Facebook’s motto for many years was “move fast and break things.” Regardless of one’s opinion on Facebook, it’s hard to argue that this wasn’t the strategically correct business decision: it enabled a period of rapid growth where they outcompeted almost every other social network and established one of the most valuable companies in the world.
This phrase and approach were repeated and copied, often blindly by many other startups. However, this isn’t a universally helpful approach. What is missed though is the context-dependence of this phrase and how specifically optimized it was for Facebook’s business model and market positioning (at the time). “Move fast and break things” is a great rule of thumb in a business where mistakes are relatively cheap and easy to fix and the rewards to being first are very high due to network effects.
If you were running, say, a nuclear power plant, a data backup service or a bank, “move fast and break things” would be a perfectly disastrous motto. These are businesses where mistakes are extremely expensive and hard (or sometimes impossible) to fix.
When you’re making a decision then, it’s important to understand the context you are operating in because the best approach is always dependent on the context.
One of the most useful tools I’ve found for understanding the context of a situation is the Cynefin framework (pronounced Kih-neh-vihn). It was developed by Dave Snowden after studying the management structure at IBM and categorizes work into four separate contexts:
Image Source: Wikipedia
The U.S. Defense Advanced Research Projects Agency has applied the framework to counterterrorism, and it is currently a key component of Singapore’s Risk Assessment and Horizon Scanning program. I used it in my book, The End of Jobs to talk about how work is changing and I’ve generally found it to be a helpful way of thinking about how to respond to problems
Simple: Building an Ikea Table
The “Simple” domain is one where the relationships between cause and effect are obvious. Anyone can apply a best practice to solve a simple problem. It’s something like putting together an Ikea table or a set of legos. In a business, something like order processing and fulfillment is generally simple. You put these items in this size box and send it to this address. There is a clear right and wrong way to do the task.
Simple problems can usually be solved either via automation (hardware or software) or very clear standard operating procedures. It’s something that can be documented easily and possibly done by a machine.
Complicated: Fixing a Car
“Complicated” occurs when the relationship between cause and effect requires analysis and investigation but is ultimately discoverable. Operating in the complicated domain requires investigation and the application of expert knowledge, but getting it done can be handled by utilizing existing expertise.
Repairing your car would be something that is complicated. If your car breaks down, it is not obvious to a non-expert how to fix it. However, a person with expertise (a mechanic in this case), can probably give you a few different “right answers” for how to repair it.
Maybe there’s a short-term option if you are in a hurry or a long-term option if you are ok with it being in the shop for a while. Maybe there’s a cheaper solution that won’t last as long and a more expensive solution that will last longer.
In a complicated environment, there is not one right answer, but there are a known range of right answers. The whole is the sum of its parts. It is the domain of known unknowns where trained experts and good practices flourish.
Complex: Caring for an Ecosystem
A complex environment is one where cause and effect are impossible to disentangle. It’s based on emergent practices. It is one where there are no right answers.
It is the domain of unknown unknowns, where more entrepreneurial skills are required.
Complicated systems don’t tend to have the same level of interlinking or dependencies where one part affects another. If your radiator is leaking, it doesn’t make your tires more or less likely to go flat. Cars are complicated machines, but an expert mechanic can take one apart and reassemble it without changing a thing. The car is the sum of its parts.
A horse, or most any other biological system, is complex. If you try to take a part a horse and put it back together, you are going to have a bad time. The horse, in particular, is going to have a terrible, horrible, no good, very bad day.
Like a horse, a rainforest is in constant flux—a species becomes extinct, weather patterns change, an agricultural project reroutes a water source—and the whole is far more than the sum of its parts.
When you try to mess with a complex system, it can trigger a huge cascade of downstream consequences. As part of an antimalarial campaign in the northern states of the island of Borneo in the late 1950’s, the World Health Organization sprayed DDT, an insecticide, to kill the mosquito population.
At first, it seemed to work great. Incidences of malaria dropped which seemed to be saving lives. However, there was an unexpected, emergent outcome: house cats were poisoned by the DDT from licking the insecticide off their fur.
As the cats died off, this created another emergent phenomenon: a rise in the rat population. The increasing rat population led to an increase in rodent-related diseases such as typhus and the plague.
Whoops? The WHO ended up stopping the DDT spraying and airdropping cats back into Borneo to try and reverse the damage they’d done.
An ecosystem like Borneo is in the “complex” domain is where the relationships between cause and effect are only clear in retrospect. The problem is solved by testing new solutions and seeing the reaction. It’s an emergent practice.
Chaotic: Putting out a Forest Fire
Chaotic is the realm of, well, chaos. If you find yourself in a chaotic situation, the goal is to act quickly and try to reduce the situation to merely complex where it can be dealt with.
First responders to a fire or terrorist attack are in the realm of chaos. There is no time to think about second order effects: you just try to put the fire out and save as many people as possible. When the situation gets to some more stable level and is merely complex, you have time to start testing longer term solutions.
Cynefin for Business
Most business decisions and systems can be fit into this framework and categorizing them appropriately tell you the right response:
- Simple problems are solved by machines (hardware/software) or simple Standard Operating Procedures (SOPs).
