Adam Pisoni recently joined Responsive.org, a new movement dedicated to helping companies become more agile, adaptive and empowering. Previously, he co-founded and served as CTO of Yammer.
Facing a rapidly changing world, companies are waking up to the fact that they need to become more Responsive Organizations. Yet even when the need for change is apparent to all, shaking things up is extremely hard, even if you're in charge. That was the lesson I learned working with leaders across many large organizations. It doesn't matter if you're trying to change previously standardized technologies or practices, standardize something previously decentralized, or shift company culture to be more innovative and customer focused. Changing deeply engrained technologies, practices or customs is never as simple as sending out a memo or holding an all hands meeting.
All organizations (even startups) end up becoming decentralized. Few are ruled by a benevolent and beloved dictator who can make unilateral decisions that will be followed perfectly. While decentralization is what gives organizations the autonomy, accountability and parallelization required to scale, this model is stubbornly resistant (and sometimes impossible) to change. Especially widespread, holistic change. Below, I explore what it takes to gain buy-in for this type of revolutionary change, and the tactics you can use to build what I call a 'new template' for your organization as it grows.
When you try to scale change, you end up fighting two opposing forces: standardization and local optimization. Standards can be both explicit, like performance management systems or central HR systems. Or they can be implicit, passed down from prior torch bearers through phrases like, “This is the way we've always done it.” Local optimization, on the other hand, can be a sign of either a healthy empowered workforce or a dysfunctional feudal system. Neither standardization nor local optimization is universally good or bad. Both provide benefits and yield failures depending on the situation.
In either case, there is a way things have always been done, and regardless of how well it worked, the process, pitfalls and benefits are understood. Change introduces risk. Sure, it might make things better, but it could also make things much much worse. Convincing people that a change is necessary and beneficial requires more than simply laying out the facts. It requires altering mindsets. Remember, there's a reason everyone is doing whatever they're doing the way they're doing it. At one time, it was believed to be the best way to do it. To drive change, you first have to convince them the new way is betterfor them.
In their book Scaling Up Excellence, Stanford Business School Professors Bob Sutton and Huggy Rao tell a story about how leaders at Xerox decades ago attempted to create a new sales playbook by combining a disconnected set of best practices they observed at various companies. This seemed like a good idea in theory. But when they went to roll out their untested collage of practices, no one would implement them.
Why? There were no real-world examples of where that whole model had been implemented and was working in practice. Without a working model to replicate, no one would trust the new model on faith alone. Sutton and Rao go on to tell a different story about howKaiser Permanente successfully rolled out a large technology change in a different way. Instead of trying to change things for all groups at once, they rolled out the new product piecemeal, in small, safer markets to prove the change was beneficial, before expanding it to even more people.
The key lesson here: People need a new template to copy.
Changes require believable models that can be witnessed in practice so that people have a reference starting point.
Even if the desired outcome is new local optimization, a clear starting line is crucial. I like Rao and Sutton's use of the word 'template' because it implies something well-defined but still containing blanks that can be filled in by those building new models based on it. Models are templates put in practice.
The people who recognize that change is necessary are easily frustrated by leaders who don't take the benevolent dictator approach and force new templates. Good leaders, on the other hand, know that there are many dangers in forcing change on a decentralized, empowered organization. It sends a signal that the top tier knows best and doesn't value the judgement of local leaders — a perilous message for the health of your company. Worse, if the local leaders aren't motivated or thoughtfully introduced to the change, they will probably fail at implementing it. These failures are too often blamed on the ideas themselves rather than lack of commitment.
So if we can't dictate that people change, and people won't change without a relatable example, how do we roll out a new template? Sutton and Rao describe a spectrum with centralized, dictatorial leadership on one end and decentralized leadership on the other (a contrast they call Catholic vs. Buddhist leadership). In their research, they found that neither approach works all the time, but instead must be balanced depending on the situation.
When we combine these philosophies, I believe we arrive at a brand new model. With it, we're able to create large-scale change in even the most decentralized organizations. What we have to start with is a new template that balances the top-down Catholic approach with the grassroots Buddhist approach to keep our companies flexible, responsive, and continually improving.
Step 1: Ditch theory. Just get started.
This may seem obvious, but I've seen many organizations big and small get stuck before they start. It's not enough to talk about what's wrong with your old systems and paradigms, or opine on some imaginary template based on new values yet to be defined. AsKhan Academy Founder and CEO Sal Khan says:
If complaining about the status quo is easy, theorizing about how things ought to be is not much harder.
At some point, someone has to figure out where and how to get started. One needn't figure everything out ahead of time, but at least enough to have a credible way of getting a small team of people to start providing feedback on the new template. Write it down. Name it. Explain that it's not final. It's just an experiment that needs use and iteration.
Step 2: Pick a partner. Start to dance.
