Making accurate decisions in the volatile world of SEO can be challenging, especially when facing complex and unfamiliar situations. In this scenario, the Cynefin framework can guide us in managing SEO projects in an increasingly VUCA (Volatile, Uncertain, Complex, and Ambiguous) or BANI (Brittle, Anxious, Non-linear, and Incomprehensible) environment. At the end of this article, I share the complete framework so you can learn to effectively handle different situations.
Origin of the Cynefin framework
Dave Snowden, a former IBM employee, is the original creator of the Cynefin framework (1999). Currently, Snowden is the Chief Scientific Officer of Cognitive Edge and founder and director of the Center for Applied Complexity at the University of Wales.
The name «Cynefin» (a Welsh word pronounced kuh-nev-in) means «habitat» and is a metaphor for the idea that each individual and system is the result of its evolutionary history. Snowden defines it as «a place of multiple belongings».
Since then, the Cynefin framework has gone through several iterations, as the naming of the different domains, which we will see shortly, has evolved over time. You can check the current state on Dave Snowden’s blog.
The 5 domains of the Cynefin framework
The Cynefin framework presents 5 situations in which an organization may find itself, which guides the way we should act. The framework compares five contexts with different degrees of complexity: clear, complicated, complex, chaotic, and confused.
To understand complexity, the structure and predictability of the elements in the environment are considered. The combination of these elements gives rise to five situations with specific characteristics that determine how to approach them, which mainly differ in the relationship between cause and effect.
When a structure can be easily understood, it’s considered simple; if not, it’s considered complicated (not complex). Regarding predictability, it can range from completely predictable behavior to completely unpredictable, with moderately predictable situations in between.
In clear situations, causal relationships between cause and effect are obvious, and the solution is easy to recognize and follow. The best way to tackle simple situations is by resorting to best practices. This is possible because the correct response is clear, known by everyone, and indisputable.
That is, if A happens → we do B.
Recommended action: Sense – Categorize – Respond.
Complicated situations also have an evident relationship between cause and effect, but only for the expert («unknown knowns»). Therefore, to make decisions, specialist knowledge is needed, as nonlinear processes and multiple variables are involved. The more complicated the system, the more important the exchange between project participants and the thorough analysis of the conditions.
Unlike the complex systems we’ll see now, in complicated systems there are no hidden variables, meaning the relationships between the parts don’t add additional information. It’s enough to know how each of them works to understand the system.
If A happens → we analyze → we do B, C, D, F… (the well-known «it depends»).
Recommended action: Sense – Analyze – Respond.
A complex environment is one of «unknown unknowns», when the number of variables is too high to reveal the relationship between cause and effect beforehand and arrive at a secure solution.
Complex systems are entities composed of multiple interconnected components that interact with each other, often involving hidden variables to the observer, generating emergent properties that are not evident when analyzing the components individually. These properties are the result of the interactions and cannot be attributed to any particular component, which makes complex systems difficult to predict and understand without a deep analysis of their interactions.
In these systems, it is recommended to apply experimentation through agile methodologies.
Recommended action: Probe – Sense – Respond.
In chaotic systems, no causal chain can be recognized between cause and effect. Old structures and processes have no effect, and new ones aren’t yet established. Rapid action (creating or removing constraints) and a leader are required to move the chaotic situation to another area. Then, in that new area, we can respond appropriately to the resulting clear, complicated, or complex situations.
Recommended action: Act – Sense – Respond.
The state of confusion, recently dubbed as Aporia, is when we don’t know which type of system we find ourselves in. Nowadays, it’s more accurately defined as a suspended transition zone between states (liminal). This domain poses a significant issue for organizations, as people are accustomed to a one-directional approach and are not familiar with an ontologically diverse perspective.
Aporia is an uncomfortable and ignorant place, but it can be a good starting point to go to when we enter chaos and from which to apply the «Aporetic Turn», where we’ll carry out different actions to transition to other domains. To reach Aporia from chaos, we must create constraints that introduce stability.
If the type of relationship between cause and effect is unclear, we must be careful not to retreat to our comfort zone and choose comfortable solutions. Addressing the bewildering and actively observing action patterns can help us explore new paths and learn from unclear future situations. For this, several useful exercises exist:
- Triopticon: is a workshop that facilitates collaboration between different disciplines and encourages the exchange of diverse opinions in an organized format. This method combines elements of formal and informal conferences, and allows groups of three people to share knowledge and experiences in different areas such as conflict resolution and innovation.
- Safe to fail probes: are small and controlled experiments to explore solutions and learn from outcomes. These tests enable the identification and expansion of successful ideas and learning from failures in a safe and tolerable manner.
- MassSense: is a tool for assessing situations in near-real-time, gathering perspectives and detecting weak and atypical signals in events or decisions. It allows the generation of hypothetical scenarios and the improvement of decision-making through self-interpretation and rapid feedback.
