Why smaller units of value are almost always better

The MVP Misconception

I’ve been thinking recently about one of the most misused terms in software: the MVP, or “Minimum Viable Product.” I say misused not because there is something wrong with the term or the concept – it comes from an intention to start small and iterate – but because MVP has become shorthand for “the first thing we’re going to ship to customers.” In reality, it often means “the smallest bundle of features that stakeholders can agree on.”

But that’s not what MVP was meant to be. 

The original intent was much sharper: the smallest thing you can ship that tests a hypothesis. The focus is on learning, not on agreement. It’s about creating something just viable enough to prove or disprove an assumption, whether that’s about user behaviour, market appetite, or the feasibility of a technical approach.

The problem is that when MVPs are defined as compromises between stakeholders, they swell in size. What should be a quick experiment turns into a mini-release. Delivery slows, costs rise, and the real purpose of the MVP is lost. By the time the ‘MVP’ reaches customers, so much has been invested that nobody really wants to learn from it – they want it to succeed, regardless of the evidence.

Thinking in ‘Units of Value’

That’s where the idea of thinking in smaller “units of value” helps. 

If I asked you how many marbles you could fit into a glass, you would probably give me a decent estimate. If I asked you how many you could fit into a swimming pool, you’d probably get it wrong by millions. 

The same applies to these units of value. The tighter we can keep them, the better chance we have of estimating them accurately.

A unit of value is simply the smallest increment that delivers a meaningful outcome. 

• Sometimes that’s a tiny feature toggle exposed to a handful of customers. 
• Sometimes it’s a rough prototype tested in a single workshop. 
• It could be a thin vertical slice of functionality that looks unfinished but proves a workflow end-to-end. 
• It might even be an experiment with the only intention of generating enough data that reveals whether some widget or other will work at scale.

What matters is not that the unit looks polished or complete, but that it delivers something useful – feedback, data, insight, or a signal that confirms you’re on the right track. And because these increments are small, they move through pipelines quickly, they reduce the risk of wasted effort, and they keep teams in the habit of shipping and learning rather than waiting for perfection.

The Mindset Shift

If MVP has become too weighed down by politics and compromise, reframing delivery around units of value helps bring the focus back to outcomes. The conversation shifts from: 

“What is the minimum feature set we can all agree on?” 

to 

“What is the smallest thing we can ship that teaches us something important?” 

In a world where the ability to learn faster than your competitors is the real differentiator, smaller is almost always better.

This mindset is the essence of the bestselling book Lean Enterprise, and we’re proud to partner with co-author Barry O’Reilly next week to host a series of exclusive executive breakfasts. Visit our events page for more details!

Why this matters more than ever for AI

One area where this goes beyond just being good advice is when building AI tools. This is because AI systems are inherently uncertain. Outputs vary, edge cases multiply, and user trust depends on nuance that you can’t fully anticipate up front. 

The only reliable way to find out whether an AI feature delivers value is to put it into the world and watch what happens. That means experimentation and iteration aren’t just desirable – they’re essential, ignore it and your project is doomed. You’ll end up with products that look promising on paper but fail in practice because no one took the time to validate value in the only way that matters: through repeated, real-world experiments.

—

At Waracle, we partner with ambitious organisations to embed an agile mindset, helping them build not just great digital products, but powerful engines for learning and growth. If you’re ready to escape the MVP trap and accelerate your delivery of real-world value, let’s talk.