The Strategy Kernel Canvas for agile teams
Have you ever felt like your agile team is getting a lot done, but not moving anywhere? Do you, your management, or your clients think the team is going in the wrong direction? When you hear how a decision is made it isn’t aligned with how you would have made the same decision?
These feelings point to the lack of alignment between stakeholders and the agile team. The key to this is a good strategy, well communicated.
At Philosophie, our product teams excel at quickly building the right things for our end users, but we often struggled executing towards strategic alignment.
This is why we created a method to quickly discuss and align on strategy. This method can be used as a jumping off point for further discussion throughout a product or organization. We call this tool the “Strategy Kernel Canvas” and think it can help you.
What is (good) strategy and why do we need it?
Strategy means a lot of things to a lot of people. I try to not be too pedantic about it, but there are a lot of wrong definitions.
John Boyd, was a great military strategist and great granddaddy of agile with his invention of the OODA loop. His presentation of “The Strategic Game Of ? And ?,” has the most succinct answer:
Survive, survive on own terms, or improve our capacity for independent action.
How do we keep surviving? This is the most difficult question for startups or groups that are battling inside a large organization.
The next issue is whether your strategy is a good one. For that, I can’t say enough good things about the Richard Rumelt book Good Strategy, Bad Strategy. It is an excellent primer on strategy itself and how to avoid the pitfalls that many (many) people have when creating their strategy.
If you have an hour (or 30 minutes at 2x), he has a presentation about good/bad strategy and the kernel here:
At a very high level, as summarized in this McKinsey article by Rumelt, the components that are necessary for a good strategy includes the following:
- A diagnosis: an explanation of the nature of the challenge. A good diagnosis simplifies the often overwhelming complexity of reality by identifying certain aspects of the situation as being the critical ones.
- A guiding policy: an overall approach chosen to cope with or overcome the obstacles identified in the diagnosis.
- Coherent actions: steps that are coordinated with one another to support the accomplishment of the guiding policy.
Why does agile need strategy?
Strategy is the north star for the organization. Agile is focused on figuring things out iteratively, with small and highly collaborative teams that work with the customer directly. Agile gives Product Owners the responsibility of setting priority, but has little to say on how they arrive or justify that priority.
With the agile mindset we assume that don’t know everything about the complexity when iterating. We do this while constantly learning from the customer (or people we are solving the problem for).
How do we use the Strategy Kernel Canvas?
The Strategy Kernel by Rumelt does a great job of guiding you to build good strategies. However, it isn’t very prescriptive on the actual process of creation, how to write it down, share it and get buyoff for it. His great work on what is good or bad strategy is the basis for the Strategy Kernel Canvas, which we will now introduce.
The purpose of the Strategy Kernel Canvas is to provide a framework for the discussion around strategy.
Just like the Business Model Canvas helps identify gaps in a business’ understanding of itself, the canvas is meant to identify divergence or broken links from the very high level to the day-to-day.
Here it is:
PDF available here.
Before, now, and future
You will notice that there are really three major horizontal sections: before, now and future. The main strategy components are within the now. This is because the strategy that you have is really about how you link the high level to now.
Top (high level) to bottom (day-to-day)
It is meant to be read from the higher level view at the top to the nitty gritty at the bottom. As you read from the top to the bottom it should make sense and feel linked. If you feel there is a break in the chain then you missed a step in translating from high to low level (or visa versa).
Complex systems, like our teams, are sensitive to initial conditions. This makes the history of the team very important. Why are we here? What did we try already?
In the case of Uber, they started in 2009 as an iPhone app in San Francisco for hailing black car limos. This was to address issues that people had getting a taxi in parts of SF. It could be horrible. I know, I lived there for 13 years. The history for Uber would include this story and other milestones or business changes over the years (e.g. funding events, country launch events, major protests). It would end with a summary of their market position as it is today (e.g. live in 58 countries, valued at over $41B).
It is the big challenge that you are addressing with your organization and doesn’t assume any particular solution yet. The most common concern for the diagnosis is how wide in scope should it be?
Looking at Uber again, what is their diagnosis? It has definitely changed from the main problem in 2009 and grown to something much larger. Today passengers are frustrated/dissatisfied with the lack of clarity, efficiency, options, and accountability in local transportation services.
Your organization’s secret. There are potentially other organizations that will be addressing the challenge that you wrote in the Diagnosis. How are you special and have an advantage over those other organizations? The guiding policy is found by an abductive process and potentially adjusted as more is found through agile iteration.
How does Uber address such a large challenge? They do so by creating a more transparent and liquid marketplace where drivers and passengers can find each other real time and reduce overall costs.
