Research Report
Organizations are spending more on data every year and trusting it less than the spending implies.
Enterprise decision making runs on data that the people using it do not fully believe.
Across 300 senior leaders at US companies with 500 or more employees, 37% say they are confident in the data behind their team's biggest decisions. Of every 100 points of weight behind those decisions, 26 sit on experience and judgment rather than on anything systematic. Most organizations are working on it: 69% have started consolidating their data into a single source of truth, while 11% have finished. That unfinished middle is where the confidence gap lives, and it is now the constraint on everything being built on top of it.
Three numbers
are confident in the data behind their biggest decisions
have a single source of truth in place across the organization
say their data is ready to support AI
Confidence in enterprise data sits low and flat. Leaders were asked to split 100 points across what their important decisions actually rested on this year, and the answers show where the shortfall gets absorbed.
Mean points allocated per respondent. Totals 100.
Board wanted a churn number in 48 hours. We gave them a number. I still couldn't tell you if it was right. Not our finest hour.
Director, Energy / utilities, 500–2,499 employeesWhat it means
Judgment is not the failure mode here. Leaders lean on it because the alternative is slower and no more trustworthy. The cost shows up later, in decisions nobody can retrace and outcomes nobody can attribute.
Two problems tie for the top of the list, and they are the same problem seen from opposite ends. Data spread across systems is why answers take a week, and a week is how long it takes because the data is spread across systems.
Share rating 4 or 5. Respondents rated each item independently.
Share ranking each item in their top three. Respondents chose three of seven, so shares total more than 100.
Good one. Had to pick which product line to pull marketing spend from. The attribution data was six weeks stale so we basically went with the one the CFO liked least.
Director, Technology / software, 2,500–9,999 employeesWhat it means
More dashboards and more analysts both rank below consolidation as the fix, at 47% and 26%. Leaders have worked out that the constraint sits underneath the reporting layer, which is a harder and less visible thing to fund.
Consolidation is close to universal as an ambition and rare as a completed state. The organizations that are underway report materially higher confidence, and the payoff arrives in trust before it arrives in speed.
Underway combines the three Q3.1 stages where work has started.
Multi-select. Respondents chose as many as applied, so shares total more than 100.
Half the org trusts it and half doesn't. The half that doesn't keeps a shadow spreadsheet and honestly some of their numbers are better. No easy answer to that one.
Director, Financial services, 10,000+ employeesMostly it's that nobody can articulate the return in a way that survives a budget conversation. We all know it matters, we can't put a number on it.
Director, Retail / consumer, 500–2,499 employeesWhat it means
51% have a central warehouse and 94% of those still run on spreadsheets every week. Consolidation projects deliver a platform long before they change how the work gets done, and the gap between those two moments is where executive patience runs out.
Rolling out AI is a top-three priority for 61% of leaders. 23% describe their data as ready to support it. Where AI is already running, it is doing work that tolerates being wrong.
Multi-select. Shares total more than 100.
Share ranking each item in their top three. Shares total more than 100.
Honestly it's not about the data, it's about explainability. I need to be able to tell my board why a decision was made. We're a long way from that.
Director, Insurance, 500–2,499 employeesIt's always the next thing. Right now everything's going into AI initiatives, which is a bit backwards given the state of the data. That's the honest version.
Director, Energy / utilities, 2,500–9,999 employeesWhat it means
69% use AI for drafting and summarizing, work where a person checks the output before it matters. 10% have handed it multi-step work with limited review. The ceiling on that second number is trust in the data, and trust in the data is what 23% currently report.
Where this goes next
The organizations reporting the highest confidence are not the ones with the most reporting tools. They are the ones that finished the unglamorous work of getting their data into one place, and they report 45% confidence against 20% among those who have not started.
Read the methodologyMethodology
Respondents qualified on three criteria: an organization of 500 or more employees, a level of Director or above, and personal involvement in at least one of how their team uses data and reporting, analytics or business intelligence decisions, or technology purchasing. Consultants and vendors selling data or analytics services were excluded. No quotas were set, so company size, seniority, function, and industry fall naturally within those criteria.
Percentages are calculated from unique respondents against the full base of 300 unless a smaller base is stated. Single-select questions are rounded using the largest remainder method and total exactly 100. Multi-select and top-three ranking questions total more than 100 because respondents chose several options. Scale questions report the share rating 4 or 5 rather than a mean. Industry and function were appended from panel firmographics rather than asked in the interview, and because no quotas were set on either, they are reported as directional.