Chief Executive Officer / United States
Decisions that cost Marrick something, and what they bought
- Stores in nine states
- 1,140
- Most in towns of fewer than 10,000 people, September 2026
- Net sales
- US$1.94bn
- Fiscal 2025, the year to 31 January 2026
- Independent pharmacies hosted in Marrick stores
- 31
- Trading in September 2026; 33 opened since January 2025
- Store robberies a year
- 47 to 14
- Fiscal 2022 to fiscal 2025, after the two-person rule
Decisions & consequences
01Hosting independent pharmacists where the counters had closed
In 2024 we mapped every Marrick town against the nearest open pharmacy and found that in 214 of our 1,140 locations a customer had no pharmacy within a twenty-minute drive. Pharmacies had been closing across Missouri, Kansas and Oklahoma since 2022, and our store managers were fielding questions about insulin and inhalers that they could not answer.
The stakes were partly commercial and partly about what kind of company we were in those towns. Stonebeck's managing partner, Anders Pruitt, proposed Marrick-operated pharmacies in sixty stores. He saw a traffic driver and a health business that could raise the company's value at exit. Our chief financial officer, Bettina Lorrimer, costed the proposal at US$34 million of capital and about US$19 million of cumulative operating losses over four years, with returns that depended on reimbursement rates set through pharmacy benefit managers.
I recommended hosting independent pharmacist-owners instead: space beside the health aisle, a fit-out allowance of up to US$150,000, nominal rent for three years and cost-recovery rent after that, with no share of dispensing income. Anders argued that we were giving away the customer relationship and the prescription data, and he was right that we were. We also gave up control of pharmacy hours and accepted that a tenant might fail inside our store. The board approved a pilot of six sites with a review after twelve months.
We recruited pharmacists through state pharmacy associations and by asking owners in nearby towns whether they wanted a second site. Recruitment, not capital, proved the constraint. We widened the over-the-counter range in host stores.
By September 2026 thirty-three hosted pharmacies had opened and thirty-one were trading. A typical site fills about 610 prescriptions a week at twelve months. Host stores' sales are 5.9% above matched stores and their customer counts 7.4% higher. Marrick's outlay has been US$4.6 million, and towns without a pharmacy within twenty minutes have fallen from 214 to 186.
Health access now sits in our site-selection model, and the board sees the town count before the sales figures. What I would do differently is the money. Our second site closed after eleven months because its owner ran out of cash waiting for reimbursements. The working-capital bridge we added in 2026 should have been in the first offer.
Two people on duty from opening to close
In fiscal 2022, 61% of Marrick's stores closed at 10pm with a single employee on duty for the last two hours of trading. We recorded forty-seven robberies that year, nine involving a weapon, and our assistant store managers, who worked most of those closing shifts, were leaving at a rate of 78% a year. Shrink had risen from 2.2% to 3.1% of sales in two years.
The stakes were our people's safety first and the stability of store management second: a company that cannot keep assistant managers cannot develop store managers.
Stonebeck's operating partner and our chief financial officer preferred a shrink programme: locked cases and contracted guards in the 120 stores with the highest losses, at about US$6 million a year, keeping the 10pm close for its sales. I proposed instead that every store have at least two people on duty from opening to close, and that 640 stores close at 9pm to help fund the second person. The disagreement was not about whether safety mattered. It was about whether safety had to pay for itself through lower shrink. I argued that it should not, and the board accepted a net labour cost of about US$11.8 million a year and the loss of roughly 0.6% of sales in the affected stores.
We introduced the rule from April 2023, region by region, with store managers choosing who covered each close. In parallel our asset protection team broke a year of shrink into its causes and found that about 38% came from receiving, markdown and damage-recording errors. We fixed those first and limited locked cases to forty stores where external theft was concentrated.
By fiscal 2025, the year to January 2026, robberies had fallen to fourteen and assistant manager turnover to 51%. Shrink was 2.3% of sales, and workers' compensation claims arising from violence had fallen by more than half.
