The smallest software projects are 10 times more productive than large software projects. Though large companies didn't like this, they benefited from it. They have been the only companies able to afford the innovation these large projects delivered.
Twenty years ago software had to be large. Every software company had to build it's own security, persistency, messaging, scaling, resiliency and display mechanisms before the business logic had a place to live. Building out all these capabilities is one of the reasons why software companies have ended up employing so many people. The huge costs of these large projects was also a reason for creating the Innovation Funnel. An approach to focus the large sums of money on the projects the company had decided most likely to succeed.
The World Has Changed
Amazon and many others now offer cloud based platforms with built in security, persistency and the other mechanisms. The price for running the platforms are very low, scalable with the swipe of a slider and charged by credit card. The barriers to entry have all but disappeared. Tiny companies are now able to compete in a strategy of asymmetric warfare against the large heritage vendors. A war these tiny organisations are winning.
However there are examples of newer, now large companies holding and winning in the new world; Amazon, Google, facebook and Twitter. All these have more than a thousand developers and are dominant in their markets. What are they doing that's different?
Common and key to the success of both startups and the new successful large players is their acceptance that they don't know what their customers want. This market humility drives the question of how do they find out what the customer wants? Answer: safe to fail experimentation to learn what the customers love.
These companies know that their success relies on building a culture that values both successful and unsuccessful experiments. Rewarding learning. They start with a question not an answer - The Secret Phrase Top Innovators Use.
It also means these organisations have continuously invested heavily in deployment, experimentation and personal learning infrastructure. To make fast, small and easy to reverse experimentation happen. It's safe to fail in these organisations. They can also run hundreds of experiments in parallel across different users to gain more insight with less risk. For example at any one time facebook is running around ten experiments at once on an any individual user.
Experimentation Benefits from Scaling
Writing large software products has dis-economies of scale. Writing software projects to experiment has scaling benefits in terms of diversification. Like finding a new drug to kill chronic disease or spreading investments across many companies, the more experiments you run the more likely you are to have one that identifies an improvement. A model well understood by pharmaceutical companies and venture capital funds.
Large software companies have the cash to fund more experiments. Even better they employ enough people to build many small experiments in parallel. They can have many teams independently experimenting on how to improve the same customer journey. Quickly deploying experiments that the improvee the customer journey and switching off those that don't. Sharing lessons learned.
Running many experiments in parallel to solve the same challenge is a strategy that gives large organisations an edge over their start up competitors. Start ups are not hampered by a lack of ideas but by their ability to identify the killer ones. They can't build experiments fast enough.
Killed by Cost Based Accounting
The challenge to this multi-experiment approach is that it appears, on the surface, to be very wasteful. Especially if you run an organisation using cost based accounting, which most heritage companies do. Where all you see is the cost of building experiments. You especially see the cost of those whose output is never released to a customer, multiplied by the cost of number of teams experimenting. It's always easier for accountants to measure cost than it is to measure value.
The Toxic Funnel
The Innovation Funnel is the classic cost based account model to address this. It's based on reducing the cost of innovation by killing ideas early and gambling on a smaller number of big bets. The problem with this is that it plays to the small nimble startup strengths again. They don't need deep pockets to play this game, they can run small numbers of experiments and they're better at it.
Large organisations need to make optimising for value generation and effectiveness over efficiency a key strategy.
High Stakes - Game it in Your Favour
The stakes are high. Having, or not having, high value features to frequently ship means the difference between gaining or losing millions of users within months. In such turbulent times this makes the cost based accounting model and it's innovation funnel highly toxic.
Large heritage organisations need to look at their strengths over startups or become irrelevant. They still may have an opportunity to win by:- *Focusing on market humility - acknowledging that they don't know what the customer wants *Building a culture of experimentation to find out *Accepting failed experiments as valuable learning *Funding many experiments to increase the probability of finding the killer feature *Funding many experiments in parallel to find killer features faster
Now it's down to the executives in these organisations to have the courage to experiment. Maybe the Innovation Funnel is a good place to start.