The 12-Week Challenge
Chapter 1: The Ultimatum
Maya Chen arrived at the CIO’s office seventeen minutes early, which gave her exactly enough time to regret accepting the 8 AM meeting invitation titled simply “Strategic Planning Discussion.”
In her eight years at Riverside University, she’d learned that meetings with innocuous titles were the dangerous ones. The meetings that actually said what they were about, like “Q3 Budget Review” or “Incident Post-Mortem,” those were fine. But “Strategic Planning Discussion” with the CIO, CFO, and VP of Business Operations? That was the academic IT equivalent of “we need to talk.”
She clutched her coffee like a talisman and waited.
At 8:00 AM precisely, the CIO’s assistant ushered her into the corner office where Robert Harrison, Riverside’s CIO for the past three years, sat flanked by CFO Patricia Winters and VP of Business Operations James Chen (no relation, though Maya had stopped clarifying this years ago).
“Maya, thanks for coming,” Harrison said, gesturing to a chair across from the executive tribunal. “We’ll get right to it. We’ve been conducting a comprehensive review of our enterprise application portfolio, and we’ve made a strategic decision regarding our ERP platform.”
Maya’s coffee suddenly tasted like ash.
Harrison slid a glossy presentation deck across the table. The cover featured a sunset image, subtle, with the title “Riverside University: Cloud ERP Transformation Initiative.”
“We’re migrating to SaaS,” Harrison continued. “Workday for Finance and HR, Salesforce for CRM, and we’re evaluating Oracle Cloud and Anthology for the student system. The board wants us to be cloud-native within eighteen months.”
Maya flipped through the deck. Slide 3: “Legacy System Challenges.” Slide 5: “Total Cost of Ownership Comparison.” Slide 8: “Implementation Timeline.” Slide 12: “Change Management Strategy.”
It was a masterpiece of selective data presentation. She recognized the consulting firm’s watermark in the footer; McKinsey had been on campus for three months, apparently building the case for burning down everything her team had spent a decade maintaining.
“This TCO analysis,” Maya said carefully, tapping Slide 5. “It shows PeopleSoft costing us $4.2 million annually versus $2.8 million for the SaaS solutions.”
“Correct,” Patricia Winters said. “Plus, the SaaS model eliminates our infrastructure overhead, reduces our admin headcount requirements, and provides automatic updates. No more PeopleTools upgrades eating six months of your team’s time.”
Maya took a breath. She’d been an IT manager long enough to know when executives had already made up their minds. But she’d also been a PeopleSoft administrator long enough to spot a flawed cost model from across a conference table.
“Can I ask about the methodology?” Maya said. “How did the consultants calculate our current PeopleSoft costs?”
Harrison glanced at Winters, who pulled up a spreadsheet on her tablet. “Hardware and infrastructure, licensing, support contracts, admin salaries, upgrade projects, and integration maintenance. All documented in the appendix.”
“And the SaaS costs?”
“Subscription fees based on user count, implementation services, change management, and training,” Winters recited. “Standard pricing from the vendors.”
Maya nodded slowly. “What about data migration costs for the SaaS implementation?”
“Included in implementation services,” Winters said.
“Customization recreation costs? We have 847 custom PeopleCode programs, 203 custom pages, and 156 integration brokers. How much does it cost to rebuild that functionality in Workday and Salesforce?”
A pause.
“The vendors assured us their out-of-the-box functionality covers 80% of requirements,” James Chen offered. “We’ll streamline the rest. Too much customization is part of the problem.”
Maya felt her jaw tighten. “Those customizations exist because our business processes exist. Student financial aid packaging, faculty load balancing, research grant accounting; these aren’t vanity features. They’re how we operate.”
“Which is why we need change management,” Harrison said. “This is an opportunity to modernize our business processes, not just our technology.”
Translation: we’re going to force-fit the university into vendor-defined processes.
Maya looked at the timeline on Slide 8. Eighteen months to replace three PeopleSoft pillars that had been configured, customized, and integrated over fifteen years. She’d seen this movie before at other institutions. It never ended well.
“Robert,” Maya said, using his first name deliberately. “I need to be direct with you. This analysis is fundamentally flawed.”
The room temperature dropped several degrees.
“Flawed how?” Harrison’s voice carried an edge.
Maya stood and walked to the whiteboard. “May I?”
Harrison gestured permission.
She drew two columns: “PeopleSoft Costs” and “SaaS Costs.”
“The consultants counted everything for PeopleSoft—every infrastructure cost, every admin salary, every upgrade hour. That’s fair. But for SaaS, they only counted subscription and implementation. Let me show you what they missed.”
