When I present modernization investment requests to my board, I need more than scary statistics about technical debt. I need a credible ROI story with realistic assumptions.
Here’s how I built the business case that got our $8M modernization program approved.
The ROI framework that works:
IBM’s research shows 15-35% annual infrastructure savings post-modernization. Microsoft Azure studies document 228-304% ROI over three years. But these are vendor numbers - boards are skeptical of vendor claims.
How I built credible internal projections:
1. Infrastructure cost baseline
- Current annual infrastructure spend: $4.2M
- Maintenance contracts for legacy systems: $1.8M
- Specialized talent premium: $600K (legacy skill sets command 40% premium)
- Total legacy-attributable cost: $6.6M/year
2. Conservative savings assumptions
- Year 1: 10% infrastructure reduction ($420K) - accounts for parallel running costs
- Year 2: 25% reduction ($1.05M) - migration mostly complete
- Year 3: 35% reduction ($1.47M) - full optimization achieved
- Maintenance contract retirement: $1.2M by Year 3
- Talent normalization: $400K by Year 3
3. Investment requirements
- Year 1: $4M (heaviest migration lift)
- Year 2: $2.5M (continued migration + optimization)
- Year 3: $1.5M (completion + stabilization)
- Total investment: $8M
4. Net present value calculation
Using 10% discount rate:
- Total savings over 5 years (extending beyond investment period): $18.2M
- NPV: $6.8M positive
- Simple payback: 2.8 years
- IRR: 31%
What convinced the board:
-
Risk framing - “75% of organizations will face moderate or high severity technical debt by 2026. We’re choosing our timeline, not having it chosen for us.”
-
Competitive positioning - “Our competitors modernized 18 months ago. They’re shipping features 40% faster. Every month we delay, we fall further behind.”
-
Talent market - “COBOL developers are retiring. In 5 years, we won’t be able to hire anyone who understands our core systems at any price.”
-
Security exposure - “Legacy systems can’t implement modern security controls. Our cyber insurance premiums are rising 15% annually because of this exposure.”
The assumptions I explicitly called out:
- This assumes no major scope expansion during migration
- Parallel running periods may extend if testing reveals issues
- Savings depend on actually decommissioning legacy systems, not running both indefinitely
Boards appreciate honesty about risks. It builds credibility for your projections.
What ROI frameworks have others used successfully?
Michelle, your financial modeling is solid. Let me add a few refinements from the finance perspective that can strengthen your next board presentation.
The payback period challenge:
Your 2.8-year payback is reasonable, but boards in different sectors have different thresholds:
- SaaS companies: Often need 18-month payback for CapEx decisions
- Financial services: More comfortable with 3-5 year horizons given regulatory stability
- Manufacturing: May accept 5+ years for infrastructure investments
Know your board’s typical hurdle rate and frame accordingly.
Sensitivity analysis they’ll ask for:
Before my board presentation, I prepare scenarios:
- Base case (your numbers): 31% IRR, 2.8 year payback
- Conservative case (20% lower savings, 30% higher costs): What’s the floor?
- Optimistic case (include productivity gains): What’s the ceiling?
- Break-even case: How bad can it get before NPV goes negative?
The break-even analysis is key. If you can show the project still works even with significant setbacks, confidence increases.
Costs you might be underestimating:
- Training and change management: Often 10-15% of project budget
- Productivity dip during transition: Engineers are slower while learning new systems
- Extended parallel running: Plan for 50% longer than expected
- Hidden integration costs: APIs to third-party systems that nobody documented
The 31% IRR positioning:
For a technology investment, 31% IRR is strong. But compare it to your company’s cost of capital:
- If WACC is 12%, the 19-point spread is excellent
- If WACC is 8% (low-rate environment), the spread is even better
Frame IRR relative to other investment opportunities the board is considering.
Michelle, the ROI model is compelling, but let me share what actually happens when you try to achieve those savings.
The implementation reality:
I’ve led two major modernization programs in financial services. Both had beautiful ROI projections. Here’s what we learned:
What delivered as projected:
- Maintenance contract retirement (once systems are truly decommissioned)
- Cloud infrastructure costs (actually came in 15% better than projected)
- Talent market normalization (this one takes longer - Year 4 in our case)
What took longer than projected:
- Parallel running: We planned 6 months, actual was 14 months. Regulators required extended validation periods we hadn’t anticipated.
- Decommissioning: Nobody wants to turn off the legacy system. It took executive mandate to finally pull the plug.
- Knowledge transfer: Legacy experts left before documentation was complete. We paid consultant rates to bring some back temporarily.
What wasn’t in the original model:
- Emergency legacy patches during migration: $400K unplanned
- Performance tuning on new platform: Original estimates assumed it would “just work”
- Third-party integration surprises: Three vendors required complete re-integration
How we still hit acceptable ROI:
-
Phased decommissioning - Instead of “big bang” legacy shutdown, we migrated by capability. Each decommissioned module delivered immediate savings.
-
Quick wins first - We prioritized migrations that would show savings within 6 months. Early wins built momentum and credibility.
-
Aggressive change management - When engineers tried to build “bridges” to keep legacy alive, we shut it down. Bridges become permanent.
-
Regular ROI reviews - Quarterly check-ins with finance to track actual vs projected. Transparency built trust.
The 228-304% ROI is achievable, but it requires disciplined execution and willingness to make hard decommissioning decisions.
Michelle, the ROI model focuses on cost savings, but I’d argue the real value is in velocity improvements. Here’s the product case.
The velocity multiplier:
Your model shows 40% faster feature delivery post-modernization. Let me translate that into business impact:
If we ship features 40% faster, we:
- Win more competitive deals (faster time to market)
- Reduce churn from delayed functionality
- Enter new markets sooner
- Respond to regulatory changes faster
How I’d quantify this for your board:
-
Deal velocity impact: If your sales cycle includes “waiting for features,” calculate deal value × days saved × conversion rate improvement
-
Churn reduction: Identify churned customers who cited “missing features” or “slow response to requests.” What’s that ARR?
-
Market expansion: What markets are you blocked from entering due to legacy constraints? What’s the TAM?
A case study from my experience:
At a previous company, we couldn’t enter the European market because our legacy systems couldn’t handle GDPR requirements. The modernization investment paid for itself entirely through the EU expansion revenue - before counting any cost savings.
The board pitch I’d add to your model:
“The cost savings alone justify this investment with 31% IRR. But the real value is strategic: with modern systems, we can ship features 40% faster, enter markets we’re currently blocked from, and respond to competitive threats in weeks instead of quarters.”
One caution:
Make sure product has a seat at the modernization planning table. I’ve seen modernization projects that achieved technical goals but missed product velocity improvements because they didn’t prioritize the right capabilities.
The 40% velocity improvement only materializes if you modernize the systems that are blocking product development, not just the systems that are expensive to maintain.