How to answer a “Should [company] do [X]?” product strategy interview question at the Director and VP level
- Nancy Chu
- 3 days ago
- 10 min read
A question like “Should Google offer fine-tuning as a service?” can leave you wondering how you are supposed to approach it. You can see arguments for yes and no, but that still does not tell you how to decide.
I think about these questions as bets. Start with the company, not the idea:
What is this company trying to accomplish, and what would have to be true for this investment to help it get there?
That prompt gives the rest of the answer a direction.
First, make a reasonable assumption about what the company wants.
Then trace the investment forward until you can explain how it helps the company make money, strengthen an advantage, expand into a valuable market, reduce an important cost, or learn something that unlocks a larger opportunity.
From there, define the conditions that would make the bet worth taking.
Then identify the assumption carrying the most weight and what would make you change the recommendation.
For Director, VP, (or L7 - L9 using Meta/Google terminology) interviews, it’s important to ensure that your thinking is visible: why this investment makes sense for this company, under which conditions, and what has to be true for the logic to hold.
Why “Should company X do Y?” product strategy questions feel so difficult
These questions are broad because the decision sits above the product itself. You are deciding whether the company should invest at all.
In an interview, you usually do not have the company’s internal data or full strategy. That makes it tempting to start listing everything that might matter: market size, customer demand, competitors, company strengths, revenue, build versus buy.
The problem is that those are inputs. They do not tell you which ones should drive the decision.
A useful way to get oriented is to treat strategy as a bet under uncertainty. Annie Duke’s book Thinking in Bets gave me useful language for this. You rarely know everything you would like to know before making a consequential decision.
So instead of trying to reconstruct the company’s actual strategy, make a reasonable assumption about what the company wants and ask what would have to be true for this investment to help achieve it. That gives you a way to decide which facts matter and which ones do not.
Start with what the company wants, then define what would have to be true
The three examples below all use the same sequence.
First, make a reasonable assumption about what the company is trying to accomplish.
Then ask: What would have to be true for this investment to help the company get there?
Follow the logic far enough that you can see the business mechanism. A useful idea by itself is not enough. Keep asking why the company would care until the chain reaches something economically or strategically meaningful.
Then pressure-test the conditions. What would have to be true for you to make the bet? What would have to be true for you to walk away? The conditions will change by question. We'll go over 3 examples in this article.
Finally, identify the assumption carrying the most weight. If that assumption changes, should your recommendation change too?
A practical way to frame the decision is: I would make this bet if X, Y, and Z are true. Of those, X is the assumption I have the least confidence in, so that is what I would validate first. If X turns out not to be true, my recommendation changes.
Example: Should Google offer fine-tuning as a service?
Take the prompt “Should Google offer fine-tuning as a service?”
In an interview, you may not know whether Google already offers something similar, how much customers use it, or what the unit economics look like. You do not need those facts to begin reasoning.
Start with what you can reasonably infer about the business. Google Cloud wants more companies to adopt its platform and expand their usage over time. A managed fine-tuning service could be valuable if it gives customers a reason to bring more AI workloads onto Google Cloud.
That gives you a useful self-prompt: What would have to be true for fine-tuning to become a starting point for a bigger Google Cloud relationship?
First, assume companies find it too expensive, slow, or complicated to customize AI models on their own. If Google can make fine-tuning meaningfully cheaper and easier, that removes a reason for customers to build the capability themselves.
Second, fine-tuning requires access to the customer’s data. If the easiest path is to bring that data into Google Cloud, the fine-tuning service does more than generate service revenue. It creates a reason for the customer to move an important workload, and the data behind it, onto Google’s platform.
Third, once the custom model and its data live in Google Cloud, the customer may have a practical reason to keep running the surrounding AI workload there too. Moving large datasets between cloud providers can be costly and slow. The finished model also needs ongoing access to the data and infrastructure around it.
