Resources

Guides for leaders deciding what to do with a legacy system.

Practical, vendor-neutral answers to the questions that come up before any modernization decision.

Abstract illustration: a vine wrapping a grey column, new growth in teal

Patterns

The strangler fig pattern: replacing a legacy system one workflow at a time

How the strangler fig pattern replaces a legacy system one workflow at a time, why it beats a big-bang rewrite, and where shared data and jobs break it.

Abstract illustration: two systems separated by a purple translation layer

Patterns

The anti-corruption layer: keeping the old system's quirks out of the new one

What Eric Evans's anti-corruption layer is, shown on a legacy invoicing example, why a parallel rebuild needs one, and when it is scaffolding or permanent.

Abstract illustration: two databases exchanging data with a reconciliation check

Patterns

Data migration vs synchronization: planning the cutover that decides success

Why data cutover needs its own workstream: mappings, backfill, sync, write ownership, reconciliation and rollback, plus dual-write vs CDC trade-offs.

Abstract illustration: two parallel lanes of requests compared side by side

Patterns

Parallel run and shadow traffic: proving a rebuild before customers touch it

How to prove a rebuilt legacy workflow before customers use it: replay vs shadow traffic, side effects, what to compare and reading a proof report.

Abstract illustration: a code path forking into toggled routes

Patterns

Branch by abstraction and feature toggles inside a legacy codebase

How to replace the parts of a legacy system a proxy can't reach: branch by abstraction, cohort toggles, a 1% to 100% release path and cleaning up after.

Abstract illustration: a bar chart with a rising trend line

Market data

Legacy software in numbers: what the global data says

Verified figures on IT maintenance spend, upkeep hours, Java 8, .NET Framework and COBOL persistence and skills gaps, plus how to set your own baseline.

Abstract illustration: a grid of systems with some highlighted

Market data

The US legacy problem: federal, financial and mid-market systems by the numbers

Verified US legacy system figures: GAO federal findings, bank and insurer core data, the mid-market gap, and rules that turn modernization into a deadline.

Abstract illustration: a path that drops away, with a smaller bridge across the gap

Market data

Why modernization projects miss their targets, and what the rest do differently

What the failure statistics for IT and modernization projects actually say, where the 70% myth comes from, and a checklist to score your own plan.

Abstract illustration: a timeline with a frozen block and a short scoped block

Market data

The cost of a code freeze: what a frozen roadmap really costs

What a long code freeze costs in delayed revenue, wasted capacity and attrition, with a simple model to price one freeze week and a worked example.

Abstract illustration: three options side by side, the third selected

Decision guide

Build, buy or renew: a decision framework for your core system

Rebuild, buy a SaaS replacement or renew incrementally? Compare the routes on cost, time, risk and knowledge, then score your system on four questions.

Abstract illustration: a clock face with a growing cost arc and bars

Decision guide

How to put a price on waiting: a cost-of-delay calculation a CFO can defend

A step-by-step cost-of-delay calculation for a legacy system: worked example, ranges, and how to add maintenance and risk without double counting.

Abstract illustration: two people carrying a system others cannot reach

Decision guide

Key-person risk: when two people hold the whole system in their heads

Why key-person risk stays off the risk register, how to measure it with the bus factor, and a 90-day plan to recover knowledge from the running system.

Abstract illustration: one large block beside a row of equal monthly blocks

Decision guide

Modernization as a subscription or project: which fits how you fund technology?

Fixed price, time and materials or subscription for legacy modernization: who carries scope risk, what boards ask about capex and opex, and exit terms.

Abstract illustration: a checklist with some items ticked

Decision guide

Twelve questions to ask any modernization vendor before you sign

Twelve questions finance, risk and engineering should put to a modernization vendor, the red flags in the answers and what a credible pilot looks like.

Abstract illustration: a live traffic waveform turning into a written specification

Engineering

Your production system is the specification: rules from observed behavior

Code reading and interviews each miss half the rules. Derive testable requirements from your running system, with evidence, stable IDs and known gaps.

Abstract illustration: a staircase of versions beside a curved alternative path

Engineering

Rails upgrade or Rails rebuild? A guide for teams stuck on Rails 4, 5 or 6

What an in-place Rails upgrade really costs, current Rails and Ruby end-of-life dates, and a four-factor table for choosing upgrade or incremental rebuild.

Abstract illustration: a build board with agent lanes and a human gate

Engineering

AI agents in legacy modernization: what they do well and where people stay

What AI agents do well in legacy modernization, the gates people must own, what the research says about AI code, and why automation still needs approval.

Abstract illustration: a routing switch with a lever to return traffic

Engineering

Rollback is a routing change: designing releases you can undo

How to design a legacy release you can undo: cohorts, acceptance criteria per step, expand/contract for data, a rehearsed rollback and a transition log.

Abstract illustration: a ledger of controls with checks and an approval stamp

Governance

Explaining a modernization program to your auditor

How SOC 2, ISO 27001, SOX and DORA change controls map to a legacy modernization program, with a sample control narrative from capture to release.

Abstract illustration: document pages linked to work items

Governance

Living documentation: why generated docs rot and what keeps them true

Why docs generated from code rot, how linking pages to work items keeps them true, and three measures of documentation health for legacy systems.

Which priority is your system holding up?

Tell us about it on a 30-minute call. We'll suggest a first scope and what a technical review would need to confirm.

Book a 30-minute call