If you're lucky, you already have the makings of a new template in the form of a team with the courage to deviate from the old one. Most likely, you found them because they're outperforming their peers. If you have or can find one of these teams, you're in luck and can skip to step 3.
If not, you may have to employ some gentle Catholic love and pick a good test group yourself. Ideally, you find willing volunteers, though it would behoove you to make sure they're not just willing but also capable. Worst case scenario, you have no volunteers and will have no choice but to “convince” a team to work with you in co-creating the new template. There are a few things to remember when doing this:
It's imperative to pick the right team.Try to choose a team that is amenable to change. This often means a team where change feels safer. Picking the largest, most successful team at your company, for example, will often be the most dangerous and risky to change. In the above Kaiser Permanente example, they rolled out their new solution to Hawaii first, precisely because it was the smallest market. I wasn't there, but I also bet Hawaii was excited to get so much attention given its size.
Be Honest and Collaborative. You're trying something new. Failure IS an option. It's important to be up-front and as honest as possible about what you know and what you don't. The people you work with on this new template aren't just guinea pigs — they're partners. Even though you may be using your authority to enlist their help, you should approach them as equal participants. You need them as much as they need you. Their role isn't simply to execute on your plan, but to help co-create the new template using the real-world feedback resulting from their own experiments.
Step 3: Document the transition
Remember, you're not just rolling out a new template, you're building a case for everyone else to use it. Document your learnings rigorously. Celebrate and measure successesand failures. Build an argument for why this is a better way. Most importantly, document how your partner group successfully made the transition from the old model to the new one based on its specific strengths.
Other groups will likely require clearer, more granular details for how they should get started than the first group did. What starting point worked? What made the transition difficult? What roles helped? What resources were needed? Finally, given that failure is still an option, be open and honest about whether the new template really isn't working when it's not. It's better to retreat sooner than later when the cost of failure is always increasing.
Step 4: Start scaling
Once you've proved that your new template produces better results, it's tempting to go big. But scaling to everyone at once may fail just as readily as if it didn't work at all. Eventually, you may have enough critical mass that it will feel natural to accelerate and potentially reach everyone, but at first you just need more partners.
Even if you believe you have a new viable template, it still won't work to impose it on anyone without their express cooperation. One thing that helps is to identify parts of the template that are explicitly open for local optimization vs. standardization. Leaving room for groups to customize the template to their needs and wants will give your next partners the ownership required for them to accept the new template as their own. Keep in mind, allowing for local optimization is usually not enough in the beginning to get new groups to buy in. This is where our secret weapon comes in: Rotation.
Step 5: Rotate
You've now got a team successfully, and hopefully happily, living in your new model. You've also found the next few teams you want to adopt the new template. The reality is they're unlikely to be as courageous, flexible and accepting as the original team was. After all, that's why you chose the original team in the first place. As outsiders, new people probably still won't get it.
New models require new mindsets that can't simply be taught, but must be experienced to be fully understood.
The safest way to replicate a new template is to take trusted people from the team you want to roll it out to and embed them for a tour of duty within the team already living in the new model. Sometimes it also helps to take people already living in the new model and embed them in the teams you want to influence.
As time passes and they begin to fully understand the advantages of the new model, they will become advocates, and you can rotate everyone back. If the new model truly is better and you've picked the right people, they will begin to spread the gospel of the new model without you coordinating cross-pollination. They will have the added benefit of understanding the model enough to provide feedback on where they see the template working well or not on their home team.
Here again we're balancing our top-down Catholic approach with a grassroots Buddhist strategy. We may have to force rotation as a top-down, temporary, safe exercise, but if we do it right, the change will start organic expansion as people go back to their home teams.
As you can imagine, you may end up doing a fair amount of rotation within your org to reach critical mass, depending on your size. This type of rotation is usually uncommon and uncomfortable for large orgs, which is why I call it a secret weapon. It will likely be uncomfortable for you as well at first until you're used to it, but it's worth it.
Discomfort is a whole lot more comfortable than failing.
Step 6: Keep Iterating
The reason you needed a new template at all is because something about how you worked before broke down. Maybe the world changed and you didn't. If you simply roll out a new template and assume you're done, you'll end up back where you started. The best models are the ones that learn and adapt over time — this evolutionary inclination should be one of the pillars of any model you mass deploy. Build evaluation, rotation, learning and iteration into your model so it can adapt before it becomes obsolete like the last one.
There are libraries of books about how to build organizations for maximum efficiency, creativity, control and empowerment, but only now are we seeing literature on how to go about changing our organizations when they fall behind.
We may have believed that we'd never need to change if we built things right from the start. Most organizations seem to only roll out foundational changes as a last resort, having already gone over the cliff. But change doesn't have to be carried all the way to death's door if we approach it iteratively. After all, the cornerstone of anyResponsive Organization is to actually build an organism that adapts continuously, conscious of the fact that the most difficult change isn't about process or technology but mindset.