Work · 2023–26

Cafepoint

Three years part-time at a London coffee equipment supplier while I did my degree. I started in credit control and ended up running the move to a new ERP system and the company’s weekly reporting.

Role
Business Analyst, then Digital Transformation Lead
When
Feb 2023 to Aug 2026, 15 to 20 hours a week during my degree
Tools
ERP, Power BI, Excel (VBA, Power Query)
  • £75,000ERP project I made the case for and ran
  • 650+customer accounts, about £8.2M a year
  • 95%of staff trained in 12 weeks
  • 9.1%average error on a 12-week cash forecast

The company

Cafepoint supplies, installs and services coffee machines, water dispensers and vending machines for offices in London and around the UK. I joined part-time in February 2023 to do credit control and analysis. Seven months later I was promoted to lead its digital transformation, and I did that until I finished my degree.

Getting paid

I rebuilt credit control around risk. Every customer got a risk band, and chasing late invoices became automatic and depended on the band. Customers went from paying in 52 days on average to 30, the collection rate reached 96% and the bad-debt provision fell by 35%.

The new ERP system

I wrote the three-year business case for a £75,000 ERP system and then ran the 15-week move for more than 60 users and 600 accounts.

  • I interviewed more than 20 staff to find out how the work was really done, and ran more than 15 workshops on requirements and training.
  • We rolled it out in phases, with trained staff in each team who helped everyone else.
  • 95% of staff were trained within 12 weeks and over 90% were using it in the first month.

Closing the books each month went from 12 days to 4.

Weekly reporting

For three years I wrote the weekly pack the leadership team ran the business from, built on more than 10 Power BI dashboards. Over that time service time fell 25% and delivery costs fell 18%. I also built a 12-week cash-flow forecast that flagged when actual numbers drifted from it. Its average error was 9.1%.

An AI project I said no to

I looked into using generative AI for customer service. The problem wasn’t the AI. Customer data was spread across systems that didn’t connect, and staff were wary of it. I wrote the memo recommending waiting until the data was sorted out.

People

I mentored five junior staff through the ERP move and afterwards, and gave the senior team a strategy briefing every month.