Hypotheses prioritization methods can be used to test ideas consistently and structuredly. An effective product hypothesis prioritization process ensures stakeholder approval and inspires the team with a fresh vision, minimizing the risk of investing in undesirable features.

Hypothesis prioritization methods are a set of principles or strategies that help define the next steps. Here are some popular methods for prioritizing product hypotheses: 

• AAARRR. This framework is a marketing funnel that reflects the main stages of a customer's interaction with a product.

• Lean Prioritization. This is a matrix tool that maps "value" to "complexity". It helps in making decisions and determining what is important, what is risky and where to focus your efforts. Entrepreneurs in startups often use this matrix to prioritize product development. 

• RICE / ICE. Each indicator is assigned a score in points and then the numbers are multiplied. The hypotheses with the highest scores are given a higher priority. 

AAARRR 

The AARRR metrics system is modeled after a conversion or marketing funnel. For ease of pronunciation, people called the system "pirate metrics.” The acronym stands for Awareness, Attraction, Activation, Retention, Referral and Revenue. Each metric corresponds to a stage of the customer's journey. By examining changes (or lack thereof) in user behavior at each stage, product managers can identify areas of decline and measure user engagement or dissatisfaction. For more details about AAARRR you can read here.

You analyze your product and correlate metrics with one of the funnel stages. Then, you correlate each key stage metric (usually conversion rates to the next funnel step) with benchmarking for your industry and product stage, or with a target value based on your unit economics calculations. The metric closest to the funnel's beginning with the most significant negative deviation from the target value should be optimized first. The logic is that each stage should consistently lead the user to a key business goal (revenue). If a bottleneck is forming somewhere, all your efforts will be wasted at this stage. Once you have identified the critical metric to optimize at the moment, you need to analyze your metric tree to see which metrics (=user actions) influence it and then form hypotheses to improve them. We recommend using ICE/RICE methods below to prioritize hypotheses affecting the same step of the funnel.  

Lean Prioritization 

This method is usefull when launching a new product, creating an MVP, or when resources are scarce. Pros: Great for identifying easy wins and immediate opportunities. Cons: Gets complicated when there are a lot of items for prioritization (hypotheses, features). 

Lean Prioritization is a 2 x 2 matrix on which "value" is plotted against "complexity". For this model to be effective, the value and complexity of each hypothesis (validation or implementation) must be quantified. 

Value represents the benefit derived from the feature by both customers and the company. Will the hypothesis about UI/UX, functionality, or marketing alleviate customers' pain points, improve their daily workflow, and help them achieve their desired outcomes? Furthermore, will the hypothesis have a positive impact on your business' bottom line? 

Complexity (or effort) is what it will take for a company to validate (or implement) this hypothesis.  

Priority Score. When these criteria are combined, they form several quadrants that objectively indicate which set of features should be developed first, which features should be considered next, and which hypotheses could be ignored.

Quadrants created using this matrix include:

• Quick wins (top left): These high-value, low-complexity functions represent the "low-hanging fruit" opportunities in the company that should be pursued with the highest priority. 

• Major Projects, High Stakes or Potential Opportunities (top right): Initiatives in this category have a good value but are considered too risky due to resource and cost implications.  

• Possible features (lower left): This quadrant typically includes features that are "nice to have", such as minor UI improvements and potential future ideas.  

• Time-sink features (lower right): These are initiatives that should be avoided because they consume valuable resources without delivering significant value. 

Lean Prioritization is a valuable tool for teams working on new products. Its simplicity makes it particularly useful for making quick and objective decisions. In addition, for teams facing resource constraints, this matrix is an effective way to identify easily realizable opportunities.  

RICE / ICE

ICE

Pros: Simple model sufficient for relative prioritization. Cons: Subjective approach; lacks data-driven conclusions. 

If you need a quick prioritization system, the ICE model is ideal for those just starting to prioritize product initiatives, but it lacks the data-driven objectivity of other systems. 

• I - impact. What impact will this feature have on the user? Impact is rated on a scale of 1 to 10, where 1 is the minimum impact and 10 is the maximum impact.

• C - confidence. How confident can we be in our assessment of the reach and impact parameters? How much data do we have to back up our estimate? Confidence is also rated on a scale of 1 to 10, where 1 is low confidence and 10 is high confidence.

• E - ease. How easy or difficult is it to implement (or validate) this hypothesis?  It is also rated on a scale of 1 to 10, where 1 is difficult and 10 is easy. 

ICE score = Impact * Confidence * Ease 

RICE 

Pros: Quantifies the overall impact on the time invested Cons: Predefined evaluation factors limit customization; may not fit your organization's needs.  RICE is a 4-factor scoring system. 

• R - reach. How many people will be affected by a feature over a certain period? For example, the UI/UX hypothesis about the feature under study is used by 60% of the total audience of 1,000 users. So, changing the UI/UX for it will affect 600 users.  

• I - impact. What impact will this feature have on the user? For example, how much feature activation rate could increase, after implementing this hypothesis about UI/UX? 3 - massive impact. 2 - high impact 1 - medium impact 0.5 - low impact 0.25 - minimal 

• C - confidence. How confident can we be in our estimate of the reach and impact parameters? How much data do we have to back up our estimate? Be honest with yourself. 100% - high confidence 80% - medium confidence 50% - low confidence.

• E - effort. How much time will it take to invest in this initiative (product, design, development)? The parameter is estimated in person-months.

RICE score = Reach * Impact * Confidence / Effort