. But he soon discovered this was not the case. On the contrary, management was competent, responsible and acting in accordance with established principles of well managed companies.

However, when it came to innovation (or lack thereof) that also turned out to be the problem. As he explained:

The best-run companies fail, not because they’re poorly managed, but because they are too well managed to innovate. (Clayton Christensen, Innovator’s Dilemma)

Think about it: Market leaders have a lot at stake. Ran by responsible people, they are drawn towards protecting what they have. Why would they want to pursue high risk endeavours, that on one hand are unpredictable and unlikely to give any meaningful returns, and on the other hand could cause them a lot of trouble? It’s not attractive. Market leaders have instead learned to value stable conditions, efficient operations and tight control.

As Christensen explains, there is nothing wrong this. Except one thing. It doesn’t work with innovation. Here, we’re supposed to:

• Develop inferior, new products with defects, inefficiencies and lack of functionality, and sell them to our customers.
• Invest in new opportunities that seem small, have no clear market and no track record of success.
• Allocate resources into something unknown and unknowable.

Of course we don’t. No responsible company would. Right?

But the world around us changes. It brings about competing products, trends or infrastructures that our current capabilities were never built for. Now, in the age of AI, change goes even faster.

By doing the right thing, by staying the course, we risk going down the wrong path. We get disrupted, quite possibly in a well-managed way.

The alternative is to go with the very projects we have learned to avoid. So what do we choose? Go under, or break our principles? This is in essence the Innovator’s Dilemma.

Is there an other way, where we can stay true to ourselves and succeed with innovation?

Building companies that last forever

In 2003 the CEO of a successful gaming company pondered what to do, after realising they were losing business to competitors, who had developed products with better performance. Everyone expected him to focus his resources on improving their capabilities — to catch up. But he didn’t. Instead, he started making changes to his products that his customers didn’t care about, to serve a market that barely existed.

Was he foolish or strategic?

The gaming company was Nvidia. A backbone of today’s AI infrastructure, they are now one of the world’s most valuable companies. This was not the case back in 2003. Losing badly to competition, CEO Jensen Huang made a crucial architectural decision: Their GPUs would use 32-bit IEEE compatible floating point, which was not a gaming industry standard, but it was better suited for general-purpose computing.

It didn’t help them win the gaming war. But it created space for something else: It enabled researchers to use their hardware. The decision seemed odd, because this “data science” and “AI” was all very niche. But Nvidia saw a learning opportunity. They wanted to see where it could lead. Later, when deep learning came, and ChatGPT some years on, they found themselves at at the epicenter of a revolution.

He was not foolish. He was strategic. About Options.

Optimising for renewal

When change goes faster and faster, companies have to optimise along two two dimensions:

Press enter or click to view image in full size

Expiry Unknown: We don’t know when our Capabilities go bad

1. Capabilities enable us to be in business: Products, organization, systems, work processes, and IP that deliver value to customers right now. For Nvidia in 2003: Their Gaming GPUs.
2. Options enable us to do something different tomorrow: Opportunities within reach given our current constraints, like new markets, adjacent markets, new revenue streams, and substantial productivity gains. For Nvidia: The change to 32 bit floating point, opening up their products for general use.

Both Capabilities and Options work as feedback loops. When a capability starts performing, we optimise it by capturing larger market shares, reducing cost and increasing margins. We allocate resources and build our brand and reputation around it. It becomes self-reinforcing.

The same goes with Options: When we get good at discovering opportunities, we tend to find even more opportunities. As we experiment, we build skills and confidence in exploring the unknown. We develop networks and relationships that surface new possibilities. The more options we pursue, the better we get at recognising their potential.

Press enter or click to view image in full size

Capabilities and Options are interdependent. Successful development of an option turns it into a capability. But there’s also a reverse relationship: Developing a capability weakens our options.

This is because capabilities make us accountable: They come with commitments that tie up resources and focus. This narrows down our space of opportunities. We can’t go out and do random things. We have things at stake, and much to lose.

An organisation can only be good at what it does regularly. In today’s organisations capabilities are the comfort zone. They’re what we’re trained to develop and maintain. Working on options is less common. In most industries companies spend less than ~3% of revenue on R&D overall, and far less on true exploration. It means we don’t on a regular basis test and learn, iterate fast, and slice and sequence work into small packages that drive business outcomes, one increment at a time. This lack of innovation know-how further reduces our options.

However, in a fast changing world, our capabilities can start their decline any time. Hence, we need options to survive. Which leads us back to our two AI projects.

Project A or Project B?