You may find that the Cynefin domains remind you of the Stacey Matrix, which is used to explain the application of agile methodologies. However, Cynefin’s success lies in its ability to elucidate and create a narrative about complexity that is more accurate and realistic than Stacey’s theory, as well as a more direct practical application.
Liminal zones and dynamics
The liminal zones (green) in the Cynefin model are transitional spaces or boundaries between the different domains of the framework. These areas are crucial because they provide the opportunity to explore and test solutions before moving to another domain where a more appropriate strategy should be applied. Additionally, liminal zones can function as moments of learning and adaptation, as situations may become clearer as they move between domains.
In the real world, agile tools and methods commonly used in complex systems (such as SCRUM or Kanban) don’t always fit neatly into a fixed domain. To address this, Dave Snowden created two liminal spaces for these methods:
- The liminal state between complex and complicated: it’s the midpoint at which we are trying to strike a balance between exploring new ideas and exploiting known solutions. To do this, we increase constraints in the process to achieve reliable repeatability and allow the system to scale. This is accomplished by merging parallel experiments, which leads to greater consistency and predictability in the outcomes. By iterating ideas in this state, we reduce the amount of experimentation in the complex domain, which helps us focus on the most promising solutions. By maintaining this balance, we minimize the risk of settling too early on a solution while also ensuring the time it takes to move from exploring new ideas to exploiting solutions is efficient.
- The liminal state between chaotic and complex: it’s none other than the edge of chaos (also called «the cliff»), where constraints and connections are reduced to achieve a certain degree of randomness. This state should not be confused with the complacency errors of the simple domain (which we will see later), as that can lead to chaotic collapse.This space allows ambiguity to generate fresh ideas and novelty. However, it’s difficult to achieve, so abstraction is used to avoid falling back into old and familiar patterns. One of the techniques recommended here is «Skunk Works,» which involves creating autonomous teams disconnected from the rest of the organization, allowing them to generate uncontaminated hypotheses for obtaining objective data.
And how can we move from one domain to another? For that, there are the dynamics of the Cynefin model.
Although the Cynefin domains are often seen as static, the model was actually designed to analyze the dynamics of situations. Understanding a situation will always have a certain degree of fluidity as the context evolves and knowledge advances.
Moving between Cynefin domains involves adapting to different situations using some of the following key strategies:
- Learning from multiple experiments: By gathering information from different sources, we can identify clearer patterns and solutions. For example, in SEO, this could be testing different optimization techniques and analyzing their impact.
- Adjusting constraints: Changing rules or policies can influence how a situation evolves and in which domain it’s located. In SEO, this could involve changing content creation guidelines and measuring their evolution.
- Setting and adjusting boundaries: Defining clear boundaries allows us to identify in which domain a situation is located and address it effectively. For example, shifting the boundaries of a complicated problem towards the complex domain can encourage experimentation and innovation, while moving the boundaries towards the clear domain could allow for more efficient and direct resource management. In SEO, it would be knowing when to discern between adopting experimental approaches versus more traditional and proven solutions. Nullius in verba is in my Twitter bio for a reason :)
- Decomposition: Breaking down a problem into smaller parts allows addressing specific aspects in different domains. In SEO, this could be dividing optimization tasks into simpler parts.
You can learn more about the movements between dynamics here.
Robustness and resilience of systems
In the Cynefin model, it’s also essential to understand two key concepts: robustness and resilience. These qualities help us face changing situations and challenges in our environment.
Robustness refers to how strong a system or process is in resisting changes or challenges without suffering significant damage. In the world of SEO, this could be the ability of a strategy to continue working well despite changes in search algorithms or the platform’s capacity to not lose its current state between updates.
On the other hand, resilience is the ability to recover quickly after facing changes or challenges. In terms of SEO, this means adapting to changes in search engine rules, market trends, or user preferences to continue growing. This has implications beyond SEO, as we will see in the next point.
It’s said that the route A (red) in the previous image is the most common, leading quickly to the realm of the obvious and focusing on solidity. However, according to Snowden, only path B (green), which involves experimenting in the complex realm, enables one to exit chaos and find patterns to advance towards the complicated environment.
The route B (green) is based on taking advantage of chaos, examining the system through experiments designed to fail safely and learning from observed behavioral patterns. Once a pattern-based solution is found, one moves to the complicated area and records the acquired knowledge.
On the other hand, the route C, known as «anticipatory awareness«, focuses on identifying early signals and facing the unknown and unthinkable of the chaotic or complex domain. In SEO, finding patterns and «listening to signals» is a fundamental part of the job.
In SEO, we seek both resilience and robustness in equal measure.
