How to ensure experiments leave the launch pad

“Flight heritage” is how the commercial space industry likes to describe a record of success established by a string of missions for a satellite or a launch system.

The phrase cropped up last week in postmortems of Virgin Orbit’s failure to launch the first commercial satellites from western Europe. “Space is hard,” said Alice Bunn, president of UK trade body UKspace, after an “anomaly” condemned the LauncherOne rocket and its expensive payload to a fiery end before it reached its target orbit.

What this blot on Virgin Orbit’s heritage underlined was the inevitable commercial tension between experimentation and predictability. Pioneering space ventures are in a constant cycle of “working out what happened and getting up and trying again”, to paraphrase Bunn. Trial and error is how they advance, but terrestrial companies are not so different. In both cases, customers would prefer not to be on board for the “error” part. The challenge is deciding when to crystallise the findings of the experiment.

Back on Earth, the pandemic was a trigger for forced experimentation at many companies that had previously assumed the status quo in how, where, and when we work was immutable. Now they have to decide when to incorporate changes in working practices into a more predictable framework. Even the World Economic Forum was obliged to try out a summer summit in 2022, before reverting to its regular January timetable. One question is whether corporate leaders will return to the Alps this week more open to the need to experiment or keener than ever to revert to pre-Covid certainties.

“It’s very risky not to experiment,” claims Costas Andriopoulos, innovation and entrepreneurship professor at Bayes Business School and author of Purposeful Curiosity. Teams that were curious, open-minded, and accepting of failure responded better to the sudden lockdowns of 2020, he believes. “If you don’t experiment for a long time, then you bet the house and you either win, or lose everything.”

That is a view shared by Phil Libin, co-founder of mobile app Evernote, who now runs All Turtles, a product studio that has adopted distributed working. The company has just implemented a seasonal structure, in which employees commit to a six-month stretch at the company, followed by a two-week holiday taken at the same time by everyone. Libin freely admits it is a trial: “It’s guaranteed that in a season or two we will say ‘some of these things don’t make sense, we have to change it’.”

All Turtles has more flexibility than many to use itself as a laboratory. “Once we become a big company then predictability of execution is rewarded much more by markets and we’ll have to experiment less,” Libin says.

But much larger groups can also pilot new ideas, without betting the house. Unilever was able to test a radical shift to a four-day working week with an 18-month pilot in New Zealand, before rolling it out to some employees in Australia last year.

Elon Musk, who, through his leadership of SpaceX, knows more than most about how hard space is, has also brought real-time testing into the open at Twitter. The debate between those who want him to leave the familiar, predictable features of their favourite social media platform alone and those who applaud a state of rolling revolution has been vocal.

The cost and risk of a “perpetual pilot” approach are clearly lower for a digital company, which can change the customer experience at the touch of a button, or test multiple options simultaneously. Yet experimentation is a mindset unconstrained by a company’s size, age, or sector. Whatever they think of Musk as chief executive, other companies could learn from Musk the experimenter’s effort to involve users in his experiments.

In the latest edition of Harvard Business Review, Rita McGrath and Ram Charan write about how digital technology can liberate managers and their teams to experiment in so-called “permissionless corporations”, characterised by flatter hierarchies, faster decision-making, and better customer relationships. They cite etailer Amazon, lift maker Kone and fund manager Fidelity as companies that have changed the pace and nature of the work they do using technology. “In the permissionless corporation, fast, inexpensive experimentation takes over from slow, involved analysis, enabling organisations to pounce on opportunities as they arise,” they write.

There are pitfalls to running a company on perpetual pilot. Andriopoulos says sometimes managers “hide behind a lot of experiments because they don’t want to move ahead and take a decision”. Eventually, Virgin Orbit and its partners will take another shot at re-establishing their flight heritage. But they should not stop experimenting. The alternative, as Libin says, would be “to be paralysed by not knowing what would go wrong”.

andrew.hill@ft.com

Twitter: @andrewtghill