Why IR office staffing has not kept up with data demand

In our interviews with 50+ IR and IE professionals across 19 states, the same story came up at institution after institution. More people want data, they want it sooner, and the IR office stays the same size. The team absorbs the difference by working faster. There is a limit to how fast one person can pull a file, reconcile a cohort definition, and send back a sourced answer, and most of the offices we spoke with have reached it.

This post covers both sides of that mismatch: what our interview data and national staffing surveys say about IR capacity, and where the extra demand is coming from.

Ad hoc requests already take a large share of the week. Across our interviews they consumed 40% to 63% of team capacity on average: 47.6% for small teams of one to three people, 40% for medium teams of four to seven, and 62.9% for large teams of eight or more. In the peak months of September, January, and May, utilization climbed to between 75% and 90%. Half of the institutions we interviewed run IR with one to three people.

The work is slower than it looks from outside the office. 73.5% of teams named vague or unclear requests as a bottleneck, 52.9% said requests routinely need extensive follow-up before work can start, and clarification typically runs two to five rounds. From request to delivery, the cycle took 3 to 14 days. If you need these figures by team size for a budget case, the IR capacity benchmarks post has the full tables.

Repeats make it worse. The same retention figure, the same FTE breakdown, and the same peer comparison arrive from different stakeholders at different times. Each one gets handled as a new request even though the underlying data has not changed.

Our interview sample is not the only evidence. The Association for Institutional Research (AIR) surveys IR and IE offices every three years. The 2024 AIR National Survey, based on 552 complete or partial responses from U.S. degree-granting institutions, found an average office of 3.8 FTE. Public four-year institutions average 5.5 FTE; public two-year and private nonprofit four-year institutions average 2.9. 31% of offices have one FTE or fewer, and most said they need at least two more staff.

The trend points the wrong way. Among the 147 offices that answered the 2018, 2021, and 2024 surveys, average staffing fell from 4.0 FTE to 3.7. AIR also found that office size rises with enrollment, but more slowly: public four-year institutions enroll nearly three times as many students as the other sectors and carry only about twice the IR staff.

AIR's 2023 survey on IR/IE office capacity asked about workload directly. 53% of respondents said their staff size is not enough for the work they have now, and 42% were dissatisfied with the size of that workload. In offices of one FTE or less, 73% said they often feel overwhelmed, against 60% in larger offices. One respondent described the office-of-one survival strategy plainly: "I've essentially limited my workload to maintain a 1.0 FTE IR office."

IR was built to support senior leaders and to file mandatory reports. AIR's Statement of Aspirational Practice for Institutional Research argues that IR should serve a broader range of decision makers, including faculty, staff, and students. Many offices took on that wider audience without the headcount to match it, so every new dean, program director, and committee chair who learns to ask for data adds to one queue.

The compliance floor stayed where it was. 47.1% of the teams we interviewed named accreditation, IPEDS, and state reporting pressure as a constraint. Those deadlines cannot move, so every ad hoc request competes with work that cannot slip. The IR team's year shows how federal and accreditation deadlines land in the same months as the heaviest stakeholder traffic.

The data is also harder to reach than it should be. 82.4% of institutions described data spread across disconnected systems (SIS, HR, LMS, finance) that someone has to merge by hand, and 35.3% were working around legacy systems. Dashboards were supposed to absorb some of the demand, yet 32.4% reported dashboards that go unused while people keep emailing IR for the same numbers.

Hiring another analyst is the obvious lever, and most institutions try it first. It is also the slowest and most expensive one. IR analysts with the right mix of domain and technical skill are hard to recruit, take months to onboard, and cost a salary far larger than most software an IR office buys. Public institutions add hiring freezes and budget approval cycles. 23.5% of the teams we interviewed named budget constraints directly, and new tools or positions often need a detailed cost-benefit case that the same stretched team has to write.

A new hire also adds capacity in a straight line while repeat questions keep multiplying. A third analyst covers a few more retention pulls and FTE breakdowns, then hits the same manual bottleneck as the first two.

Headcount is only half the argument. Our Institutional Intelligence Gap research found that 85% of institutions depend on a single person for the institutional knowledge behind their numbers, and 55% sit in what we call the "Large Gap" tier, where the definitions that run the place exist in one head and nowhere else. A new hire does not close that gap. Once trained, they become the new single point of failure. Capacity and knowledge concentration are two angles on the same structural problem, and neither one is solved by headcount alone.

Most of the queue is retrieval and assembly, not insight. The analyst who already knows the answer still has to pull the data, clean it, calculate it, and format it. A model connected to your data and your definitions can do that part in seconds, and it can ask the clarifying questions (which cohort, which term, which definition of retention) before the request reaches a person.

The model has to be built for higher education. A general chatbot cannot tell whether "enrolled student" means headcount or FTE at your institution, and it cannot query your SIS. Clema's data request management handles intake as a conversation, drafts answers from your institutional data with sources shown on every figure, and routes anything new to an analyst with the scope already pinned down. The analyst reviews and sends instead of pulling and assembling.

This changes the cost curve: answering the tenth copy of a retention question costs a review, not a rebuild. If you are not ready for a new tool, the data request best practices post covers the intake and automation rules that help any team, whatever software it runs.