- Complicated problems are solved by more robust SOPs and expertise
- Complex problems are problems a company or organization has never faced before and how those are tackled is ultimately a question of culture.
- Chaotic problems are likewise a question of culture and are different mainly in that they require extremely fast response (e.g. Mayor Rudy Giuliani responding to the September 11, 2001 terrorist attack on New York didn’t have time test solutions, something had to happen fast).
Red: The Domain | Green: The Appropriate Type of Response | Blue: The Appropriate Tool to Respond
Simple: Machines and SOPs
This is the realm that is most effectively done by machines or simple Standard Operating Procedures (SOPs). This could be software code in a digital business or physical machines on a manufacturing line.
Example: A prospect submits an inquiry form asking to be sent further information, there is no reason for a human to do that. Email marketing software or off-the-shelf tools like Zapier can automatically trigger a response to go out.
For tasks that cannot be done by a machine or where it is uneconomical to do so, a simple SOP works well.
Example: Order processing and fulfillment: You put these items in this size box and send it to this address according to the order information.
Complicated: Experts and SOPs
Some SOPs are simple whereas others can get more complex, something more akin to a decision tree that requires some expert insight as to which of a few possible actions is best.
Example: a customer happiness SOP might have canned responses for frequently asked questions (simple), but also a guideline like “Use your judgment on what is best for the company in the long-term for any decisions costing less than $100.” This is not something a machine can do. But, like a mechanic fixing a car, an experienced customer happiness person would be able to make intelligent decisions about this.
Complex: Company Culture and Principles
The Cynefin guidelines for complex problems is “probe, sense respond.” Said another way, dealing with complex problems is mostly about tinkering: trying little solutions and seeing what happens.
Every day, individuals in a company will encounter situations that no one in the company has encountered before and have to make decisions about them. They will have to probe and tinker about the best way to respond. The question of how they should think about this and where to probe is a question of company culture.
Example: Facebook’s “move fast and break things” would be an example of a cultural belief or principle. When an engineer at Facebook ran into a problem they’d never seen before then this principle would tell them to just try and figure it out on the fly and ship a solution.
Someone working at a data backup company might have the exact opposite cultural principle: “move slow and triple check everything.” when they encountered a problem that they had never seen before then they might move very slowly, check with multiple colleagues, test it in a sandbox environment, etc.
First Context, Then Response
Responding to problems in a business appropriately requires appropriately fitting them into these categories. There’s no reason to take a simple problem and make it complicated where it requires an expert to do it: just right the code or SOP.
If customer orders are going out to the wrong address, you don’t need to re-evaluate your company culture or encourage entrepreneurial problem-solving. You need to fix the bug in your fulfillment software or standard operating procedure and make sure people are following it. It’s a simple problem that needs a simple solution.
On the flip side, you can’t write a SOP for every possible problem a business is going to face. At the level of complex problems, you have to rely on creating a company culture where people will solve problems in a way that leads towards the company’s long-term goals.
If the error in the fulfillment SOP is one of many errors due to people not keeping documentation up to date and and follow it, then you need to address the proximate cause (the error) at the simple level by fixing the bug. But you also need to address the root cause (the lack of documentation and resistance to creating it) at the complex level by creating a culture of effective documentation.
Any business can put its various parts into this framework. Some activities in the business are simple, easily expressed in “if this, then that” statements. Others are complicated: requiring expert knowledge but ultimately solvable. Still, others are complex, dynamic, and ever-changing.
The lines between these are blurry and most roles are going to deal with various pieces of them.
For example, the process of getting sales and expense information into an accounting software is probably fairly simple: you can connect your accounting software via APIs to your payment processor and computer code (AKA machines) will pass the data along.
The question of how to organize that information gets more complicated. You probably don’t just want a huge list of transactions, you want to group them in some way on your profit and loss statement. Money spent on books, courses and conferences may all get put under a category for “training and education.”
However, it could be equally correct to put a conference expense under “marketing” – maybe the conference was more about networking and finding potential partners or customers than it was learning from the content.
There are a number of “right” ways to do this but some are clearly wrong. You wouldn’t label a conference as a legal expense. This is the realm where a more complicated SOP with some decision trees along with expert guidance can be effective.
At another level, how to structure a Profit and Loss statement gets to be complex. There are a number of viable ways to do it, but what way of doing it allows the business to best understand the overall health of the business? As the business model changes, what may have worked in the past may become counterproductive.
One famous example of re-imagining a Profit and Loss Statement is the creation of Amazon Web Services (AWS). AWS was produced by Amazon taking the internal infrastructure it built for its eCommerce business and offering it externally as a service.
Amazon effectively turned something from a cost center (the cost of supporting their eCommerce sales) into what is probably their most profitable business unit. In a sense, they just moved it around on their Profit and Loss Statement. If Amazon had continued to think about it as a cost center rather than a profit center, they would have approached re-investing in it very differently.
Next time you’re thinking about a problem, ask yourself: what context does it fall in and what is the appropriate response?