Adobe handed employees a red box before a manager had approved the idea

A method that lives in one team dies as a personality. The people in that room can run it because the person who built it is standing there to correct them. Everyone else is still waiting for permission, or for a next step they can take without that person.

Adobe's interview with Mark Randall, published on 10 December 2014, is the company's account of making one method portable. Randall was vice president of creativity. Kickbox had been running since the year before. More than 1,000 people in different countries had a box, used on projects, apps and product changes that might otherwise not have been built. That week he was due to speak at the Lean Startup Conference in San Francisco.

The object is a red box. The interview lists a candy bar, a Starbucks gift card, a debit MasterCard loaded with $1,000, then maps and instructions. The company gives time and resources. Where the idea goes is the employee's call. People generate ideas, cut them to one and test that one with customers, asking "Would my proposed solution solve problem X if we were to build it?" The evidence, not the hunch, is what they bring back.

The money is not a stage gate. It starts, he says, with trust. In the seed phase people manage themselves. "They are the CEO of their idea." He calls the seed phase scalable, because a committee is not approving each small spend. Six colour-coded levels each have a checklist, and the list opens the next level. He calls that self-gating. Finish level 6 and you have beaten the red box, which earns further funding and a blue box. The blue box comes at the end, not at the start.

He tells the origin as a teaching problem. Teams around his own products were using lean and agile methods, and others asked him to show the rest of the company. He did not know how to train that. Colleagues told him he could colour outside the lines. The frame that worked was to treat would-be innovators inside Adobe as his customers, and the box as a product for them. The box teaches that reframing, and he says the programme was built with its own method. The challenge he names is "Find a way to unleash more innovation." His answer was more innovators, including people outside product and design. The first project may fail. "Whether their first project succeeds is less important than the experience and skills gained." Those skills are a concise idea, an internal look, a customer test and the use of that evidence to ask for support. He wanted this in an organisation of 11,000, on an ordinary Tuesday. The shift he describes is that people stopped treating the job as the assigned task alone, and could connect what they wanted to make with the customer's problem.

1. Put a small sum in the person's hand before the pitch. The card holds $1,000 and is described as supplies for starting, not as a claim after approval. Money that moves only once a manager has said yes puts the gate back.

2. Let the employee finish a level without booking a meeting. Six checklists, and the next level opens when the list is done. A stage that waits on a committee selects for people who are good at waiting.

3. Score the skill, not only the survivor. He says the first project matters less than the experience of testing with a customer and using what came back. A showcase of winners misses the reason he gives for the box.

4. Hand it to people whose title is not innovation. The aim is more innovators, including people outside product and design. Recruiting only the teams who already ship repeats a skill the company has.

A lean startup session is a place to run one seed the way the box does, with the person who holds the idea and the person who can fund the next step looking at the same customer evidence before the money gets larger.