I was chatting with Gary Cokins on LinkedIn last week about using Activity-Based Costing (ABC) for university financial management and I mentioned that by using traditional financial accounting methods a lot of costs could be missed, specifically cross-subsidization and overhead.
Cross-subsidization occurs where the costs of teaching subjects are covered by other schools, think elective subjects, and therefore hidden from the Dean or Head of School managing their own budget. This is particularly complicated when you think about the hundreds or thousands of subjects taught and the many-to-many relationships with all of the academic programs. The other big cross-subsidization is where teaching subsidizes internal research.
Overhead is another big one, and it’s growing, and the whole reason for creating ABC in the first place was to deal with the allocation of overhead using causal logic across the organization. Although Deans/Heads of Schools might not be able to directly control overhead allocation, and therefore they usually manage their own direct costs when it comes to budgeting, it is absolutely essential for institutional financial management to understand what this overhead supports and to make sure that margins calculated include ALL costs.
The obvious issue here is that financial decisions could be made based on inaccurate or incomplete data, which could make things worse!
The conversation also extended into our work with AI enhanced ABC models and how AI can be used to analyze our ABC models. So, it got me thinking, could I use our existing (although still a work in progress) solution to write this blog post?
Just to provide a bit of context, the demo model is built the same way as our typical Higher Education cost models, just using dummy data. The standard model will cover the entire institution and every academic program and subject instance (when, where and how it’s taught). It is built up using data from a range of source systems including:
- Financial Management – The entire GL for the whole institution, that we will also reconcile to.
- HR/Payroll – Every individual employed by the institution and what they do e.g. teaching, research, community support, administration etc. For teaching specific subjects we break it down into time for teaching, marking, student support, admin etc. – so it’s very detailed.
- Student Management – We don’t include individual students, but we do include every academic program and subject taken and the numbers and mix of students.
- Timetabling – who is teaching what and when.
- Facility Management – not essential but some institutions include it for analysis of room usage and spare capacity.
So, the model we are about to query has a lot of very detailed data it and because it uses causal allocation logic, it represents the actual operations of the institution and is like rocket fuel for AI engines. In this case we’re using Snowflake Cortex as the engine behind Pilbara Intelligence. Also, all of our client institutions have multiple years of these models so we can analyze trends over time as well.
The Pilbara Intelligence solution is an agentic AI, this means it has an AI SQL agent for querying the underlying database, so we are working with real numbers and perform all calculations using real numbers, we don’t want hallucinations with financials! Plus, it has the Large Language Model for natural language querying and results.
Let’s take it for a test drive!
Prompt: Can you review the Pilbara Insights university model and identify costs that would normally be hidden using traditional financial accounting and write a blog post on your results.
Uncovering Hidden Costs in Higher Education: What the Pilbara Insights University Model Reveals
Traditional financial accounting in higher education often obscures the true cost structure of academic operations, making it difficult for institutions to understand where money is actually being spent and which programs are truly profitable. The Pilbara Insights University model provides a fascinating case study in how modern activity-based costing and granular financial tracking can illuminate costs that would otherwise remain hidden in aggregate financial statements.
The Problem with Traditional University Accounting
Most universities rely on traditional financial accounting that aggregates costs into broad categories like “Academic Salaries,” “Administrative Expenses,” and “Facilities Costs.” While this approach satisfies regulatory reporting requirements, it provides little insight into:
- The true cost of delivering specific courses or programs
- How resources are actually allocated across different academic activities
- Which programs are subsidizing others
- The real cost impact of different staffing models
What the Insights University Model Reveals
After analyzing the comprehensive financial data from Insights University’s 2024 operations, several categories of previously hidden costs emerge:
1. Cross-Subsidization Patterns
The data reveals significant cross-subsidization between different course types and funding streams. For example:
- Research programs show substantial losses, with Research (Internal) generating only $0.62 in revenue for every dollar spent, and Research (External) generating $0.48 per dollar spent
- High-performing domestic teaching programs like certain Psychology and Medicine courses generate revenue-to-expense ratios of 3-6x, effectively subsidizing research activities
- Commercial and out-of-scope activities show mixed results, with some commercial services generating healthy profits (1.66x ratio) while others operate at significant losses
This level of detail would be completely invisible in traditional accounting, where research and teaching costs might be aggregated into broad “Academic Operations” categories.
2. Granular Staff Cost Allocation
Traditional accounting typically shows total salary expenses by department. The Insights model reveals:
- Precise allocation of permanent vs. casual academic staff across specific courses and programs
- Cross-faculty teaching arrangements where, for example, Psychology courses utilize staff from Health Sciences, Education, and Science & Technology faculties
- Hidden administrative overhead embedded in academic delivery, with detailed tracking of preparation time, administrative hours, and student engagement time
For instance, the Bachelor of Psychology program shows that permanent academic staff represent only 26-32% of total academic costs, with the remainder split between casual academics (19-20%) and professorial staff (both permanent and casual). This granular view enables much more sophisticated workforce planning and cost management.
