What I've learned finding value that was hiding in plain sight.

A business I worked with had a familiar problem: high cost per acquisition, and a journey that leaked people right after they'd converted.

Two different teams, two different scorecards - one on getting people in, one on keeping them once they'd arrived. Neither team was doing anything wrong. Nobody owned the join between the two.

Redesigning the acquisition journey and the onboarding experience together, as one continuous thing rather than two handoffs, delivered 40% growth in acquisition and a 50% reduction in cost per acquisition in the same programme.

Neither number moved because of a bigger budget. They moved because the join stopped leaking.

Where does your customer's experience quietly cross a boundary between two internal teams?

The budget was there. The business case looked attractive. Nobody would have questioned the decision.

One of the most commercially valuable recommendations I've made in my career was telling a business not to buy the enterprise platform they were already leaning towards.

Not because it was a bad platform. Because the organisation wasn't yet set up to use most of what it did, and was never likely to be.

A smaller, right-sized platform delivered what they actually needed - at an estimated saving of over £1m across two years.

Transformation isn't about recommending the biggest solution available. It's about recommending the one the business can actually absorb.

What's the last piece of technology your business bought that turned out bigger than what you were ready to use?

Years before generative AI was mainstream, I built a simple rule into how commercial decisions got made: no insight counted as reliable until it had been checked against at least two independent sources.

Internal data and experience. External insight from partners or agencies who saw the market differently. And then, as a third check - not the first - AI.

Used that way, AI's job isn't to generate the answer. It's to challenge the other two, or get challenged by them. Several long-held internal assumptions didn't survive that process. Some genuinely held up.

The organisations getting the most value from AI right now generally aren't the ones asking it for answers. They're the ones using it to stress-test the answers they already have.

If AI disagreed with your most confident assumption tomorrow, would you know which one to trust?

Severity gets attention because it's dramatic - something breaks, someone escalates, it gets fixed. Persistence is quieter. Nothing's ever bad enough on any single day to interrupt anyone's meeting. It just never quite gets better either.

I built a reporting framework around a rule most dashboards don't have: it wasn't how bad a metric was that decided who saw it next. It was how long it had stayed bad.

Clear within a week - stays with the team, no escalation needed. Still amber a month later - it moves up automatically, no one has to notice and decide to raise it. Still unresolved after a quarter - it reaches board-level visibility, whether or not anyone flagged it.

The system didn't need someone paying attention to catch the slow leaks. It was built to catch them on its own.

Most dashboards are good at surfacing the fire. Almost none of them are built to notice the thing that's been quietly smouldering for months because it never got bad enough on any single day to matter.

What's been sitting at "fine, but not quite right" in your business for longer than anyone's willing to admit?

It's assuming every market is ready for the same solution, at the same time.

Head office tends to measure success as standardisation - one platform, one process, one rollout plan. Local teams tend to measure success as solving the problem they actually have today. Those aren't always the same thing, and the gap between them is where most programmes quietly stall.

I once inherited a multi-million-pound transformation programme that had stalled for exactly this reason - technically sound, but built around an ideal operating model that didn't match how any single market actually worked. Rebuilding it wasn't a technology fix. It was building enough local flexibility into the governance that markets could get there at different speeds, without losing the shared standard that mattered.

The programmes that succeed usually aren't the ones with the strongest central mandate. They're the ones with governance strong enough to allow variation without losing direction.

Where in your business is "the plan" further ahead than the people expected to deliver it?

Switching provider used to mean picking up the phone. Every quote, every enquiry, went through a sales team who had no way to scale beyond how many calls they could physically take.

Nobody was doing their job badly. There simply wasn't a self-serve option in the industry at all - so a manual, sales-led process was the only process anyone had.

Building the industry's first online quoting tool - tested and refined against real customer behaviour, then extended all the way through to full online contract acceptance - didn't just add a channel. It removed a ceiling that had been quietly capping growth the whole time.

It also didn't work as a digital-only build. Sales and digital had been operating separately - digital shipping changes, sales working leads, neither comparing notes on what the other was seeing. Moving both teams onto a shared weekly cadence, reviewing the same numbers together, was what let the tool actually get optimised rather than just launched and left.

89% more enquiries. A 127% higher conversion rate. Within a year, 60% of all switch enquiries were being handled digitally, without a phone call.

The bottleneck was never the sales team. It was the absence of any alternative to them - and the absence of a rhythm that let sales and digital improve it together once it existed.

Where in your business is a great team quietly acting as a ceiling on growth, simply because there's no other way to get through the door?