Most organisations will choose Project A, the project that gives only a 10% return. They choose it, *because *it has a small margin. They avoid Project B because it’s so valuable. Here’s why:

1. **Focus: **If Project A slips by just 10% on time or cost, it’s a loss. Hence, it requires tight execution and control, otherwise we shouldn’t do it. The paradox, however, is that the more we focus on the control regimes we need to make such projects successful, the more likely it is that we’re also focusing on such projects.
2. Catch-22: Moreover, tight control regimes often can’t handle high-value projects. They typically kill them from the get-go by demanding upfront answers to questions that can only be found through testing and iteration, which wont happen unless the project is started.
3. Mindset: We’re assume Project B is more risky because it has a higher upside. But innovation doesn’t work like traditional investments. You don’t invest €1 million upfront. You invest €50,000 to explore whether Project B has potential. You talk to users, build prototypes and run experiments. You learn. If it fails, you’ve lost €50,000, not €1 million. Innovation is staged (you can even explore several investments in parallell).

Wait, we’re not done yet. There is one more reason why so many organisations go with Project A, and it eclipses all the other reasons.

They primary reason they choose Project A is that they are unaware of the existence of Project B.

They are unaware of Project B because they don’t explore their options*.*

Change the success criteria

In capability-optimised organisations, success means delivering predictable results in a controlled, stable and efficient fashion.

So what about discovering an option, or creating the conditions for the organisation to find its best options? Will that make a board member, manager or employee equally successful?

In most organisations, the answer is no. The reason is simple: To explore options, we also have to make learning about the unknown an equally legitimate measure of success. If you want innovation to happen on purpose, as an integral part of your strategy, this change needs to come from the top.

It can be very powerful: When Satya Nadella took over as CEO at Microsoft in February 2014, the company was in serious decline. Microsoft was caught up in old products and systems, unable to respond with the innovations markets wanted. This complacency had created a “know-it-all” culture. Nadella famously stated that Microsoft now has to become a “learn-it-all” company instead.

He started doing his own experiments, then told his direct reports what he had learned. He expected them to do the same with their direct reports, and so on, until the practice proliferated across the corporation.

The transformation was remarkable: Microsoft embraced cloud computing with Azure (now generating over $33 billion quarterly), opened its platforms to competitors like Linux and Apple, acquired LinkedIn and GitHub, and partnered with OpenAI to lead in AI. These were innovations that would have been impossible under the old siloed, Windows-centric know-it-all culture.

Microsoft’s market cap grew from around $300 billion in 2014 to over $3 trillion, surpassing Apple as the world’s most valuable company in January 2024.

Getting Started

There is also a more practical side to Options. To find and qualify our “Project Bs”, we need to increase the bandwidth of our search. This is what the options loop as all about.

How do you get full bandwidth? You empower the organisation to explore:

1. Pre-authorise learning initiatives: Enable people to do small-scale, low cost experimentation without having to ask for permission or approval.
2. Allow parallell development: Allow new initiatives to be built on the side of existing systems. Establish policies and guides, but don’t require green light from the functions they will depend on long-term. Provide easy access to resources and data.
3. Use Autonomous Teams: Use cross-functional teams that have the know-how to explore and build without having to ask for help or approval. Provide metered funding, that corresponds to the business value they create.

It has several advantages:

1. People who want a project to happen must first validate key assumption about it. The decision process becomes both educational and, frankly, fun, for everyone.
2. Bringing real-life data about proven business impact, gives significantly higher accuracy when deciding on what big bets to go for
3. People step up and take more responsibility for value creation, when they’re actually asked to do it.

It’s the responsibility of management to decide which options to develop further, and which ones to leave behind. By allowing their organisation to do the exploration, to be creative, they get more comprehensive strategic insights, that will help them make much better choices.

When they do want do double down, they will have a decisive advantage over those new companies trying to disrupt them: Once an option is validated, they can scale up faster and more effectively than any startup. They already have customers, distribution, brand recognition, capital and operational capacity. Startups must prove the concept and build the scaling capability and fundraise *and *convince the world they are for real. (It begs the question whether today’s incumbents are making it unnecessarily easy for the startups to disrupt them).

To get started with options we need to learn to use new practices, to work together in new ways, and to see our organisation and the outside the world in new ways. This requires a change in mindset, a cultural change. A fast-track to making this happen is to mix people and cultures. Bring in people from a culture that has already done it. Ultimately, you’d like them to have an entrepreneurial background, and also know their way around established organisations. Use such cross-pollination to build your new culture — bottom up.

How far could it go?

Established companies aren’t just businesses. They are part of the structures that shape how our economy and society function. If they weaken and lose trust, it makes society more brittle.

AI raises the stakes by challenging the premises of our capabilities faster than ever. In this environment, continuity comes from renewal, not preservation.

AI won’t save an organisation that can’t renew itself. But an organisation that can will thrive through any technological wave. That’s the real productivity unlock.

How far could this go if we succeed? Imagine your company is able to renew itself faster than the rate of external change. It can live forever. This is called *Organisational Longevity, *a larger topic that I will cover in my coming posts.

In the meantime: Open up. Stay curious. Build options.

Credits: Features / Options: Kent Beck. Project A, Project B: Tom DeMarco