Case Study: The Storefront

This is the whole group applied to one example. A product person was asked to walk our online store, the Storefront, and say how to make it the best. What came back was a flat list of surface bugs. This doc takes that exact assignment and shows what the answer should have been, using every step from Product Sense through Prioritization.

It’s the same store the Strategy group diagnoses, so you can see the two halves of the craft meet: strategy sizes where value leaks; product sense decides what to build about it. Read this as the model. When you’re handed a product and asked “how do we make this the best,” this is the shape of the answer we expect.

What We Asked For, and What We Got#

The brief: walk the Storefront and tell us how to make it the best store of its kind. Here’s the response, in full:

A couple of these are real. But step back and ask the three questions from Product Sense: who is this store for, what are they trying to do, and would fixing all six sell one extra order? The list can’t answer any of them. It’s a walk-through of one person’s friction on the pages they happened to land on, sorted by nothing, aimed at nothing. It never names a shopper, never states the goal, never notices that orders have been flat for two straight halves. It’s the symptom-collector’s default output, and it’s what this group exists to replace.

Here’s the same assignment, done properly.

Step 1 — Comprehend#

What is this product, and what’s it for? An online store: it takes a visitor from “I might want something” to “I bought it, and it’s on the way.” It’s an established store, so the question is less “what should exist” and more “where is it weak, and where could it be the best.”

What’s the goal? Completed orders. And the uncomfortable fact the cosmetic list ignored: orders are flat at ~1,900/week and have been for two halves, on traffic we already have. So the goal is sharp: convert more of the shoppers we already get into paid orders, worth about +350/week to break the plateau. Every judgment below is measured against that.

Step 2 — The Personas#

Three people use this store, with different jobs and different definitions of a good day. The original review saw one undifferentiated surface. A product thinker sees this:

Who do we prioritize? The shopper. Their checkout completion is the north star, they sit on the exact step the goal is made of, so their friction costs us orders directly. The operator matters (their promos drive demand) and ops matters (a botched order is a refund), but the plateau is a conversion problem, and conversion is the shopper’s journey. Build outward from them.

Step 3 — The Real Pains#

Map each reported item to the pain underneath it, add the pains nobody reported (found by looking at where the funnel actually leaks), and rank by impact on the goal:

The two biggest opportunities in the store, checkout abandonment and slow pages, were never on the original list, because nobody walked the store as a shopper trying to pay, or looked at where the funnel actually leaks. That’s the cost of collecting symptoms instead of walking journeys.

Step 4 — How We Win#

Fixing pains keeps the store competitive. It doesn’t make it the best. Winning is a separate bet, and the Storefront’s seam is the crossing from full cart to paid: the checkout, where one in five shoppers quit rather than re-type an address, and the exact step the goal is made of. So the candidate bet:

But it doesn’t get built on conviction. It’s a Rock and it’s the most attached-to idea in the room, so it gets earned first: confirm shoppers actually abandon over the typing and not the shipping price, confirm they’ll trust saved payment, and watch out, this exact store already had a “one-tap add-on” idea that tested badly (shoppers read it as pushy, refunds rose). A bet that survives that scrutiny is worth a lane of its own. A bet that skips it is just the loudest idea, and the loudest idea being wrong is how a quarter gets burned.

Step 5 — Quality Attributes and Where Attention Goes#

The dimensions this store is actually judged on, and how the current attention maps to them:

What we’re over-investing in: cosmetic polish on the pages you land on first, image sizes, menu icons, footer links, filter niceties. Visible, easy, satisfying, low-impact. It’s where attention pools because it’s where the reviewer’s eyes happened to be.

What we’re neglecting: the step where money is actually made (the checkout where 1 in 5 quit) and the shopper who doesn’t complain (the one who hits a slow page and leaves). These are the highest-impact areas in the store and they generate almost no feedback, because the people they hurt don’t file tickets. They leave. The absence of complaints was mistaken for the absence of problems, which is exactly how orders sit flat at 1,900 while everyone stays busy.

Step 6 — Goal, Metric, and the Prioritized Roadmap#

Goal: convert more of our existing traffic into paid orders; break the ~1,900/week plateau.

Primary metric: completed orders per week (and checkout completion rate, its last step). Supporting: page-load median (guards the top of the funnel), search success rate. Guardrail: as we speed checkout up, the return rate must not rise. Gaming check: we count paid orders, not carts created, so nobody wins by nudging shoppers into carts they never buy.

Before committing, the plan gets the sufficiency test: does it actually add up to the goal?

The six-item flat list, now a sequenced plan with a reason behind every rank:

The winning bet runs in its own lane, off this queue. Full one-tap checkout is a Rock, and a Rock never wins an impact-for-effort contest against sand and pebbles, so if it had to compete here it would lose every sprint forever. Instead it’s sequenced deliberately across sprints and earned first. Saved address & pay (item 1) is the incremental step toward it, and gets completion to 88%; the full one-tap bet is the path to 95%+.

Same raw material. The difference is entirely in the thinking: a goal to measure against, personas whose jobs are understood, pains ranked and sized before they’re trusted, a plan checked that it actually closes the gap, a bet earned before it’s built, and an order with a defensible reason at every line.

The Difference, in One Line#

The first review answered “what did I notice?” This one answered “who is this for, where is it weak, how does it become the best, and does our plan actually get us there?” Same store, same afternoon. One produced a to-do list that, fully shipped, would have left orders exactly where they’ve been for two halves. The other produced a plan that closes the gap.

That’s the bar. When you’re handed a product and asked how to make it the best, this is the answer we expect.

Back to: Product Sense  ·  Related: Diagnosis, Bets