The discovery-delivery ratio nobody sets on purpose

"How do we balance discovery and delivery?" is one of the questions I hear most from clients.

And here's the thing most teams miss: there's no single right ratio. The split should shift - sometimes dramatically - based on two things. Your stage, and the risk you're actually facing.

Early stage: roughly 80% discovery, 20% delivery

When you're early, you're navigating unknown territory. You don't yet know who the customer really is, which problem is worth solving, or whether anyone will pay to have it solved. The dominant risk is building the wrong thing entirely.

So most of your energy should go to discovery - validating that a problem is real and worth solving before you pour months into a solution. Building fast at this stage doesn't help you. It just gets you to a confident, expensive wrong answer faster.

As you mature: the ratio flips to roughly 20/80

Once you've found a problem worth solving and a solution that works, the picture changes. Now the problem space is well understood. You're optimizing and extending something proven, not pioneering into fog.

At that point, heavy discovery becomes the wrong bet - you'd be re-validating things you already know. Delivery should dominate, with discovery running quietly in the background to catch the next shift before it hits you.

The real failure mode is the default

Here's what actually happens on most teams: nobody sets the ratio at all. And when nobody decides, it doesn't land at a thoughtful middle. It drifts to almost pure delivery.

Why? Because delivery is visible and measurable. Story points, velocity charts, a burndown that goes down. Discovery looks like "talking to users" and "not shipping," so it's the first thing cut when the quarter gets tight. The team feels productive - and quietly gets very good at shipping the wrong things efficiently.

That's the trap. An unmanaged ratio isn't neutral. It's a slow slide into a feature factory.

Running both at once is an operating-model choice

The teams that hold a healthy ratio don't alternate between discovery months and delivery months. They run both continuously - discovery and delivery as two tracks moving in parallel. That's dual-track, and it only works when a product manager, designer, and engineer own the outcome together as a product trio, rather than passing work down a line.

Which is the real point. "How much discovery vs delivery" isn't a scheduling question you answer once. It's a property of how your product organisation is wired - and rewiring that, so discovery stops being the thing that gets cut, is exactly what Product Model Transformation is for.

So the honest answer to the balance question is a better question: what's your riskiest unknown right now, and does your current ratio reflect it? If you can't remember choosing the ratio you're running, you didn't. It chose you.