Identifying the most important problem to solve

The #1 job in Product is to identify the most important problem to solve. But it also can be an intimidating and overwhelming exercise. Here’s how I do it.

1. How’s the business performing? I start by developing an understanding of how the business creates value: the marketing and sales funnel, engagement with the product, revenue generation and retention. I’ll then compare our actuals against our forecast to see if we’re falling behind.¹

2. Where do we need to focus the most on getting better? I’ll divide these into two buckets.²  

   1. The first is more operational: we have what it takes to get better, we just need to improve incrementally (better process, a couple key hires, small optimizations in our funnels, etc.). These might turn into OKRs for operational teams or maybe a product squad but unlikely to be the Most Important Problem we are looking for.

   2. The second is more structural: we’re not going to move the needle just by trying harder, we need a concerted effort to redefine our product or our value prop or our way of operating. This is where Most Important Problems tend to live.

3. What customer problems can we solve that would result in improving our structural focus areas? This is where we go do our customer discovery. We’re analyzing our data, talking to customers, running design sprints, iterating on design prototypes, etc. What we find, at the end, is our Most Important Problem. If we solve this, we’re not only creating value for customers, we’re also moving the needle for the business in a way that otherwise would be unlikely to happen through business-as-usual effort.

What I like about this approach is that, by starting with the business needs, I know that whatever customer problems I then uncover are also going to be valuable to solve for the business. A lot of PMs will make the mistake of starting with customer discovery and never bothering to check to see how much impact they would have to the business, even if they solved the customer problem to perfection³ .

The Workshop

This is a newsletter-only section where I share a half-baked idea in hopes that y’all who are smarter than me can work it out with me.

Any reactions to Apple’s WWDC from earlier this week?

My first reaction was how un-Steve Jobs the announcements were. Letting iPhone users change the color scheme of their app icons (often in ways that are just ugly)? Shoving LLMs (of inconsistent quality and accuracy) into all kinds of nooks and crannies just to show Wall St that Apple is hip to AI too? And the asks-you-every-time “do you want to ask ChatGPT?” pop-up dialog that achieves a privacy goal, yes, but at the expense of the user experience?

My second reaction was how good at “we can do things no one else can because we also own the hardware and OS” Apple is. The AirPod background noise microphone improvements, the super cool Calculator + Apple Pencil app (you can literally draw math and it’ll calculate, graph, and solve equations on the fly as you write), the way their advantage in silicon chip design & manufacturing means they can support on-device AI computations (for superior privacy and latency) in a massively better way than others.

My third reaction was curiosity in how all this is going to play out amongst the big AI companies. Yes, Apple chose OpenAI / ChatGPT 4o. But they also made clear they designed the integration to be modular, that they can swap out OpenAI for something else that is better in the future. Great strategy move by Apple. The commodification of AI (which really does seem like the trend, IMO) puts pressure on OpenAI to either accept a future where they (like Nvidia / Intel / AMD) are constantly competing to build a better mousetrap or they become an aggregator of demand in their own right and benefit from the defensible network effects produced by that. If you’re interested in the underlying strategy framework, read this 2017 Stratechery post about the Smiling Curve.

1  And if we’re an early stage startup and we don’t have it, create your own. I’m serious. Open up a spreadsheet, drop in the funnel and other business levers into the rows, and months across the columns. Drop in the actuals for the months past, and then start wiring together the formulas for the future that reflect how the funnel and business levers impact each other. For example, if we have a row for traffic and sign-up conversion, then we can have a row for registered users that is the multiplication of traffic * sign-up conversion. And from there, we can convert registered users to revenue through that conversion rate, etc.

2  In truth, there is a third bucket, which is more internal: the reason why we’re struggling is due to our culture and internal processes. We’re too slow, we’re afraid to try things, our founder is dictating what we should do and they don’t have good enough product sense, etc. Sometimes the Most Important Problem lives here.

3  Or, compared what they want to do to the opportunity cost of solving a different problem instead.