Making Better Product Decisions: Approaches and AI Prompts for Product Managers
One of the most important skills for any product manager is making good decisions. I’ve found that having a structured approach to decision-making not only improves outcomes but also helps communicate your rationale to stakeholders.
The key thing is to be clear about the outcome that you want to achieve. Understanding what you want to happen will help in framing the decision in the right way. It’s one of the reasons why I’m a big fan of my friend Martin Eriksson “Decision Stack” framework as it makes sure that decisions ladder up to the vision and objectives that a business is looking to achieve.
Let me share some other approaches that I’ve found helpful when make decisions:
Type 1 vs Type 2 decisions
This framework distinguishes between Type 1 — irreversible and high-stake — decisions and Type 2 decisions which are reversible. With Type 1 decisions, the door closes behind you once you’ve made a decision, whereas Type 2 decisions allow you to re-enter if needed.
S.P.A.D.E.
S.P.A.D.E. is a useful framework to use when you need to make tough decisions that impact multiple teams or require significant trade-offs:
- Setting — Define the decision scope, timeline, and value metric impacted.
- People — Identify who needs to be involved in the decision-making and their roles (approver vs consulted).
- Alternatives — Generate multiple options beyond the obvious binary choice, assessing the viability of each option.
- Decision — Make the decision based on data and voting, if required.
- Explain — Communicate the decision and rationale to stakeholders.
I’ve found S.P.A.D.E. particularly helpful in rapidly getting to the core of the tradeoffs to decide on.
Annie Duke’s approach
Annie Duke is a former poker player turned decision-making expert and author. Duke’s approach is really valuable for decisions with a high degree of uncertainty, a regular occurrence in the world of business and product management. I particularly like the ‘three Ps’ which Duke covers in her book “How to Decide”:
- Preferences — Identify your preference for each outcome. The simplest way to do this is to list the potential outcomes in a decision tree, outlining decision outcomes and their likelihood (see a good example below).
- Payoffs — For most outcomes there will be things you could gain (upside) and things you could lose (downside). These gains or losses are called payoffs. If an outcome moves you toward a goal, the payoff is positive. If an outcome moves you away from a goal, the payoff is negative.
- Probabilities — Estimate the likelihood of each outcome materialising. Such estimations will be based on guesses and Duke stresses that good decision-making requires a willingness to guess.
Duke has also written about ‘pre-mortems’: group activities designed to identify upfront the ways in which a product or project could go wrong. In my experience, pre-mortems can be really useful in identifying your riskiest assumptions and changing your plans/decisions.
Duke recommends using the pre-mortem to set what she refers to as ‘kill criteria’ — a set of signals that will indicate that things aren’t going as planned and that we should act on. For example, if we see that we can’t get a first sales deal for the product within 3 months post-launch, we should kill it. The point is to link these signals to specific actions, so that we can act as soon as we see the signal.
Using AI to make decisions
In addition to the approaches I mentioned above, I increasingly use ChatGPT or Claude to help my decision-making. AI can help sense check my decision options or offer relevant tradeoffs to consider. Here are four effective prompts I’ve used to help make decisions:
1. Generating alternative options:
I’m considering [describe situation]. Beyond [current options], what are 3–5 additional options I might not have considered?
2. Evaluating decisions:
I’m deciding between [Option A] and [Option B] for [goal or context].
My criteria are [list decision criteria].
Please evaluate both options against these criteria, noting any additional considerations I might have missed.
3. Considering short vs long-term impact and tradeoffs:
I’m evaluating [Option A] versus [Option B] for my product. My core product objectives are [list key objectives]. Please create a clear comparison table showing:
SHORT-TERM (1–2 years):
- Key benefits of Option A
- Key limitations of Option A
- Key benefits of Option B
- Key limitations of Option B
LONG-TERM (5+ years):
- Key benefits of Option A
- Key limitations of Option A
- Key benefits of Option B
- Key limitations of Option B
Then highlight the 2–3 most critical tradeoffs I should consider when making this decision, focusing on where short-term gains might create long-term constraints or vice versa.
Please help me evaluate these tradeoffs by:
- Identifying potential inflection points where short-term gains might become long-term liabilities
- Suggesting ways to capture short-term wins while preserving long-term optionality
- Analysing what market or user signals would indicate we’ve made the right choice
- Recommending checkpoints to reassess this decision over time
4. Identifying decision blind spots:
Here’s my planned decision: [describe decision].Based on this, what potential risks or unintended consequences should I be aware of that I might have overlooked?
The key is using AI as a ‘co-thinker’ rather than delegating the decision itself. Like with all quality prompts, the more context and references you can provide to the AI, the better.
Main learning point:
No approach is perfect for every situation. The goal is to get to the heart of the decision fast, with the right level of insight and conviction. Be clear on what you’re trying to achieve, then apply an approach that provides enough structure without unnecessary complexity.
What decision frameworks have you found most helpful in your product work?
