Agile and DevOps have taken the tech world by storm, or have they? Less than 20% of organizations have adopted Agile practices up, down, and across their businesses. Organizations that go deep and wide see 60% higher revenue and profit growth. That’s a massive incentive. Yet, with all the thought, time, and money put into scaling Agile, why aren’t more organizations more broadly Agile? Why haven’t the armies of consultants pushing SAFe, LeSS, Nexus, Scrum@Scale, and The Spotify Model made a more significant dent? What gives?
We at Merely are passionate about Agility. We’ve devoted our time and treasure to creating software that empowers an Agile culture. We are testing hypotheses to the questions above. One observation is the presence of Agile development teams driven by rigid and opaque strategies from leadership.
Today you’re going to get a very transparent look into how we apply Agile principles and practices to craft and execute strategies, one facet of a wholly Agile Organization.
Embrace Uncertainty
The first and most crucial step is to drop the pretenses and posturing that manufacture false certainty. Each product in our portfolio wishlist has the potential to address and service hundreds of thousands of people. No set of requirements can reasonably serve that many potential customers. There are too many variables. Moreover, a fair number of those products in our wishlist don’t yet have one-to-one analogs in the marketplace. To say anything like, “We will build these features with those specific requirements and execute that group of marketing campaigns, then we will have 1,000 new customers.” is nothing more or less than speculation, especially in this early stage of our product’s lifecycle.
Predictions are essential for learning, but pretending they are certainties limits our ability to extract value from them. Our response is to see ourselves as scientists. We are pursuing truths through the scientific method. Let’s refresh ourselves on the rudimentary form of this method.
There’s no better way to turn simple observations into powerful truths, especially when vital information is missing or contentious. Now that we have a proven method for dealing with uncertainty, we need to tune it to our environment and purpose. We want to quickly iterate through our observations, analysis, predictions, and experiments to take advantage of the low cost to change our products and business.
That’s where Agile comes in.
Amidst genuine uncertainty, we can act safely by creating and iterating through thoughtful experiments. For us, the best way to get there is by approaching questions, hypotheses, predictions, tests, and adaptations as a team.
Collaborate
In our experience, senior leadership often handed down strategy through the hierarchy. While created by bright, thoughtful, and passionate leaders, these strategies had a quality problem. We usually found core elements impractical, lacking substance, and tightly coupled with teams who weren’t given a chance by leadership to provide meaningful input. We recently posted an infographic that tells a story rooted in lousy strategy.
Competing objectives stem from a lack of clarity. And while there are many possible antagonists to breed firefighters, one is strategic misalignment, which is one of several predictions we are testing.
In a survey of thousands, a whole quarter says their organization’s strategy is entirely unclear. Entirely. Never mind the degrees of clarity or if the strategy is clear but unsound.
So then, where can we start? What can we do to get at arguably the biggest problem, the disconnect between strategy and execution? We need to do so much to create a genuinely sound strategy, but let’s leave the details to the good Doctor Rumelt.
Our first step will be cooperatively creating the strategy with those accountable for the execution. This kind of collaboration must be more than a facade. It should never be a token gesture or fluffy.
The benefits are economic and cultural. You get fast feedback, diverse thinking, and shared ownership. Even if the resulting strategy is mediocre, we will feel the gains through aligned and coherent execution.
For a software team to be Agile, it is necessarily cross-functional. The team must have the skills or ability to acquire the skills needed to deliver working increments. When the Agile team plans the work for each sprint, it is a whole team effort.
We take the same approach with strategy.
Respond to Challenges
Goals, missions, visions, and slogans are not material actions addressing specific challenges. The most important thing a strategy should do is define how we respond to the challenges we face in the market.
To do that, we will put the right group of people together (senior leaders or not) and fire up the scientific method, starting with collecting diverse observations of the challenges we face in the market. Second, we all ask questions about the nature of these challenges. We look for shared observations to aggregate, refine, and abstract. Third, we draft and refine testable explanations for success against the challenges. Fourth we make predictions as policies and actions that we can test.