She began writing under the SaaS column:
Migration Costs (Year 0-2):
Data migration and validation: $800K-1.2M
Customization gap analysis and recreation: $1.5M-2.5M
Integration rebuilding (47 systems): $600K-900K
Business process reengineering: $400K-700K
Organizational change management: $300K-500K
Dual-run period (6-12 months): $250K-400K
Hidden Ongoing Costs:
Premium support tier (required for higher ed): +$180K/year
Integration platform (Workato/Boomi): $120K/year
Lost productivity during transition: $500K-800K (one-time)
Institutional knowledge loss: unquantified
Re-training existing staff: $200K
Consultant dependency (first 3 years): $400K/year
Subscription inflation (historically 7-12% annual): compounding
“That’s conservative,” Maya said. “And I haven’t counted the opportunity cost of your entire IT department being consumed by implementation for two years while other initiatives stall.”
Patricia Winters was taking notes. James Chen looked uncomfortable. Harrison’s expression was unreadable.
“Now let me show you what modernized PeopleSoft costs,” Maya continued.
She drew a third column: “PeopleSoft Modernized.”
One-Time Modernization Investment:
Cloud infrastructure migration (GCP/AWS): $200K
Infrastructure-as-code implementation: $80K
CI/CD pipeline development: $60K
Observability platform: $40K
Team upskilling (training/certifications): $50K
Automation framework: $70K
New Steady-State Costs:
Cloud infrastructure: $480K/year (vs $720K on-prem)
Reduced admin overhead (automation): -$200K/year
Managed database services: $180K/year
Modern observability tools: $45K/year
Remaining support/licensing: $650K/year (existing)
“Total modernized cost: approximately $1.2 million annually by year two,” Maya said. “That’s 57% less than the consultants’ SaaS projection, and 71% less than they claim we’re spending now.”
She capped the marker. “And we keep our institutional knowledge, our customizations, our integrations, and our team. Plus we’d be done in months, not years.”
Harrison leaned back. “That’s quite a claim, Maya. You’re telling me we can cut our PeopleSoft costs by 71% in a few months?”
“No,” Maya said. “I’m telling you the consultants inflated our current costs by counting every possible expense, then deflated the SaaS costs by ignoring migration reality. Real PeopleSoft modernization versus real SaaS migration? We win on cost, risk, and timeline.”
“The board wants cloud-native,” Winters said. “PeopleSoft isn’t cloud-native.”
“PeopleSoft on Google Cloud Platform with Kubernetes orchestration, infrastructure-as-code, automated scaling, and cloud-native observability isn’t cloud-native?” Maya countered. “We’re not talking about lifting and shifting our data center to AWS. We’re talking about re-architecting PeopleSoft using modern DevOps practices. The application platform is irrelevant if the operational model is modern.”
Harrison studied her. “You really believe you can do this?”
“I know we can,” Maya said. “But I need time to prove it.”
“Eighteen months?” Harrison asked.
“Twelve weeks.”
The room went silent.
“Twelve weeks?” James Chen said. “To modernize an entire ERP platform?”
“To prove the concept,” Maya clarified. “Twelve weeks to migrate one environment to cloud infrastructure, implement infrastructure-as-code, build CI/CD pipelines, establish modern observability, and demonstrate the cost model. Twelve weeks to show you that PeopleSoft modernization is faster, cheaper, and lower risk than SaaS migration.”
Harrison exchanged glances with Winters and Chen.
“And if you can’t prove it in twelve weeks?” Harrison asked.
“Then we proceed with your SaaS plan,” Maya said. “No resistance, no sabotage. I’ll personally lead the transition and help you find homes for my team.”
“And if you do prove it?”
“Then we commit to full PeopleSoft modernization instead of migration. We make it the strategic direction, not a stopgap. And we fund my team’s transformation properly.”
Harrison picked up the McKinsey deck and dropped it in his desk drawer. “You understand what you’re promising? In twelve weeks, I need to present to the board. I need concrete evidence, including cost numbers, performance metrics, automation demonstrations, and the works. Not theory. Proof.”
“I understand.”
“And your team? Are they on board with this? You’re asking traditional PeopleSoft admins to become DevOps engineers in three months.”
Maya thought of Jake, her senior DBA who still wrote Oracle PL/SQL by hand. Of Tom, her application server admin, who had seventeen years of tribal knowledge and zero documentation. Of Sarah, her junior developer, who was brilliant but had never touched Git.
“They’ll rise to it,” Maya said, with more confidence than she felt. “Or we’ll find out who belongs in a modern IT organization.”
Harrison stood and extended his hand. “Twelve weeks, Maya. Starting Monday. I want weekly status reports and a final presentation in week twelve. Make this count.”
Maya shook his hand, feeling the weight of what she’d just committed to. Twelve weeks to transform her team, her platform, and her career.
She walked back to her office in the IT building, her mind already racing through task lists and architecture diagrams. She needed to call a team meeting. She needed to assess their current state honestly, with no illusions or excuses. She needed to figure out how to teach a team of administrators who’d spent careers clicking through PIA how to write Python and Terraform.
But first, she needed to build the real TCO model. The one she’d sketched on Harrison’s whiteboard was directionally correct but approximate. If she was going to stake her credibility on these numbers, she needed them to be bulletproof.
Maya opened her laptop and created a new spreadsheet: “PeopleSoft vs SaaS: True Total Cost of Ownership.”