Thinking patterns for PMs

A typical decision-making process involves the following steps

• Problem formulation
• Creating a solutions space
• Reducing the solutions space to potential solutions
• Choosing the right solution

The patterns along with the checklists can help product leaders reduce the risk of being wrong.

The list of patterns is by no means comprehensive. Where it makes sense, the patterns have been clustered together under a common category. The patterns could be applicable during multiple stages of the above process. Sometimes patterns could be combined and applied simultaneously. Not all patterns are relevant for every problem and one needs to apply judgment to identify the most relevant pattern(s) for a problem. One could reduce the risk of having to pick the right pattern by applying all relevant patterns and comparing the results to choose the most appropriate one.

• Constraint-free Thinking – This category includes patterns that liberate the thinking process from the boundaries and constructs of the current problem by relaxing self-imposed constraints. These constraints often hinder thinking but are not critical for the problem. Patterns in the category include – Long term – Timeframe of desired outcome is a common self-imposed constraint that subconsciously influences thinking. We have often heard Jeff say – “Long term thinking is a huge lever”. This pattern encourages stepping back and asking if we would do anything different if the time horizon was much longer than what was being considered during the decision making process.

Example : Early in the Amazon’s history, They had to decide if we would allow negative customer reviews on the website. There was pressure from the publishers to remove the negative reviews.  It is not difficult to imagine that if they were thinking short term, we could have made a different decision. Another example is our decision to allow third-parties to sell on the same detail page as retail. In both cases we took it as an article of faith that helping customers make purchase decisions would translate into customer trust in the long term and used this conviction to optimize for the long term.

• Inspiration and Reuse – This category includes patterns that bring humility to our thinking. These patterns encourage us to get inspired by other successful executions and reuse where it makes sense. One way to be right is to take advantage of something that has already proven to be right. Patterns in the category include – Benchmark – This pattern encourages a “divine discontentment culture” and helps leaders raise standards when making decisions. Being customer centric does not preclude us from comparing ourselves on discreet customer facing attributes against companies with great customer experience. Benchmarking allows leaders to calibrate their products against world class and set higher standards. Leaders can look externally (examples include multiple benchmarking programs run by our Benchmarking team) or internally (e.g. across teams, geographies) for comparisons. In some cases, benchmarking can be automated and done continuously. Our Competitive Monitoring Tools are good examples. Reuse – This pattern encourages using existing solutions (systems, processes, and standards) that already work well for the exact problem being solved. The solution can be internal or external. Reusing an already successful solution eliminates the need to solve the same problem again, thus reducing likelihood of mistakes.
• Measurement and Analysis – This category includes patterns that encourage using data when making decisions. Jeff’s 2005 Shareholder Letter talks about this pattern – “Many of the important decisions we make at Amazon.com can be made with data. There is a right answer or a wrong answer, a better answer or a worse answer, and math tells us which is which. These are our favorite kinds of decisions.” Jeff goes on to say – “The above decisions require us to make some assumptions and judgments, but in such decisions, judgment and opinion come into play only as junior partners. The heavy lifting is done by math.” Patterns in the category include – Math-based – Sometimes the key elements of the problem are susceptible to measurement and analysis. There is a right answer or a wrong answer. In some cases there is historical data that allows us to accurate predict the answer. In such cases, this pattern encourages using data and math to find the right answer instead of relying on debate and judgment. Using math to do the heavy lifting increase our odds of being right.

• Precision and Clarity of thought – This category includes patterns that help ensure precision and clarity of thought on the things that matter. Patterns in the category include – Working Backwards – Starting with the customer and working backwards helps drive precision, clarity and alignment on what matters most to the customer. This pattern encourages using the Working Backwards methodology. The customers can be internal or external. The methodology forces the team to be clearly and crisply articulate who their customers are, the problem being solved, the desired customer experience and why would their customers care about the experience. Having clarity on the desired customer experience and working backwards from it increases the likelihood that decisions would optimize what is most valuable to the customer thus increasing their likelihood of being right. Objective and Principles – This pattern encourages coming up with the primary objective of any solution and few generalized guiding principles specific to the problem domain. Leaders clearly articulate their current objective and then step back to come up with tenets that would be helpful for the class of problems with similar objectives. Stepping back from the details of the problem allows coming up with a set of principles that are more closely aligned with the mission, vision and the core values. At the same time, the principles are relevant to the problem domain providing leaders with a powerful set of tiebreakers when solving specific problems in the domain. These principles can be perfected with time. Since these evolve with time, learning from prior decisions gets encoded as principles for future reference. Leaders can rely on these time tested principles to increase the likelihood of their decisions being aligned with what matters, thus increasing the likelihood of them being right.  Future leverage is a nice side effect of this pattern.
• Restatement – This category includes patterns that help with reducing the complexity of the problem. Patterns in the category include – Thought experiments – This pattern encourages using thought experiments to reduce complexity and risk. Wikipedia describes a thought experiment as “a proposal for an experiment that would test a hypothesis or theory but cannot actually be performed due to practical limitations; instead its purpose is to explore the potential consequences of the principle in question.”

For example, the need to come up with a scalable way to add leadership to our distribution system as we grow to a $100B company can be restated as “the need to hire a capable and senior operations leader for our FCs every month”. This makes it easier to visualize the scale of the problem and helps in effective decision making. It is possible that absent this visualization, the leader might decide on a process that does not meet the significant scaling need.