Turning around a failing medication-management product.

I was headhunted by a Director of Product I'd worked with before, for a complex technical product he didn't know how to hire for. When the leadership team grew concerned about the delivery of the medication-management (eMARs) pilot, I was the person they asked to audit it.

The audit found an outsourced build with oversized scope, incomplete design, weak acceptance criteria, no KPIs or data plan, no risk register, and safety concerns from the Quality team that hadn't been incorporated. I then took the product on myself.

Medication administered by non-clinical carers in people's homes is safety-critical by definition. A big-bang launch wasn't an option.

• Audited the build and rescoped it to a genuinely safe pilot.
• Set up a working group and steering committee, and went back to Quality on the real risks.
• Added instrumentation and reporting, so we could see what was happening, and user research, so we understood why.
• Piloted with five clients, with a paper backup, before scaling.

To protect the investment, I let the pilot go live on schedule with design I knew was weak: some medication data was captured unstructured, and scheduling didn't match how care is actually delivered. I put controls around both, so they were safe, but safe isn't the same as easy to undo. A poor interface can be improved later; the shape of your data is far harder to change, because every record and report that follows is built on it. Now I'll cut scope to hit a date, but I won't compromise on the data model.

Phase 1 reached 70 branches and 9,000 clients, saving £560k a year and around 25,000 branch hours. As the platform and company grew, it scaled to 123 branches and £7.3m of benefit, becoming the company's most successful product.