From post-it maths to a single source of truth

The Client

An innovation centre based within a large publicly funded Australian university. The centre runs a portfolio of cohort-based founder programs for both students and the wider community, alongside a regular calendar of events. Like most innovation centres of its kind, it runs largely on external grants, which makes accurate and timely reporting a condition of survival, not a nice-to-have.

The challenge

Each year the team produced an annual impact report and a steady stream of grant acquittals, drawing on a wide range of reportable metrics across its programs: founder and team counts, coaching hours, satisfaction scores, participation rates, funds raised, event attendance and enrolments.

The numbers were sound. The process behind them was not. Data originated in a scatter of disconnected systems, from form tools and event platforms to calendars, university systems and manual head counts in the room. All of it was then copied, recalculated and re-keyed by hand into a master Excel workbook with dozens of tabs and no linking formulas. Totals were worked out with a calculator and what one team member cheerfully called "post-it maths".

Reporting that was meant to happen regularly slipped because of the effort involved. The same data was entered twice, once during a program and again for reporting, multiplying both the workload and the risk of error. Overlapping spreadsheets lived all over the shared drive with no single source of truth, and the knowledge of how each number was produced lived in people's heads rather than any documented process.

The approach

The engagement opened with a facilitated discovery workshop. Rather than start with tools, the team mapped the full data journey for every reportable number: where each figure originated, where it was aggregated and who was responsible. That surfaced the real problem, which was not a shortage of data but the manual handling between systems.

Three principles guided everything that followed. Build in parallel and never disrupt the live process until the new system was proven. Chase quick wins first, because momentum and voluntary adoption beat forced change in a small, time-poor team. And design every field backwards from the end report, so each piece of data traced to a metric that actually appeared in an acquittal.

The solution

The team consolidated reporting onto a single Airtable base as the central hub, built around a lean set of linked tables covering people, programs, enrolments, events, surveys and mentoring.

Event registrations were connected directly into Airtable so attendee data flows in automatically, with no middleware and no manual exports. The application and survey forms were rebuilt on a tool that writes to several linked tables from a single submission, replacing a fragile workaround that had relied on extra staging tables and automations. Adding a new program form became a quick task rather than a technical project, and a match-or-create rule keeps records clean as the same founder moves between programs.

Live dashboards replaced the manual report tabs, so headline figures can be read on demand. Underpinning it all is a documented metric playbook that records, for every number, its definition, source, collection method, calculation and owner, so anyone can reproduce or verify a figure without tribal knowledge.

The results

The centre moved from a fragile, manual process to a single automated hub fed from its source systems. Event data that once meant exporting and re-keying now flows in automatically. Double-handling was designed out rather than tolerated. Figures that previously took weeks to assemble can be pulled live from dashboards, and the playbook means continuity no longer depends on any one person remembering how a number was built.

Just as importantly, the solution respected the team's reality. It is cost-conscious, suited to mixed technical confidence and can still export to Excel where grant bodies require it. The legacy spreadsheets stayed running in parallel until the new system had proven itself across a full program cycle.

The data was never the problem. The handling between systems was. Fixing that gave the team back its time and its confidence in the numbers.