Safety staffing is now set before the labour budget rather than inside it. What I would do differently is consult earlier on hours. In towns where a factory shift ended at 10pm, the earlier close cut off customers who had shopped with us for years. We restored the later hour in twenty-three stores, with three people on the close, after complaints that store managers had predicted and that I had not asked them about.
Cancelling part of a holiday buy when import costs moved
In spring 2025, as US import tariffs changed in a series of announcements, our merchandising team estimated that landed costs on about 31% of what we bought would rise by roughly US$46 million a year. Most of the exposure sat in general merchandise, seasonal goods and own-label household items sourced in Asia, and much of our holiday merchandise had already been ordered for delivery from August.
Our customers' budgets were the first stake. A Marrick shopper typically spends under US$20 a visit, and many households in our towns live on fixed incomes or hourly wages. Our lenders were the second: after the 2024 recapitalisation, a large shortfall in margin would have narrowed our covenant headroom and drawn attention we did not want.
Stonebeck's directors favoured passing the higher costs through across the range to protect EBITDA. Our chief merchant, Lowell Haskett, wanted to hold every price and absorb the cost, arguing that customers would remember who raised prices first. I recommended a third course: hold prices on about 300 everyday household items, reprice discretionary and seasonal goods, cancel part of the holiday buy and move some own-label sourcing. We gave up seasonal sales, paid cancellation charges and accepted an EBITDA figure below budget.
We cancelled US$19 million of holiday orders, paying US$2.2 million in charges, and cut the seasonal assortment by 28%. We moved forty-one own-label items, mainly paper goods, cleaning products and food storage, to suppliers in the United States and Mexico; in several cases they cost more than the old price from Asia but less than the new landed cost. We repriced about 2,100 discretionary items, some of them twice as tariff rates moved again.
In fiscal 2025 comparable sales rose 3.1%, with consumables up 3.8% and customer transactions up 2.1%. Seasonal sales fell 11%. Gross margin fell sixty basis points to 30.8%, and adjusted EBITDA was US$141 million against a budget of US$152 million. Covenant headroom never fell below 25%.
We now review sourcing concentration by country every quarter and write cancellation terms into seasonal contracts. I would act faster next time. We waited six weeks for clarity that never came, and an earlier cancellation would have cost less than half as much.
A smaller dividend recapitalisation, with my own interest declared
In March 2024, with Marrick's net debt at 3.1 times EBITDA and lenders keen to lend, Stonebeck proposed a dividend recapitalisation of US$280 million, funded by a US$160 million incremental term loan and a sale-leaseback of the 186 stores we owned. The fund was in its sixth year, and returning capital was a reasonable aim.
The stakes ran beyond the debt ratio. Pro forma net debt would have risen to about 4.3 times EBITDA, and the sale-leaseback would have created leases of fifteen to twenty years in towns where our own projections showed population falling. Our 410-store remodel programme, budgeted at US$120 million over three years, depended on cash that the larger structure would have consumed.
I recommended a dividend of US$120 million from the term loan alone, no sale-leaseback of stores and a ring-fence around remodel capital. Anders Pruitt and his Stonebeck colleagues on the board disagreed, arguing that real estate was not our business and that remodels could be phased. The two Marrick family directors split. I also had a personal interest: my management equity would have received about US$4.4 million from the larger dividend and about US$1.9 million from the one I proposed.
I declared that interest in writing, asked Constance Quell, our independent chair, to have the non-management directors retain their own financial adviser and abstained from the vote. My team built an estimate of how many owned stores we would want to close or relocate over twenty years and what carrying those leases would cost, and the adviser tested it.
The board approved a US$120 million dividend financed by the term loan and declined the store sale-leaseback. Against my recommendation, it also approved a sale-leaseback of our Salina distribution centre for US$58 million, with the proceeds committed to remodels. Net debt rose to 3.9 times EBITDA and had fallen to 3.6 times by January 2026. By September 2026 we had remodelled 236 stores, whose sales rose 6.8% in their first year.
Recapitalisation papers at Marrick now include a section on store-network flexibility. My mistake was one of timing. I brought the lease analysis after Stonebeck had begun talking to lenders, which made it look like resistance rather than evidence. I should have built it a year earlier.