From the returned parcel to the product page

Mandate

Build one customer product across web, app, stores and the after-sale service, and turn trade-ins and repairs into something customers could use online.

Outcomes

• Cut click-and-collect cancellations from 11.2 per cent to 2.4 per cent while collection grew from 17 to 24 per cent of online orders.
• Launched bilingual resale across Canada in July 2023; 38,000 items resold for C$4.5 million in 2025–26.
• Reduced the online return rate for new clothing from 21 per cent to 16.5 per cent through fit guidance built on return reasons and garment measurements.
• Redesigned the paid care membership; card disputes fell from 1,900 to 140 a year and membership recovered to 166,000.
• Released a new checkout in January 2025 that 22 of 24 disabled testers could complete, against 9 before.

Decisions in context

I joined Sablemere in March 2021 as its first chief product officer, after three years in which the company had been one of my clients at Brisevent. Revenue was C$560 million, online sales were growing quickly after the store closures of 2020, and the website, app, store tools and a nascent repair and trade-in service were run by four separate teams. My mandate from the chief executive was to make them one product and to make its promises dependable. The order of work followed the complaints: availability first, then the membership, then resale, then the storefront replacement that exposed our accessibility failures. I lead 148 people, work in shared teams with engineering and present to the board six times a year. By January 2026 revenue was C$742 million, with 31 per cent online and resale established in every province.

Reason for the move

Current role. Completing the US resale launch; open to executive conversations from the second half of 2027 and to one independent directorship with the board's consent.

Mandate

Turn a store-inventory tool into a product suite on which retailers could run click-and-collect.

Outcomes

• Launched click-and-collect orchestration in 2019; 61 clients used it by the end of 2020.
• Supported a ninefold rise in clients' collection orders in spring 2020 without a platform-wide outage.
• Raised net revenue retention from 104 to 112 per cent.
• Brought the store-associate app to WCAG 2.1 Level AA in 2020 after a client's blind associate showed us she could not use it.

Decisions in context

As vice-president of product I was responsible for everything Brisevent sold, with a team of 52 across Montréal and Toronto. The company had grown on a single product, the store-counting app I had built as a director, and clients wanted to run whole fulfilment journeys on it. We launched click-and-collect orchestration in 2019, deliberately requiring clients to reach a count-accuracy threshold before switching it on for a store. That condition irritated several clients and slowed our sales. In March 2020 it proved its worth: when stores closed, collection volumes rose ninefold within weeks, and clients who had counted first could keep their promises. I presented the product plan to our board every quarter and learnt to explain a roadmap to directors who cared more about retention than features. Net revenue retention rose from 104 to 112 per cent. I left to run the whole customer product at one retailer.

Reason for the move

Wanted to own a retailer's whole promise to its customers rather than supply one part of it; Sablemere, a client, offered its first chief product officer role.

Mandate

Build a product that told retailers how far they could trust their store stock figures, and what to count to improve them.

Outcomes

• Built a risk-ranked cycle-count app used in about 1,900 stores across 31 retailers by 2017.
• Launched an availability confidence score that clients used to decide which stock figures to show online.
• Raised average store record accuracy at early clients from 71 to 92 per cent within a year of adoption.
• Cut clients' click-and-collect cancellations by an average of 58 per cent in the year after adoption.

Decisions in context

I moved to Montréal in June 2014 to work on the problem that had stopped my ship-from-store pilot in Toronto: store stock records nobody trusted. Brisevent was then a small company with one inventory product and a dozen clients. With a team of 14 I built an app that told associates which shelves to count each morning, ranked by the likelihood and cost of a wrong figure, and a confidence score that clients could consult before showing a store's stock online. By 2017 the app ran in about 1,900 stores across 31 retailers, and early clients had raised record accuracy from 71 to 92 per cent. Most of our users were associates on hand-held devices, so I spent more time in stockrooms than in our own office. I also recovered the French I had lost since school, because half our clients and most of our associates' feedback arrived in it.

Reason for the move

Promoted to vice-president of product.

Mandate

Launch the chain's first transactional website and grow it into a channel the stores would support rather than resent.

Outcomes

• Launched the online store in English and French in 2010.
• Grew online sales from C$18 million in 2010–11 to C$96 million in 2013–14.
• Reduced online clothing returns from 29 to 24 per cent with garment measurements and fit notes.
• Stopped a 2012 ship-from-store pilot after 18 per cent of orders were cancelled, and documented why for the board.

Decisions in context

Delacourt & Finch hired me in September 2009 to help launch the first online store of a family-owned department-store chain with 22 stores. We went live in 2010 in English and French, because seven stores were in Québec, and online sales grew from C$18 million in the first year to C$96 million in 2013–14. I became head of online store product in 2011, with a team that grew to 26. The fit notes and garment measurements we added reduced clothing returns from 29 to 24 per cent, a figure I had been waiting to improve since my catalogue days. The defining episode was the 2012 ship-from-store pilot, cancelled after 18 per cent of orders failed because stock was not where the system said. I wrote the paper that stopped it and explained the cause to the family board. During these years I also completed a part-time graduate diploma in human–computer interaction.

Reason for the move

Wanted to fix store-inventory accuracy at its source, the problem that had stopped the ship-from-store pilot.

Mandate

Replace the warehouse management system and make stock records reliable enough for retailers to order against.

Outcomes

• Replaced the warehouse management system across both distribution centres in 2007.
• Raised location-level record accuracy from 91 to 99.2 per cent through cycle counting.
• Improved order fill rate for retailers from 93 to 97.5 per cent.
• Built a dealer stock-check portal used by more than 1,200 retailers by 2009.

Decisions in context

I joined Quillhaven in August 2004 as a systems analyst at its Mississauga distribution centre, which with a second centre in Calgary supplied bicycles, skis and camping equipment to about 1,900 independent shops. The company's warehouse management system was failing, and its stock records were accurate enough to reorder against but not to promise from. I became warehouse systems lead in 2006 and ran the C$6.8 million replacement across both centres, live in 2007. We introduced daily cycle counting by location, and record accuracy rose from 91 to 99.2 per cent, while the fill rate on retailers' orders rose from 93 to 97.5 per cent. My favourite work was a small stock-check portal that let a shop owner see what we held before promising it to a customer. More than 1,200 retailers used it. It convinced me that I wanted to build for the person at the end of the chain.

Reason for the move

Wanted to design for shoppers rather than warehouses, and joined a department store's new online team.

Mandate

Reduce the time and cost of processing returns and improve slotting in the distribution centre.

Outcomes

• Cut the time from receipt of a return to refund from nine days to four.
• Introduced triage of returns into resaleable, repairable and liquidation streams.
• Showed that 38 per cent of clothing returns were for fit, leading to brand-specific size charts in the 2003 catalogue.

Decisions in context

My first job after graduating was as an industrial engineer at Assiniford, a Winnipeg catalogue retailer that still sold clothing and household goods by printed catalogue and telephone to customers across the Prairies and the North. I was posted to the returns centre, which received about 1.1 million parcels a year. I opened hundreds of them in my first months and timed every step from the loading dock to the refund. Introducing triage into resaleable, repairable and liquidation streams, and moving refunds earlier in the sequence, cut the cycle from nine days to four. My analysis showed that 38 per cent of clothing returns were for fit, largely because one size chart served every brand, and the 2003 catalogue printed brand-specific charts for the first time. The business was shrinking as customers moved online, and I left to learn the systems that decide what a company can promise.

Reason for the move

The catalogue business was shrinking, and I wanted to work on the systems that decide what stock a business can promise.