Practice · Migrations

The selection is easy.
The cutover is not.

ERP, cloud, mainframe exit, POS and data center migrations, evaluated by someone who ran the selection and then had to answer for the cutover on Monday morning.

$130MLargest ERP program
7Platform migrations led
14KPrice changes/day at cutover
1,000+Stores consolidated

Why This One Is Different

The selection is easy.
The cutover is not.

Most migration advice stops at the signature. The vendor is chosen, the contract is signed, and the hard part — data conversion, parallel running, the weekend the orders have to keep flowing — is handed to whoever is left holding it. That is where migrations fail, and it is the part no market quadrant scores.

This practice is led by someone who has been accountable on both sides of that line: running the selection as the advisor, and then owning the cutover as the CIO who has to answer for it on Monday. A $130M global ERP program, a mainframe retirement, a post-merger POS consolidation across 1,000+ stores, and a cloud cutover that held real-time inventory at 14,000 price changes a day.

That history shapes the advice in one specific way: we score implementation risk as seriously as license cost, because the overrun is almost always on the implementation side.

Engagements

Selected, converted,
and cut over.

OrganizationCategoryScopeResult
Adtalem / DeVryGlobal Head, Digital & Shared Services · 2008–12ERP · GlobalLed the $130M global ERP implementation across a $2.4B S&P 500 education group (Ellucian / SunGard Higher Education), plus the $8M Becker ERP and eCommerce selection spanning Lawson financials, Sterling Commerce, an Oracle data warehouse and IBM B2B.$130M ERP
KEH CameraCTO / CDO, then CIO · 2020–24ERP · Cloud · CommerceSelected and built the full order-to-cash stack — Acumatica Cloud ERP, PIM, customer MDM, a pricing and CPQ engine and Shopware commerce — with full data conversion, then moved the business to a headless hybrid-cloud architecture with CI/CD. Cutover held real-time inventory at 14,000 price changes a day.14K Prices/Day
True Value CompanyDirector, Enterprise Architecture · 2013–18Mainframe Exit · DataMigrated the wholesaler off the mainframe and built a data warehouse with KPI analytics, alongside PIM and loyalty implementations, across 4,400+ stores.Mainframe Retired
Pet Retail BrandsInterim CTO · 2018–19POS · HRIS · Post-MergerConsolidated post-merger retail systems across 1,000+ stores: migrated Oracle Retail POS and Loyalty from hybrid to Oracle Cloud and project-managed the Workday HRIS implementation with the organizational change management behind it.Post-Merger
Adtalem / DeVryInfrastructure · 2008–12Platform MigrationMigrated from Sun to x86 at Rackspace, which made replication and disaster recovery possible for the first time, then ran successful recovery tests against it.Sun → x86
McDonald'sIQ4hire · 2001–03Data Center MoveMoved McDonald's into an Exodus data center and outsourced IT operations, then transitioned to ACS for full application and infrastructure support across 30,000+ locations in 121 countries.$600M ProgramSaved $72M
SIRVADirector, Global Head · 2004–08ConsolidationConsolidated data centers across a $2.35B global moving and relocation group, with disaster recovery and swing space built out and recovery-tested.Consolidated

What We Evaluate

Where migrations
actually go wrong.

The integrator, not just the platform

Two customers on the same ERP get opposite outcomes depending on who implements it. We score the integrator, the named team and their reference cutovers separately from the software — because that is the variable that actually moves.

Data conversion, scoped honestly

Conversion is routinely under-scoped because nobody wants to own how bad the legacy data is. We make it an explicit line in the evaluation rather than a discovery in month seven.

What runs in parallel, and for how long

Big-bang versus phased is a risk decision, not a preference. It drives cost, staffing and how bad a bad weekend can get — and it should be settled before selection, not after.

The second year

Year-one pricing is the pitch. We model the steady state: support tiers, version upgrades, the modules not in the initial quote, and what the renewal looks like once the switching cost is real.