These two concepts are fundamental to the model as each domain has different levels of robustness and resilience. Clear and complicated systems tend to be stronger, as the relationships between causes and effects are easier to understand and predict. However, they may be less resilient when faced with drastic changes.
In contrast, complex and chaotic systems tend to be more resilient (or «antifragile» as Taleb would say), as they’re better prepared to adapt to changes and challenges. To effectively manage these systems, a more flexible and dynamic approach is needed, based on trying different strategies, monitoring their impact, and adjusting as necessary.
How to apply the Cynefin framework to SEO in 5 steps
The Cynefin framework is a pragmatic decision-making model. The main idea is to try to bring the unknown to the clear zone and its approach consists solely of resolving the current situation. Moreover, this framework helps you apply constraints that prevent jumping into chaos. If you don’t take this into account, problems may arise in project management or even lead to failure.
Now let’s take a look at the key elements of the framework and how we can apply them to SEO.
1. Don’t confuse a simple strategy with a complex or complicated situation
Some people tend to distort and simplify the existing situation. Others, as a consequence of excessive complacency, don’t see new ways to approach the situation. This can lead to disastrous decisions that can jeopardize the success of the SEO project. Be careful with the Dunning Kruger effect here.
For example, an SEO specialist might assume that adding keywords to content/metas is enough to improve rankings in search results. However, this clear strategy ignores the complexity and multiple factors that influence ranking. That is, it’s not acknowledging the limitations of a basic best practice.
It’s also possible that the specialist has used a clear strategy in other instances with success, but that strategy can’t be applied in the current context.
In the CRAWLO video/podcast episode titled ‘CRO, SEO, and Artificial Intelligence (AI) to Improve Your E-commerce’ I explain these concepts.
Generally, SEO is a complex system where there are very few things we know for sure, and where we must deal with other complex systems, both internal (organization, departments, people in organization departments…) and external (market trends, competitors, consumer behavior…).
When I talk about liminality, I mean it in its full sense, not just the liminality of the Cynefin framework, but also the ability to interconnect, relate, and transition between disciplines. One must feel comfortable at the threshold between different worlds. Like lungfish between aquatic and terrestrial environments. Like Reese’s Peanut Butter Cups, where someone came up with the brilliant idea of mixing chocolate with peanut butter, which Pharrell Williams called ‘the most incredible invention of all time’ in this video. Jokes aside, Pharrell is so great thanks to his liminal approach to music, and he demonstrates it in that video.
We need to be Renaissance men/women or T/PI/M/Comb-Shaped professionals, also known as generalist professionals.
I talk about this in this conference I gave at SEOnderground on Data Science applied to SEO. I also recommend the book «Range: Why Generalists Triumph in a Specialized World» by David Epstein.
2. Don’t approach complex situations with strategies for a complicated situation
In a complicated system, you find expert knowledge and thorough analysis. However, if you apply this approach to a complex system, there’s a significant risk of getting lost in analysis loops and inefficient processes (analysis paralysis).
For example, an SEO agency might try to create a comprehensive optimization plan that addresses all technical aspects (or assumed technical aspects that appear on a checklist). Although this approach may seem exhaustive, it can also lead the agency to get lost in the complexity of the situation and the client to waste money on implementations that may not work or may not even be necessary.
Doing this is not recognizing that SEO is an iterative process that requires continuous adjustments and adaptations based on results.
Instead of focusing on a detailed and rigid plan, the SEO agency should adopt a more flexible, adaptable, and empirical approach that allows them to efficiently address the complexity and uncertainties of the SEO world.
To achieve this, I propose encouraging serendipity through experimentation. Serendipity plays an important role in the context of finding the «best way» to do something or in advancing knowledge. It refers to the fortuitous discovery of something valuable or useful while searching for something else. In other words, it’s the unexpected finding of solutions or ideas that occur by chance.
In the case of clear and complicated systems, the disciplined and structured method is often the standard approach to discovering best practices. However, serendipity can also play a significant role in these advancements, as sometimes the most meaningful discoveries can be the result of accidental events or situations.
By including serendipity in the equation, we acknowledge that knowledge and best practices are not always obtained exclusively through systematic and predictable approaches. These fortuitous discoveries can challenge preconceived ideas and offer new perspectives or solutions that might otherwise have been overlooked.
Critical thinking should be promoted too, where the opinions of experts and managers, both internal and external, are challenged to combat entrained thinking.
In the video/podcast I shared earlier, I also discuss this issue.
3. Be able to recognize a chaotic situation
If the system leans towards a chaotic state, a quick reaction and appropriate leadership are needed to control the situation as quickly as possible. If one acts hesitantly or the leader decides on suboptimal constraints, this can cause long-term damage. This may happen due to an overconfidence, not paying attention to metrics, procedures, or the evolution of the structure itself.