The policies and actions ought to be constraining. Our purpose is to create focused execution and meaningful discrete tests. Here is how we evolve observations into testable action.
Challenges / Won’t Do / Will Do
For this exercise, we use a Liberating Structure like 1-2-4-All for identifying a clear and concise set of challenges. Then, repeat for finding actions that would work against our purpose as, “Won’t Do.” Finally, we go through the same process for identifying and clarifying our “Will Do” actions.
Once the group is satisfied, we begin mapping the responses to the challenges. There are no hard and fast rules here, but we aim to create virtuous cycles where each action and forbiddance work together to address challenges. The more direct connections they have, the better.
This activity produces a high-level description of what we believe to be our most significant challenges and productive responses today.
But we aren’t done. These short statements lack connection to execution, and that’s a strategic anti-pattern. For example, “Opinionated Software” may work against the needs of those we hope to serve—the specific nature of the opinion matters to our like-minded Agile enthusiasts.
Policies & Actions
Armed with a general characterization of challenges and responses, we can start putting some meat on the bone. We find it helpful to create policies alongside actions that provide examples of realizing them.
We mustn’t see the initial actions as immutable. They serve as a starting point and help us see how we might iterate on the practical realization of policy.
To relate this to the scientific method, one of the working observations is that so-called Millennials (of which Merely is entirely composed) are now the majority of buyers. And the overwhelming majority of Millennial buyers are turned-off by requirements to speak to sales.
Our policies of “Whole Focused Products,” “Natural Scaling,” and “Low-Touch Sales” are our strategic hypotheses to counter our competitor’s complex gatekeeping. The related actions are specific predictions that we are testing.
We then sum up what we believe are the three most impactful strategies. Today, those are:
- Focus on the Agile Niche
- Dead Simple Pricing
- Self-Service by Design
We will need to address many more challenges with specific responses to succeed in our vision, but we need to start with focused hypotheses and experiments within a practical timeframe.
Test & Iterate
Preferring a plan over adaptation is fragile, and there’s nothing that prevents this truth from destroying any benefits we derive from strategic planning. We will learn from successful Agile product managers and apply the same empirical process to strategy. Product Managers and Marketers both understand this process very well within their domains. A/B testing, product analytics, cohort segmentation, and campaign analytics are robust and mature. A product manager given great user acquisition (UA) from marketing who’s figured out retention but lacks conversion will start experimenting with ways to increase conversion. And a thoughtful product manager with a responsive team will find gains given enough time.
Now that we have actions to meet our challenges, we test our path. We ask ourselves, “What outcomes would validate or invalidate our actions?” These outcomes are often lagging indicators. We call them “Targets.” Even if it is a lagging measurement, a target should still have a meaningful result within three months. We aim for one to five targets, preferring fewer.
For example, this quarter, our targets are: Increase early access users' willingness to buy, decrease the time between deliveries, and increase social media followers. We express the target as a relative change. Like so: “Decrease time between deliveries 50%.” Next, we theorize about which leading indicators will predict the target outcome. Metrics like social media or feature engagement rates make for good predictors. These metrics should change quickly after deploying a feature, campaign, or process. Naturally, we call them “Predictors.”
Finally, we ask ourselves, “What coarse-grained areas can we conduct experiments in to try to move those predictors?” These areas might be features, marketing campaigns, business model changes, UX improvements, or processes.
The result is something like this:
Most strategic efforts end at this stage until the next annual strategy session. But that’s not our approach. Each sprint, we look at what we’ve done and what effect it produced in our outcome-oriented roadmap. Each sprint planning, we create new work constrained by our strategies in response to the outcomes. As soon as meaningful data exists, we can adapt our experiments, predictors, targets, or the strategy itself.
An output-oriented roadmap that plots features delivered over time would be a complete waste with this approach. And that’s a good thing. Because maybe, they’ve always been a waste of time and talent.
