团队重组不可避免：Heidi Helfand 谈五种变化方式，以及怎样把它做好

Helfand spent 20 years in fast-growing software-as-a-service startups, including joining one as its 10th employee and watching it grow past 900 people. The books she read told her to keep teams stable and move through forming, storming, norming and performing. That did not match her experience, so she set out to document how teams really change and whether others saw the same thing.

She acknowledges that much management advice focuses on delivering on time and building things customers love. But there is also a lot of company-building work in the people structures as an organization goes from one team to many. In her view it would be convenient if everything were just product-market fit and customer conversations, but we are humans together, and that work does not get enough attention.

Helfand's favorite approach is transparency. She cites Spotify's Christian Lima, who visualized a large infrastructure team's reorganization on whiteboards and invited people to give input. Inspired by that, Procore did the same with about 80 people in a platform organization that was splitting from two clusters of teams into three. Rolling out the boards felt scary, she says, because it touches people's day-to-day work.

The boards showed each team's name and mission, open hiring slots, and people's names in their current positions. Employees found design mistakes, argued that a team belonged elsewhere, and saw roles they wanted to be considered for. She says this beats a reorg decided in a back room, and that having a starting point makes meetings more productive. She calls the approach whiteboard reteaming.

Even more open is the self-selection model used at Redgate Software, as described by her colleague Chris Smith, where teams pitch what they need and people volunteer. Helfand notes that people can be delighted by work others find dull, so choice matters. Her takeaway is that there are different grades of transparency, and Procore chose the whiteboard variant.

Lenny raises the executive worry that involving everyone causes distraction and endless disagreement. Helfand answers with William Bridges' Transitions, which describes an ending, a neutral zone and a new beginning. She says learning it made later changes easier, and that leaders can help by painting the benefits and purpose of the new reality.

She also stresses being clear about who decides. She learned the RIDE framework from Procore's chief people officer, Pat Wadors: who requests the change, who gives input, who decides and who executes. Some changes, such as an acquisition, are not up for team input, while others, like splitting one team into two after a retrospective, can be teams' own choice. She adds that decision-making need not simply equal hierarchy.

She admits reteaming and reorgs are hard and not always great, but argues for focusing on the people layer because they are inevitable. On process, she says to time box the discussion and bias the schedule toward shorter, because including more people makes it more distracting and the aim is to proceed expediently.

Helfand lists the five patterns. In one by one, a person joins or leaves the company. In grow and split, a team gets bigger and divides in two or more. Merging combines teams and tends to be a shrinking pattern. Isolation, or innovation by isolation, creates a team off to the side as a beneficial silo with process freedom, and switching moves someone from one team to another at various cadences.

She distinguishes reteaming from a reorg. In her view "reorg" carries traditional baggage and implies large, top-down changes. When someone named Sue wants to learn how web operations work and moves to that team, calling it a reorg did not feel right. Reteaming covers these patterns at all levels of the organization.

Helfand tells the story of Expertcity, where she was the 15th employee. The first product was a marketplace for tech support, a bit like an eBay of services, built around screen-sharing technology. She was an interaction designer, then called a navigation designer. One day the CEO told her to stop working on the marketplace because it was being killed after bringing in about six dollars the month before, and to go to the beach so she would not start work that would need maintaining.

She cried and objected, but later was invited onto a small team off to the side. The founders and product people had become students of market validation, following Four Steps to the Epiphany, which was a do-or-die shift. That team did not have to follow waterfall, deployed more often and built GoToMyPC. She calls it the pivot that saved the company, and the team later folded back into the main organization to build GoToMeeting and GoToWebinar.

She argues new products need faster iteration loops than mature product lines, and that sometimes a beneficial silo is exactly right despite advice to desilo everything. Isolation also suits emergencies, such as a team that left their groups for a couple of weeks to fix performance problems. She adds that the pattern is old, citing the book Teamwork and the story of the Chicken McNugget, which was developed by a small team in a different plant reporting to a McDonald's executive.

Asked how startups-within-companies actually succeed, Helfand describes SecureDocs at AppFolio. A team was created off to the side, worked in daily sprints instead of the two-week cadence of others, and had its own marked area in a big open room that it claimed and named. SecureDocs later branched into its own entity and was recently acquired. She also mentions a Citrix team that incubated a product idea in a garage.

Her recipe is to separate the location, and to have leaders tell others not to disturb the team, which she says is key. Teams should have process freedom and report to someone with real decision-making authority so decisions are not reversed, without heavyweight bureaucracy such as elaborate quarterly business updates. To make it easier to move people into such a team, she suggests building redundancy in advance through pairing and shared ownership so nobody is chained to a system.

Lenny highlights the importance of an executive sponsor. Helfand agrees and warns that isolated teams can go wrong, such as when someone sells a feature, friends build it, and others are left to maintain it without having been involved. Her image is balconies and basements: every pattern can be done badly, and everything takes effort.

For one by one, she says help new hires feel they belong, for example with a first pair on day one and by getting them to talk about themselves, citing Daniel Coyle's The Culture Code. People already there also need attention and coaching, particularly if they were not involved in a decision such as an outside manager joining. For grow and split, signals include longer meetings, harder decisions at 13 people than at five, and diverging work in standups. Splitting can create new dependencies, for instance with only one product manager or designer.

Merging can happen when teams consolidate, when a company downsizes, or when one company acquires another, as AppFolio did to move faster into software for law firms. For merged teams she recommends an activity called story of our team, where each person stands in order of joining and the group builds a timeline of milestones and proud achievements. It creates a shared history and helps people look ahead to a shared vision.

Switching, she says, is tied to learning, development and fulfillment. Tuckman's model skips a stage she calls stagnating, when people have been together too long, and a move can feel like a new job within the same company. Switching also builds safety nets: with pairing and switching, as practiced at Menlo Innovations, knowledge is not locked in one person's head, which matters for critical systems like rent payments at AppFolio. She warns not to switch people around haphazardly without a say in their team.

She recalls AppFolio engineers who missed pairing after a split and started their own rotation between teams, which brought learning and joy. Linking it to autonomy, mastery and purpose from Dan Pink's Drive, she says the goal is not just products people love but companies people love.

Lenny quotes a manager who said there is no perfect structure, only the best idea you have at the time. Helfand agrees, calling much of this problem trading as conditions change, and compares it to the Toyota Kata of grasping the current condition and moving toward a target condition. She asks leaders what kind of leader they want to be, whether announcing a change by email or involving people, and notes that involving people everywhere is not always possible.

Her first anti-pattern is the percentage approach, assigning people 10 percent here and 20 percent there, which she says does not add up and causes context switching. Another is the poof-they-are-gone pattern, where people suddenly appear or vanish with no communication.

A third is spreading high performers across other teams. She says AppFolio's Jon Walker tried it with a team that had great energy and cadence, and it did not reproduce the effect. He felt he had destroyed something, and she adds an image from Damon Valenzona of a team as a band making music together.

Helfand acknowledges that sometimes you want to keep a great team together, but a company with a bigger vision has to grow beyond ten people. People ask how to preserve culture, and her answer is that it does change as the company and world change. She shows a photo of herself holding a melting ice cream cone as a reminder to appreciate the good moments, and recalls skipping through the halls during such a time.

She points to Ichak Adizes' Managing Corporate Lifecycles, which describes stages from birth through go-go and maturity to death, and says teams age and renew in a similar way. Lenny recalls Sheryl Sandberg telling Airbnb staff that constant change is the best-case scenario because the alternative of not growing is more painful. Helfand echoes her former CTO Jon Walker's line that it is always great to be at a successful company, while adding that not all change is good.

Lenny passes on John Cutler's question about why Helfand is such a good listener. She calls listening a muscle and describes the three levels from her co-active coach training. Level one is internal listening, thinking about lunch while someone talks. Level two focuses on the other person, and level three is global listening, noticing the environment and signals like a sudden touch to the neck or a reddening face.

In the lightning round she recommends Leading Intelligent Teams, Liberating Structures and Transitions, likes asking interviewees why they want to join, and offers the motto of asking how you can be kind to yourself. She works with companies through workshops, talks and retainer consulting.

Asked about the unofficial director of fun title, she describes AppFolio's hack days modeled on Atlassian's ShipIt Day. Teams formed through a self-selection marketplace and built things such as geocaches around Santa Barbara, a vintage arcade machine and a trebuchet, with goofy traveling trophies. She says working with different people builds relationships that make later reteaming easier because people are not strangers.