For example, if the SEO consultant applies strategies that aren’t appropriate for the situation (due to being outdated or having little experience), forgets to review the quality of established processes, or tries to apply unpremeditated changes in response to a Google update… you know where this leads.
One way to act is not to act. Or the famous sayings: «An ounce of prevention is worth a pound of cure» by Benjamin Franklin, «You can use an eraser on the drafting table or a sledgehammer on the construction site» by Frank Lloyd Wright, or the popular «Measure twice, cut once.»
4. Learn what to measure in each domain
In the clear domain, KPIs are useful and necessary, as they focus on concrete results. For example, in SEO, a KPI could be the number of non-brand organic visitors arriving at the website. These indicators should be clear and easy to understand.
In the complicated domain, it’s important to set flexible goals that allow adaptation to different situations. For example, in SEO, instead of aiming for a specific number of users, the goal could be to improve traffic quality. Here, standards are key to allowing variations within certain limits.
In the complex domain, it’s more appropriate to measure direction and speed of progress (vector). In SEO, this could mean monitoring organic visibility, both overall and specific. In this context, visionary goals, like being a leader in your market niche, can provide direction but should be used with caution.
The image above was taken from Tom Critchlow’s newsletter, called The SEO MBA, which is highly recommended and aligns perfectly with the principles I want to convey in this article.
5. Create barriers to avoid ending up in chaos
We need to build a robust SEO system, capable of withstanding changes, adversities, and risks without falling into the realm of chaos. To achieve robustness, I propose several actions:
- Implementation of Guardrail Metrics: I first heard this concept in the book «Trustworthy Online Controlled Experiments» by Ron Kohavi. We need to establish untouchable but consensus SEO KPIs with other departments. If a change affects one of these SEO KPIs, the organization must backtrack. For example, if a change significantly affects Core Web Vitals, and you can demonstrate that an impact on these metrics affects your website’s visibility, then the change must be undone or not implemented in production.
- Creation of unit tests to validate that the platform’s state remains the same between releases.
- Design processes based on patterns found in the clear domain (best practices).
- Regular internal and external audits: Perform regular technical and content audits of the website to identify and address potential issues before they negatively affect SEO performance. For example, most of my clients have in-house SEO teams or even a global agency and a local SEO agency for each market, but they still need another point of view.
- Premeditatio malorum: Preparing contingency plans to face potential problems, such as changes in search engine algorithms, website technical failures, attacks, or reputation crises, is an approach that has some connection to Stoicism. This practice involves anticipating and reflecting on possible negative events to be better prepared for them. Contingency plans should include specific actions and designated responsible parties to address each type of problem, using a RACI matrix.
- Diversification of traffic sources: Don’t rely solely on organic search engine traffic. I believe this doesn’t need further explanation.
- Plan retrospectives to reflect on performance, decision-making, and the processes that led to a specific outcome. If you have experienced chaos, it’s important to encourage experts to challenge your point of view once the crisis has subsided.
- Continuous training and updating: Ensure that you and your team are up-to-date on the latest trends and best practices in SEO. Continuous training and updating will allow you to quickly adapt to SEO changes and reduce the risk of falling into the realm of chaos.
No matter how exceptional the solution we apply is, it won’t achieve success or be useful if it’s not employed in the appropriate context. The focus of this framework is on the importance of adapting different methods to varied situations. Although the framework is quite robust, the solution can be detrimental if used in an inappropriate context.
Quoting from Dave Snowden’s own article explaining the framework in HBR: «Good leadership requires openness to change at the individual level. Truly skilled leaders won’t only know how to identify the context they’re working in at any given time but also how to change their behavior and decisions to adapt to that context.»
It also don’t matter how much you know about your discipline if you aren’t able to achieve change.
Cynefin enables not only identifying the essence of the problem we face but also providing the process to follow. In other words, it assists project leaders in determining whether a problem can be solved with best practices (clear), good practices (complicated), or if it requires inquiry and experimentation (complex, chaotic).
Also this study: European Commission, Joint Research Centre, Rancati, A., Snowden, D., Managing complexity (and chaos) in times of crisis: a field guide for decision makers inspired by the Cynefin framework, Publications Office, 2021, https://data.europa.eu/doi/10.2760/353
And I also recommend reading the following books:
- «Thinking in Systems» by Donella Meadows.
- «Liminal Thinking: Create the Change You Want by Changing the Way You Think» by Dave Gray.
- «Never Split the Difference: Negotiating as if Your Life Depended on It» by Chris Voss.
- «Good Strategy/Bad Strategy: The difference and why it matters» by Richard Rumelt.
- «The Crux: How Leaders Become Strategists» by Richard Rumelt.
As promised, here I share Dave Snowden’s framework:
And a video where he explains it himself: