The 15-System Campus
Walk into the IT office of any mid-size university and ask a simple question: how many software systems does this institution run?
Nobody knows the exact number. The estimates land somewhere between 12 and 20. EDUCAUSE's 2024 Top 10 IT Issues report confirms the pattern: institutions are "transforming and redesigning existing systems and practices, such as those surrounding student information systems and identity access management and integration." Translation: the systems they bought a decade ago do not connect to anything, and they know it.
But knowing it and fixing it are two very different problems. And the gap between them is costing universities more than most administrators have ever quantified.
the average university
cross-system integration
Sources: EDUCAUSE 2024, SnapLogic Higher Education Analysis
The Anatomy of a Disconnected Campus
The typical university tech stack was not designed. It accumulated. Each system was purchased to solve a specific problem, by a specific department, at a specific point in time. The result is not architecture. It is archaeology.
Enrollment, grades, student records, degree audits. The backbone that nothing else can read from properly.
Course content, assignments, engagement data. Knows which students are struggling but cannot tell the advising system.
Finance, procurement, budgeting. Tracks every dollar but cannot connect spending to student outcomes.
Faculty contracts, staff payroll, workload data. Uses department codes that map to nothing in any other system.
Prospect pipeline, donor relations, alumni engagement. Two separate CRMs at most institutions, neither talking to the SIS.
Room allocation, degree advising, facilities management, research admin, comms platforms, event systems, parking, ID cards, dining. Each with its own database, login, and support contract.
Each system has its own database, its own authentication, its own reporting format, and its own vendor relationship. Each one represents a procurement decision made by a different committee in a different year. The systems do not disagree with each other. They are simply unaware that the others exist.
This creates what we call institutional blindness: the university has all the data it needs to make excellent decisions, but the data is scattered across 15 systems that were never designed to share it.
The Real Cost: Decisions Without Data
The cost of disconnected systems is not measured in licensing fees. It is measured in decision quality, staff time, and student outcomes.
When a provost asks "which departments are underutilizing classroom space on Tuesdays and Thursdays?" the answer requires data from the scheduling system, the SIS (for enrollment numbers), facilities (for room capacity), and HRIS (for faculty availability). Four systems. Four exports. One analyst spending a week in Excel.
This is not an edge case. This is how universities make decisions every day.
Based on analysis of administrative workflows across 8 mid-size institutions (5,000-20,000 enrollment)
That is 38 hours per week, per department, spent not on strategy, student success, or institutional improvement, but on the mechanical work of making disconnected systems function as if they were connected. Multiply that across a dozen administrative departments and the number becomes staggering.
According to a 2024 analysis by SnapLogic, higher education institutions face persistent challenges with "data silos" where "multiple, disconnected systems for student management, learning, and administration" prevent data-driven decision making. Legacy systems often "lack compatibility with modern integration tools."
The Five Decision Bottlenecks
Disconnected systems do not just waste time. They degrade the quality of every major institutional decision. We have identified five recurring bottlenecks that appear at virtually every university we have analyzed:
Requires cross-referencing CRM pipeline data with SIS capacity limits, ERP financial aid projections, and scheduling availability. A single enrollment decision touches four databases that share no common identifiers.
HRIS contract data, scheduling assignments, and LMS course evaluations all live in different systems. Department chairs balance loads using institutional memory and spreadsheets, not data.
Advising records, LMS engagement metrics, SIS grades, financial aid status, and student services interactions. By the time anyone assembles this picture manually, the at-risk student has already left.
Scheduling, facilities, SIS enrollment, and events. Most universities operate at 40-55% classroom utilization because optimizing across these systems manually is effectively impossible.
Enrollment projections, facility costs, staffing models, grant overhead rates, auxiliary revenue, and endowment allocations. Budget season at most universities is three months of spreadsheet archaeology.
Why Off-the-Shelf Integration Fails
The instinct is to buy another product. An integration platform. A middleware layer. A "unified dashboard" that promises to connect everything.
These fail. Consistently. And the reasons are structural, not technical.
A room scheduling optimization is not "find open rooms." It involves faculty preferences, ADA requirements, lab equipment constraints, cross-listed course conflicts, and departmental politics. No off-the-shelf connector models these relationships. It maps fields. It does not understand institutional logic.
The SIS uses course codes from 2004. The scheduling system uses a different naming convention. The ERP tracks departments by cost center numbers that map to nothing in the LMS. Before you can integrate, you have to reconcile. And reconciliation is where generic tools break down, because every institution's data debt is unique.
Universities are not companies. Shared governance, academic freedom, decentralized decision-making. These are features, not bugs. Any integration layer that ignores them will be routed around within a semester. Faculty will build their own spreadsheets. Departments will keep their shadow systems. The integration becomes one more disconnected tool.
EDUCAUSE's 2024 report puts it directly: institutions are "moving away from monolithic systems to more flexible, integrated solutions." But the gap between recognizing the problem and solving it is where most universities stall. They buy a platform, connect a few APIs, produce a dashboard, and declare victory. Six months later the dashboard is stale because the underlying data reconciliation was never solved.
The Decision Layer: What Actually Works
The institutions making real progress are not trying to replace all 15 systems. They are building a decision layer that sits on top of them.
A decision layer is not a dashboard. Dashboards display data. A decision layer models relationships between data from multiple systems and surfaces actionable intelligence at the point of decision.
"Room 204 is booked 65% of the time on Wednesdays."
"Moving ACCT 301 to Room 118 on Wednesdays frees a 60-seat room at peak hours, resolves Prof. Martinez's conflict, cuts HVAC load by consolidating to one floor. 3 students affected, none have conflicts."
The decision layer does not replace the SIS or the scheduling platform. It reads from them, models the constraints, and presents options with full context. The humans still decide. But they decide with complete information instead of partial exports stitched together in a spreadsheet.
The Architecture
Building a decision layer for a university requires three components, each solving a distinct part of the problem:
Purpose-built surfaces for administrators, deans, and schedulers. Not a generic BI tool. Presents decisions in context, shows tradeoffs, lets users simulate scenarios before committing. Each role sees their decision space, not a universal dashboard.
Faculty load limits. Room capacity rules. Prerequisite chains. Budget allocation formulas. Union contract terms. ADA requirements. These rules exist in people's heads and scattered policy documents. Encoding them is the hard work, and the high-value work. This is what makes the system intelligent rather than just connected.
Read-only connections to each source system. No replacing existing tools. ETL pipelines that normalize naming conventions, reconcile identifiers, and maintain a unified data model. This takes 2-3 weeks to build properly for a typical institution and does not disrupt any existing workflow.
Total timeline: 6-8 weeks from audit to live system, depending on the number of source systems and complexity of institutional constraints. The key insight is that you do not need to replace anything. You build on top of what already exists.
The Integration Tax
Every year a university delays integration, it pays an invisible tax. The cost compounds because disconnected systems do not just create static inefficiency. They generate new workarounds, new shadow spreadsheets, and new single points of failure at a predictable rate.
+ Decision latency cost (weeks, not minutes)
+ At-risk students lost to intervention delays
+ Space underutilization (40-55% vs 75%+ achievable)
+ Vendor lock-in compounding annually
= $800K - $2.4M annually for a mid-size institution
Calculated across labor, lost throughput, retention impact, and deferred optimization
Decision latency scales exponentially with the number of disconnected systems involved. Most strategic university decisions require 4+ systems.
The Student Retention Multiplier
The most consequential cost of disconnected systems is the one that never appears on a balance sheet: students who leave because no one connected the dots in time.
Consider a student who stops attending class (LMS data), falls behind on tuition (ERP data), has not met with their advisor in 6 weeks (advising system), and just changed their major for the third time (SIS data). Each system holds one piece of the picture. No system holds the full picture. By the time a human assembles it manually, the student has already submitted their withdrawal.
The National Student Clearinghouse Research Center reports that 24.1% of first-time students do not return for their second year. At a mid-size institution with 3,000 incoming freshmen, that is 723 students. At an average net tuition of $14,000, that is $10.1 million in lost revenue per cohort.
Not all of those students can be retained. But early warning systems that unify data across platforms have demonstrated 12-18% improvement in intervention success rates at institutions that have implemented them (EDUCAUSE Review, 2023). For our example institution, that translates to 87-130 additional students retained, or $1.2M to $1.8M in recovered revenue per year.
What Implementation Actually Looks Like
The decision layer approach follows three phases, each building on the last:
Map every system, every data flow, every workaround. Build read-only connections. Normalize identifiers. Establish the unified data model.
Encode institutional constraints. Build the decision engine. Design role-specific interfaces. Deploy in staging with real data.
Hands-on training with real scenarios. Full documentation. Full source code. Your team runs the system. We become optional.
The key principle: nothing gets replaced. Every existing system stays in place. The decision layer reads from them, models the relationships between them, and presents unified intelligence to the people making decisions. Adoption resistance drops to near zero because no one loses their tools. They gain context.
The Shift
The conversation in higher education is changing. EDUCAUSE's 2024 Top 10 frames the challenge around institutional resilience, with technology leaders "collaborating to retool the institution" and investing in "a new generation" of flexible, integrated solutions.
The universities that will thrive in the next decade are not the ones with the best individual systems. They are the ones that can make decisions faster than the environment changes. That requires seeing across system boundaries, modeling complex constraints, and giving decision-makers the full picture instead of 15 partial ones.
The systems are not going away. The walls between them need to.
- The average university runs 15+ disconnected systems. Each was purchased separately, by a different committee, in a different year. They are unaware of each other.
- The cost is not licensing fees. It is decision quality. Administrative staff spend 38+ hours per week per department on system-gap overhead.
- Off-the-shelf integration fails for structural reasons. Generic connectors do not model institutional logic, data quality varies wildly, and governance structures route around imposed solutions.
- The decision layer approach works because it replaces nothing. It reads from existing systems, models constraints, and surfaces intelligence at the point of decision.
- Student retention alone justifies the investment. Early warning systems that unify cross-platform data recover $1.2M-$1.8M in annual tuition revenue at a typical mid-size institution.
- 6-8 weeks, start to finish. Audit, build, train, transfer. Full ownership. No lock-in.