The important link from the Guiding Policy above to the day-to-day Actions/Process below. What are the main hypothesis your organization has that you are actively working on. In the parlance of Cynefin and SenseMaker these are safe-to-fail probes that are you currently working on. You should be rotating in Bets as you validate/invalidate the current set of Bets.
Uber’s Bets could include that drivers don’t need to be employees, no pre-scheduled trips (now allowed), network effects matter, people think about ordering food from the same app they get rides (so far unsuccessful), etc. Uber is a really large company so there are some Bets they are constantly tweaking and validating and others that they are rotating through as they fail.
What you assume isn’t right about the world or that doesn’t really matter to your success, but others think is?
Uber’s big Bet Against is that they can ignore local transportation regulations.
Actions / Processes
What the team are doing day-to-day to make the strategy happen. I include processes in this space since they have an impact on how a project moves forward and can change how the people you are building for experience your product. Weekly sprints with CI/CD would be something I would put in this section.
Uber’s Actions and Processes can be seen in what they implement in public so far. This has included:
- Apps for all platforms with GPS to book rides
- Buzz marketing to capture the attention of passengers
- High bounties on new drivers that are disgruntled taxi/black car drivers
- Cashless payment — also tipping and other fees should be hidden from the customer
- A marketplace needs to be managed for liquidity like a marketplace — thus surge pricing
- Rating of drivers by passengers (and visa versa) will keep the best drivers in the market and kick the bad ones out
- Pay for new users to the service, not per ride as exposure is most important in new markets
- Run each city/country as a separate office to adapt more quickly to the local conditions
The list goes on from here and is changing constantly as they try new Bets.
Success / Survival
The upper and lower bounds of how the organization can continue to exist. How will we know that this strategy is being executed properly? How do we know it is a good strategy? What is the bare minimum for our organization to continue to exist?
For Uber to be successful they need growth to keep marketshare and to have a functioning marketplace. Fluid marketplaces need the number of suppliers and number of consumers to be equal as much as possible with the ability to easily transact. As you get further and further away from equality the marketplace suffers. They are also looking to replace any other transportation services for the rider.
For a minimum bar they need to see that a marketplace can be functioning against competitors and in the regulatory environment. This is probably why they made the decision to sell their China branch to DiDi Chuxing.
Questions from workshopping the canvas
How do I decide what the scope of the Diagnosis is?
I have found that it really depends on the stage of the organization…
For startups, it is a very narrow challenge that they are trying to solve. Or at least it should be.
For a group that is within a larger organization, we should assume there is a certain overall organizational challenge. It isn’t always helpful for the group to rehash the particular transition from high to low level from that point. The Diagnosis should be why the group taking the strategy exercise exists within the larger organization.
For consulting groups, it is really all about the strategy of the engagement. Once you understand what the main challenge is you are doing everything you can in the engagement to address it.
What if I start with a lot of current Actions? Or just a Bet?
When dealing with something lower level that you want to understand at a higher level you can use Why Laddering.
You have gone too far in ‘why’ laddering if you hit the equivalent of reality bedrock. As in, things are the way they are because the laws of physics apply. Interestingly this is what Elon Musk uses this as his basis for understanding what should and shouldn’t be possible with the first principals method.
CFAR has started work on a method called Double Crux which is a more structured way to do Why Laddering. I’m very interested to see how this can help in these cases.
What if I have problem breaking the Guiding Policy into Bets? Or Bets into Actions?
Do the opposite of Why Laddering, which is How Laddering. This is very similar to “How might we…” questioning.
You have gone too far in the How Laddering for the Strategy Kernel Canvas if you are starting to plan individual tasks for someone. This should be left to whatever agile planning process you have.
How often does the Strategy Kernel Canvas need to be updated?
Whenever one of the sections changes. Especially the Bets.
Divergence or a break in the chain from the Diagnosis to Actions is the most important thing to look for when performing a regular audit of the strategy.
Should I do this instead of other canvases?
No, the Strategy Kernel Canvas doesn’t replace other canvases and the benefits they individually bring in workshopping, visualization, and documentation.
I find that doing a Business Model Canvas (or a Lean Canvas if you prefer) and Rich Picture help set the stage for a great Strategy Kernel Canvas.
You should display the Strategy Kernel Canvas for everyone to see where you would post other canvases.
Align your agile teams with the strategy before you have a problem
Strategy that isn’t communicated effectively in an organization is one that exists independently in each person’s mind. Until you have alignment across your organization you will continue to have agile teams that go off the rails.
If you would like to see the first talk I gave about the Strategy Kernel Canvas at a Product School meetup, you can watch it here:
The slides and links referenced are also available online.
If you have any questions, are interested in a Strategy Kernel Canvas office hours, or want to hear more about what Philosophie does please email me at [email protected]