ANDREW JOHNSTON | Private equity’s next returns are hiding in the data it already owns

Andrew Johnston, cofounder and partner at Swiss-South African process intelligence company Verdant Data, argues that private equity investors are underusing the transactional data that portfolio companies already generate. Finance, sales and operations systems log every order, change, invoice and payment with a timestamp, and together these records show where cash stalls, where staff time goes and where small failures accumulate into material losses. The article contends this evidence rarely feeds into value-creation plans.

The argument is framed by a tougher returns environment. Johnston writes that South African private equity can no longer rely on quick exits, cheap financing or multiple expansion, leaving operational improvement as the lever funds control. He cites S&P Global Market Intelligence's 2026 survey of global PE fund managers, in which 72% ranked operational improvement as the top value-creation lever, and the 2025 Southern African Venture Capital and Private Equity Association industry survey, produced with EY-Parthenon, which found regional firms expecting operational improvement to deliver 47% of near-term exit returns. S&P Global Market Intelligence's Kevin Zacharuk is quoted saying firms are "constrained by fragmented data and limited visibility into the metrics that matter most."

Johnston's method is process mining, which reconstructs how work actually flows from system trails rather than from procedure manuals. He illustrates with a hypothetical fund that has acquired a distribution business with R5bn in annual turnover. Analysis of the order-to-cash process could show that many orders are amended after submission, creating rework, credit notes and invoice corrections, while an average debtor-days figure masks a long tail of invoices older than 90 days. Because each step is timestamped and valued, Johnston argues these problems can be priced, and the output should be a costed plan of the five or 10 initiatives most likely to move earnings or working capital, with owners and supporting evidence. He warns that automating a broken process only repeats errors faster, and recommends fixing point-of-entry controls before automation, with funding released stage by stage.

Johnston also argues the data remains useful after the deal closes, turning a static value-creation plan into a live scoreboard tracking cash conversion, cycle times, rework and automation against acquisition targets. He says the analysis can be applied during due diligence, where a quick assessment takes about 10 days, during ownership to prioritise interventions, and at exit to document measured improvements for the next buyer. A back-test on a portfolio company's historical data, he suggests, is the easiest way for a fund to validate the approach before relying on it in a live transaction.