That creates the door-buster logic. Fine-tuning can be valuable as the starting point for a much larger relationship because it can lead naturally to data storage, model serving, AI operations, and future software built on Google Cloud.
You would also pressure-test the opposite case. If customers can fine-tune models easily on their own, if they do not care where that work happens, or if the service creates heavy support and compute costs without leading to broader Cloud usage, the bet becomes much less attractive.
So your recommendation could sound like:
I would make the bet if managed fine-tuning solves a real capability gap for customers and creates meaningful downstream adoption across Google Cloud. If it behaves like a one-off service with little effect on retention or expansion, I would not prioritize it.
The point is not whether you guessed Google’s actual strategy. What you are demonstrating is creating a logical set of conditions that would make the investment attractive or unattractive.
Example: Design Slack for plant enthusiasts
Now take a stranger prompt: “Design Slack for plant enthusiasts.”
Before thinking about plants, start with Slack's parent company Salesforce. What would Salesforce have to want for investing in this use case to make sense?
One reasonable assumption is that Salesforce wants to expand the number of business workflows its software can help companies manage and automate. Today, many of those workflows happen entirely inside software. Sales teams work with customer records. Service teams work with support cases. Employees collaborate through messages and documents.
But many large industries also have important workflows driven by what is happening in the physical world. A manufacturer may need to respond when a machine overheats. A logistics company may need to respond when a shipment changes location or temperature. An agriculture business may need to respond to changes in soil, weather, or crops.
If Salesforce wants its software and AI to automate more of those workflows, it needs to learn how to combine human collaboration with signals coming from the physical world. Doing that well could let Salesforce support more valuable workflows in industries such as manufacturing, logistics, agriculture, and field service.
Now come back to plants.
Plant care has many of the same ingredients on a much smaller scale. People communicate about what is happening, while photos, soil moisture, temperature, light, and sensor readings provide information about the physical environment.
That gives you a condition for the investment: I would make the bet if Salesforce wants to expand into enterprise workflows that depend on physical-world data, and if a plant use case gives the company a cheap, low-risk way to build and test capabilities that can later transfer to those larger markets.
If Salesforce does not see those physical-world workflows as an important growth opportunity, or if what it learns from plants would not transfer, there is much less reason to invest.
The plant community is the test case. The larger bet is whether learning to connect software, AI, and physical-world signals helps Salesforce expand the kinds of enterprise workflows it can own.
Example: Improve Google Flights after booking
Now take the prompt: “Improve Google Flights’ post-booking experience.”
Before thinking about features, start with Google. What would Google have to want for investing in the post-booking experience to make sense?
A more useful company-level assumption is that Google wants to understand user intent and preferences well enough to make its products more relevant. Travel is especially valuable because a flight search reveals unusually strong intent. Where someone is going, when they are going, who they may be traveling with, and what kind of trip they are planning can help Google understand what that person is likely to need next.
That gives Google a reason to care about the post-booking experience beyond simply being more helpful. Before booking, Google learns that someone is considering a trip. After booking, the intent becomes much more concrete. The destination and dates are known, and the next set of needs starts to emerge.
Those needs might include where to stay, what to do, where to eat, how to get around, what to pack, or how to organize the trip. Each interaction can help Google learn more about the traveler’s preferences while also giving Google more opportunities to make Search, Maps, Gemini, and other products more useful for that person.
So the condition becomes: I would invest in the post-booking experience if it helps Google build a richer understanding of a traveler’s intent and preferences, and if that understanding can improve the relevance and value of Google’s broader ecosystem.
Then pressure-test it. If the post-booking experience does not teach Google anything meaningfully new about the user, if those signals cannot improve other Google products, or if travelers do not return to Google for the decisions that follow booking, the investment becomes less attractive.
The larger bet is whether Google Flights can turn a high-intent travel moment into a deeper understanding of the user that makes the rest of Google’s products more useful.
A strong product strategy recommendation makes the conditions visible
The three examples work because the recommendation comes at the end of a causal chain.
