Up to half of enterprise data team capacity in Australia and Singapore is consumed by single-use data work that will never be reused, new research from Quest Software has found. The study describes the pattern as a 'hidden tax' on capacity at the moment organisations most need it for AI initiatives.
The research, conducted with Corinium Global Intelligence, surveyed 154 Chief Data and Analytics Officers and senior data leaders in February and March 2026. All respondents came from organisations with more than 500 employees and at least US$100 million in annual revenue.
Nearly every respondent (97 per cent) said their organisation delivers data products. Yet only 25 per cent operate a structured data products programme with standardised, repeatable workflows. Three-quarters operate in an unstructured, ad-hoc way.
Among organisations without structured programmes, 45 per cent said between a quarter and half of data team effort goes to non-reusable projects. A further 17 per cent said more than half of team capacity is absorbed by one-off work. Some 62 per cent expect faster delivery if projects were built as reusable data products.
"C-level leadership wants to move faster in the age of AI, and boards are backing AI initiatives, but the underlying data foundations in many organisations have not caught up," said Susan Laine, Chief Data Technologist at Quest Software. "The result is a hidden tax on capacity at exactly the time they need it most."
The findings also point to a compliance dimension. Compliance is absorbing up to 30 per cent of IT team budgets, and governance reviews were cited by 59 per cent of ad-hoc operators as the primary bottleneck slowing data project delivery. That figure was significantly higher than for organisations with structured programmes.
Respondents cited skills, culture and competing priorities above budget as the biggest barriers to structured delivery, suggesting the transition is as much a change management challenge as a technical one.
Brendan Mathias, Director of Analytics and Data Science at Cochlear, contributed views to the report on the duplication that domain-centric products can create. Within a domain, attribution rules are agreed and reporting is trusted, he said, but enterprise decision forums can disagree on reported numbers when there is no shared definition.
Forty-four per cent of respondents have active AI or generative AI projects underway that depend on foundational data preparation, while 45 per cent are experimenting with pilots not yet in production.