ERP is the decision that defines the program.

Why ERP is the central decision.

The ERP is where financial transactions, inventory movements, customer orders, vendor invoices, and the chart of accounts meet. It is the system of record for almost everything that produces a journal entry. In a carve-out, the seller's ERP contains both seller and Newco data, runs on seller controlled infrastructure, and is governed by seller policies. None of that survives the long term ownership change. The ERP decision is the decision about how Newco will operate, close the books, run procurement, and report financials for years after Day One.

The ERP decision drives the TSA. If Newco stays on the seller's ERP through the TSA, the TSA includes seller IT operations, license sublicense, support, and data segregation. Each of those creates pricing pressure and dependency risk. If Newco moves off the seller's ERP, the migration project drives the IT separation timeline. Either way, ERP is the gravitational center of the program.

The decision also drives the operating model. ERP choice cascades through procure to pay, order to cash, record to report, plan to inventory, and hire to retire. A different ERP often means a different set of business processes. Process change accompanies platform change and often costs more than the platform itself. Buyers who plan only the technology budget miss the larger process change budget.

The decision needs to be made early. Procurement, implementation, data migration, and parallel running take 9 to 24 months depending on path. ERP decisions made after signing usually mean the TSA carries the seller's ERP for a year or more. ERP decisions made during diligence usually allow Newco to land on its own platform within the TSA window. The IT context is in the carve-out IT separation playbook.

The four ERP paths every Newco considers.

Path one is clone and migrate. Newco creates a copy of the seller's ERP instance, separates the data, and runs the copy under Newco ownership. The path preserves process continuity and is the fastest for large complex environments. It carries license and contract risk because most enterprise ERP licenses do not allow free cloning. The seller must agree to the clone and the vendor must accept the license arrangement. Negotiated correctly, this path can land in 9 to 15 months.

Path two is greenfield on the same platform. Newco stands up a new instance of the same ERP product with a fresh configuration. The path produces a cleaner instance, often simpler than the inherited configuration. It carries a longer timeline, typically 12 to 18 months, because the implementation is a real project rather than a copy. It works well when the seller's configuration was over engineered for the Newco scale and a simpler instance suits Newco.

Path three is greenfield on a different platform. Newco selects a new ERP platform appropriate to its standalone scale and operating profile. The path produces the cleanest Newco environment with no residual seller dependency. It carries the longest timeline, typically 15 to 24 months, and the highest change cost. It works well when Newco is materially smaller than the seller and the seller's enterprise platform is oversized. PE backed Newcos in mid-market scale often choose this path.

Path four is stay on the seller's ERP indefinitely. Newco signs a long term TSA or a managed services agreement and never moves. This path is rarely the buyer's first choice but appears in industries where regulatory or operational complexity makes migration uneconomic. It locks in pricing pressure from the seller for the life of the relationship. It produces stranded cost exposure for both sides. It is the path of least resistance and the path with the highest steady state cost.

How to choose between the paths.

The decision rule starts with Newco scale relative to the seller. If Newco is 20 percent or less of the seller by revenue, employees, and transaction volume, the seller's enterprise ERP is almost certainly oversized for Newco. Greenfield on a simpler platform usually produces lower steady state cost and better fit. If Newco is closer to half the seller's scale, clone and migrate is often the right answer because the seller's platform is already sized appropriately.

The second factor is industry specificity. Highly specialized industries with deep ERP customization, like aerospace, defense, regulated manufacturing, or large healthcare systems, often have configurations that took years to build. Greenfield rebuilds in those industries are expensive and slow. Clone and migrate is usually the right answer if licensing permits. Less specialized industries with standard configurations can move to new platforms more readily.

The third factor is buyer thesis. PE buyers with a 3 to 5 year hold often choose greenfield on a leaner platform because the operating model in the value creation plan calls for smaller, simpler systems. Strategic buyers integrating Newco into their existing footprint typically prefer to migrate Newco onto the buyer's own ERP, which is a fifth path that lives outside the four general paths above. The buyer's destination drives the source decision.

The fourth factor is timeline pressure. If the TSA window is short and the seller is unwilling to extend, the fastest viable path wins regardless of cost. If the TSA window is generous and the extension fees are reasonable, the path that produces the best steady state wins. The economic comparison between TSA cost and migration cost is covered in the TSA exit versus extension cost analysis.

Data, history, and the close cycle.

Whatever ERP path Newco chooses, the data migration is the work item that most often slips. The general ledger needs opening balances on Day One. The accounts payable and receivable subledgers need open items. The inventory module needs item masters, locations, and quantities. The customer and vendor masters need addresses, payment terms, and tax codes. The fixed asset register needs cost basis and depreciation schedules. Each of these needs reconciliation between the source and the target.

Historical financial data is a separate question. Most regulators require Newco to retain access to the historical books even after the new ERP is in place. The buyer can take an extract of historical data, host it on a Newco system, and serve it from there for the retention period. Alternatively, the buyer can pay the seller through the TSA for read access to the historical ERP. The first option is more durable. The second is faster to set up.

The first close cycle on the new ERP is the most stressful operating moment after Day One. The trial balance must tie to the opening balances. The subledgers must reconcile to the general ledger. The intercompany eliminations must work. The reporting package must produce financials that the audit committee will accept. Programs that planned the data migration and rehearsed the close usually pass the first close. Programs that left the close to land on its own usually fail it.

Parallel running for one or two close cycles before cutover is the safest pattern. The cost is incremental staff time across two systems. The benefit is that the cutover happens after the new ERP has produced clean financials at least twice. The pattern is widely understood. The discipline to schedule it is often missing because the project pressure to cut over compresses the parallel window.

Where ERP programs typically fail.

The most common failure is choosing the wrong scope. Buyers who try to redesign business processes alongside the ERP migration usually do neither well. The migration takes longer, the processes do not fully stabilize, and the project misses milestones in both dimensions. The disciplined approach migrates the existing process to the new platform first and redesigns the process later, once the system is stable. The two phase approach almost always finishes faster than a combined program.

The second common failure is underestimating the integration footprint. The ERP touches dozens of upstream and downstream systems. Each integration is a small project. Many integrations were built years ago by people who have left. Documentation is uneven. The integration inventory needs to be built early and validated against operations. Programs that discover integration scope mid project usually miss the cutover date.

The third common failure is the change management gap. Users trained two weeks before go live make mistakes in the first close. The mistakes compound into reconciliation problems that take months to clear. The training program needs to start sooner and run longer than most projects budget. Training is not optional and not cheap. Programs that fund it adequately produce smoother cutovers.

The fourth common failure is letting the TSA cover an unbounded ERP service. Sellers will gladly provide ERP services through the TSA for as long as the buyer pays. The pricing is rarely favorable. The buyer needs a hard exit date in the TSA with extension fees that escalate sharply if the buyer misses it. The pressure of the exit date is what keeps the migration on schedule. The mark-up and extension fee dynamics are in TSA mark-up benchmarks.

More on carve-out systems.