Three engagements, without the names

The organisation had removed its intermediate data stores without delivering the replacements. The entire steering setup sat in a local file fed by shared workbooks. The deadline was fixed: six executive reports for a set date, five weeks later. Direct data access was refused, and the official solution cost around USD 20,000 — close to CHF 16,500 — and would have taken three to four months to install.

A deliberate, documented workaround built inside the client’s own environment, then replaced by a proper, sanctioned data route as soon as one existed. And thirty-minute daily check-ins for a month, so nobody discovered a gap on deadline day.

Deadline met. The best measure of success was a silence: “nobody was complaining about it, which is really positive for me” — against an early prototype openly criticised three months before.

A budget tracking file built five years earlier by someone who had since left, protected by a password nobody holds — not even the finance analyst who inherited it, and who detests it. No row can be deleted.

Rebuilding the file’s logic live, in front of the team: where the figures come from, what is duplicated, how currencies are handled. Then reversing the roles — the workbook is only for entry, while calculation and control move into the steering setup, fed automatically by the finance system.

One input template per programme, a central configuration file, a consolidated report and a verification page. The client’s response: “I’m in total agreement, and I think my team is coming to that realisation too.”

Five regions, two instances of the procurement system that do not talk to each other, therefore two supplier masters. Invoicing spread across several versions of the finance system, including a local one in China. Contracts scattered across three locations, with inconsistent mandatory fields. Data maturity measured at level 2 of 5: no documented procurement data strategy, almost no reporting to stakeholders, no quality control beyond what the systems do natively, and no integration between the HR system and the ordering system — everything was done by hand. Against that, quantified savings targets: JPY 2.5 billion of P&L impact, roughly CHF 13 million.

Two trajectories set out in the room, without hedging: standardise every system over three to five years, or connect the existing sources and get results in six months across a fifth of the scope. The second was chosen. Then three prerequisites made contractual conditions — a workstation, a named account, a licence tied to that account — with one sentence that never moved: nothing starts until those three are settled. And a quality threshold acknowledged rather than promised: around 80% accuracy, enough to decide on.

A four-pillar maturity assessment delivered on one page, then seven procurement indicators put into production: savings, supplier base rationalisation, share of spend under control, catalogues, purchasing card. Three dashboards in service — steering, performance across three levels of reading, and transactional visibility over both instances. And an audit page open to all, letting the client check the quality of its own data itself.