3. Activity-Based Overhead Allocation
Rather than spreading overhead costs uniformly, the system categorizes expenses as:
- Direct costs tied to specific research centers or academic programs
- Faculty support costs that directly support academic delivery
- School direct costs for specific schools within faculties
- General overhead for institution-wide services
This reveals that traditional overhead allocation methods may significantly misrepresent the true cost of different programs. For example, the Faculty of Health & Medicine shows $54.5M in direct school costs versus only $6.8M in faculty support costs, indicating that most resources are deployed directly to academic delivery rather than administrative overhead.
4. Resource Utilization Inefficiencies
The detailed tracking of delivery hours, preparation time, and student engagement reveals hidden inefficiencies:
- Disproportionate preparation time for some courses relative to actual delivery hours
- Underutilized capacity in certain programs where delivery hours are low relative to allocated resources
- Cross-subsidization of low-enrollment programs by high-enrollment courses within the same faculty
5. Revenue Stream Dependencies
Traditional accounting might show total tuition revenue, but the Insights model reveals:
- Commonwealth Grant Scheme (CGS) dependency varies dramatically by program, with some courses receiving 2.8x their expenses in CGS funding while others receive none
- HECS revenue concentration in specific program types, creating vulnerability to policy changes
- International student revenue clustering in particular faculties, creating both opportunity and risk concentration
The Strategic Implications
These hidden costs have significant strategic implications for university management:
Resource Allocation Decisions
With visibility into true program profitability, universities can make more informed decisions about:
- Which programs to expand or contract
- How to price different course offerings
- Where to invest in additional capacity
Staffing Strategy
Understanding the true cost structure of permanent vs. casual academic staff enables:
- More strategic workforce planning
- Better negotiation with academic unions
- Optimization of teaching loads across faculties
Cross-Subsidization Management
Rather than allowing implicit subsidization, universities can:
- Make explicit decisions about which programs to support
- Set appropriate pricing for profitable programs
- Develop targeted strategies for loss-making but strategically important activities
Risk Management
Visibility into revenue stream dependencies enables:
- Better preparation for policy changes affecting government funding
- Diversification strategies for over-dependent programs
- Contingency planning for international student market volatility
The Technology Behind the Transparency
The Insights University model demonstrates how modern data architecture can transform financial transparency in higher education. The system integrates:
- General ledger data with detailed account classifications
- Academic product information including course structures and delivery modes
- Student enrollment and completion data for accurate cost allocation
- Staff time tracking for precise resource utilization measurement
This integration enables real-time visibility into costs that would traditionally only be visible through expensive, time-consuming activity-based costing studies.
Conclusion: The Future of University Financial Management
The Pilbara Insights University model represents a significant evolution in higher education financial management. By making hidden costs visible, it enables university leaders to:
- Make data-driven decisions about resource allocation
- Optimize the balance between teaching, research, and service activities
- Develop more sustainable financial models for the future
As universities face increasing financial pressure and accountability demands, this level of financial transparency will likely become not just advantageous, but essential for institutional survival and success.
The question for other institutions is not whether they can afford to implement such systems, but whether they can afford not to. In an era of constrained public funding and increasing competition, the universities that thrive will be those that truly understand their cost structures and can optimize accordingly.
The insights presented in this analysis are based on 2024 financial data from the Insights University model, encompassing detailed cost and revenue tracking across 14 faculties, 31 schools, and hundreds of individual course offerings.
Back to Blog Post
Well, that was very interesting, and only took a few minutes to do this analysis. I have to call out this particular nugget:
“This integration enables real-time visibility into costs that would traditionally only be visible through expensive, time-consuming activity-based costing studies.” – I’m presuming Cortex is referring to traditional, manual ABC models and not our much faster, more efficient way of building ABC models 😎
Another important note is that the intelligence is in the ABC model, and the AI agent is the new interface to provide management and executives with their own personal analyst that they can call on whenever they want. That said, it’s still very important to have a “human in the loop” to verify the output of the AI, because AI is not perfect, just like people are not perfect, so we still need to check their work.
The major advantage here is that as an Academic or Financial manager/executive, you can brainstorm issues directly with your personal analyst very quickly and at any time of the day or night.
So this was a successful experiment to use AI to not only perform the analysis but to also write about it. It was able to find examples of hidden cost and also discuss the strategic implications. The next step for the manager would be to dig into some of these specific areas, again using your AI personal analyst and when you have identified several scenarios hand it over to the human analysts to validate and provide the final report.