Google Cloud wants broader platform adoption. Fine-tuning matters if it can pull an important workload and the customer’s data onto Google Cloud, then lead to more ongoing usage.
Salesforce wants to expand the workflows its software and AI can support. A plant use case matters if it helps Salesforce build capabilities that transfer into larger enterprise markets.
Google wants to understand user intent and preferences well enough to make its products more relevant. A better post-booking experience matters if it reveals useful signals that improve the broader Google ecosystem.
In each case, the product idea becomes strategically meaningful because you can explain how it advances something the company cares about.
That also makes the assumptions easier to see. You can identify what you know, what you are assuming, which assumption carries the most weight, and what new information would change the recommendation.
Use the company goal to decide what belongs in the answer
A coherent strategy answer is selective. The company goal tells you which facts, risks, customer needs, advantages, and market dynamics actually matter to the investment.
That is why starting with the company changes the answer. You are no longer trying to touch every category in a framework. You are looking for the few conditions that determine whether this particular bet creates meaningful value for this particular company.
Once those conditions are clear, the rest of the answer has a natural through line. Each point either strengthens the case for the investment, weakens it, or helps you decide which assumption to validate first.
How to start your next “Should company X do Y?” product strategy answer
Before you analyze the product idea, answer this first:
What is this company trying to accomplish?
Then ask:
What would have to be true for this investment to help the company get there?
Follow that chain until you can explain why the company would economically or strategically care. Then write down:
I would make this bet if…
I would not make this bet if…
Finally, identify the assumption that separates those two answers.
That gives you a concrete way to approach a broad product strategy question without pretending you know the company’s internal strategy. You are making your assumptions explicit, reasoning from the company’s goals, and showing exactly what would make the investment worthwhile.
FAQ
How should I answer a “Should company X do Y?” product strategy interview question?
Treat it as a decision under uncertainty. Identify the conditions that would make the investment strategically worthwhile for that specific company, then make a recommendation and name the assumption most likely to change your answer.
What framework should I use for a “Should company X do Y?” product strategy question?
Start with what the company is trying to accomplish. Then ask what would have to be true for the proposed investment to help the company get there. Follow the causal chain until you can explain why the company would economically or strategically care, then identify the assumption that matters most to your recommendation.
What makes a product strategy answer Director or VP level?
A senior leader’s answer shows judgment about what matters to the decision. It connects the company’s goals, advantages, customer demand, risks, and tradeoffs into a clear recommendation and explains what would change that recommendation.
Do I need to choose yes or no in a product strategy interview?
You should make a recommendation. The stronger answer also explains the conditions under which that recommendation holds and what new information would cause you to update it.
How is a product strategy interview different from product sense?
Product sense often asks you to identify a valuable customer problem and design a product around it. A “Should company X do Y?” product strategy question asks whether the company should make the investment in the first place. That requires company-level reasoning before moving into solutions.
How do I show strategic thinking without sounding vague?
Make your assumptions concrete. State what has to be true, which assumption matters most, what evidence supports it, and what would change your mind. That gives the answer a clear decision path.
Where to start
If a product strategy question leaves you unsure how to approach it, start with two prompts:
What is this company trying to accomplish?
What would have to be true for this investment to help the company get there?
Then keep following the logic until you can explain why the company would care enough to make the investment.
For related examples of leadership-level product thinking, read Why your product sense answer doesn’t feel like a slam dunk at the Director and VP level and How to 10x Google Gemini.
About Nancy Chu
Nancy Chu is a product leadership coach who works with senior product leaders preparing for Director and VP-level, or L7-L9 roles at companies like Google, Meta, Microsoft, Stripe, and more. Her coaching focuses on the thinking underneath strong executive communication, especially when the problem is ambiguous and there is no obvious right answer.
Her clients have landed offers with total compensation packages up to $2.9 million in the first year. See more client wins.
Comments