David StrathwickChief Product Officer

Chief Product Officer United States

Deciding what to build and what to switch off

Every promise our app makes is kept by someone in a lumber yard.

01

Choosing which customer problem to fund

Each year our product organisation can fund about twelve substantial pieces of work across eight teams. Requests outnumber that by roughly four to one, and most arrive described as features. I ask for each to be rewritten as a problem with a named customer and a named associate: the contractor waiting at the gate, the household customer told an item was available when it was not, the cashier re-keying a quote. Every proposal states the evidence behind it, usually yard visits, order data and call logs, together with the measure that would tell us to stop. Our finance team checks the value estimate and store operations confirms the labour it assumes. I keep a fifth of capacity unallocated for problems that surface mid-year, because a spring building season rarely follows the January plan.

02

Tools for the associates who keep the promise

In a home-improvement retailer, many customer features are only half built until an associate has a matching tool. An order-ahead button needs a pick list; an availability badge needs a count someone trusts; a delivery window needs a driver's route that reflects it. I treat store and yard associates as primary users with their own research, design and release schedule. Our designers test handheld screens with gloves on, in rain, at the far end of a lumber shed. We measure associate minutes per order as carefully as customer conversion. When a feature adds work to a shift, I ask store operations to agree the hours before launch rather than discover them later in overtime reports. Associates forgive an unfinished tool that improves every fortnight; they do not forgive one that adds a step without explanation.

03

Pickup and in-stock promises the building can support

Household pickup and contractor order-ahead now produce about 17 per cent of our sales. Both rest on an inventory count designed for replenishment, not for promising a particular board to a particular customer at seven tomorrow morning. Long lumber, sheet goods and bagged concrete move in ways a count cannot follow in real time. I would rather show a customer a narrower promise than a broader one we break. Our current work assigns an availability confidence to each item and location, using recent count accuracy, sales rate and the hours until pickup. Items below the threshold show an invitation to check with the store instead of a quantity. Fewer items look available, which our merchants dislike, but in the pilot regions cancellations at the counter fell from 6.1 to 3.4 per cent of pickup orders.

04

Loyalty and trade-account economics

A rewards programme is a pricing decision that marketing happens to operate. I review its economics with our chief financial officer every quarter: what the points cost, who earns them, how much of that spending we would have received anyway and how much unredeemed balance sits on the books. For trade customers the question changes. A framing contractor buys what the job requires and compares quotes; points rarely alter that choice. Reliable delivery, accurate invoices and dependable credit terms do. Since 2023 our larger trade accounts have received delivery credits and a dedicated order-desk line instead of points, while smaller contractors and household customers remain in the points programme at a lower earning rate. I keep the people who lose from each change visible, because a redesign that looks costless has usually not been examined closely enough.

05

Retiring features without stranding the people who use them

A retailer's app collects features faster than it removes them, and each one keeps consuming engineering time, associate training and customer attention. Every January we name the features we intend to retire alongside the work we intend to fund. A retirement follows the discipline of a launch: we identify who still uses the feature, what they will use instead and what we will do for anyone without a good replacement. When we retired the desktop quote builder in 2023, about 900 trade accounts still used it, some daily. We moved their saved quotes, gave each account a named order-desk contact for ninety days and kept a read-only version for a year. Thirty-one accounts asked for more time and received it. Removing features is how our teams make room for the work customers and associates need now.

A leadership conviction

A button on a contractor's phone is a commitment made on behalf of someone loading lumber at six in the morning. Before I fund a feature, I want to know who will carry it out, what they will stop doing to make room for it and how we will notice if the promise starts to fail. Retiring a feature deserves the same care as building one.

Start a conversation