Six questions settle a utility ERP program before you write the RFP or ask the board for money. Answer them in order, because each one sets the next. Decide the customer information system (CIS) question first, freeze the FERC chart of accounts second, and time the go-live against the rate case third. Then budget the data cleanup, test field service on a storm day, and build the shortlist by segment. A tier 1 utility program needs 36 months from decision to go-live, so the budget cycle that starts this question is the one that funds the answer three years out.
Decide the CIS question before anything else on the budget
The CIS usually costs more and carries more risk than the ERP, so it sets everything downstream. Whether you keep billing, replace it before the ERP, with it, or after, fixes the vendor list and the schedule before any other choice. Of the five vendors that show up on utility shortlists, only SAP and Oracle own a billing engine. The five-vendor comparison covers who owns what; the point for the budget is that the sequencing decision is larger than the vendor decision.
Write down the sequence and the reason before the RFP. Four options exist, and each prices and schedules differently.
| Sequence | What goes live first | Main cost it adds | When it is right |
|---|---|---|---|
| Keep billing | ERP only; CIS untouched | A permanent CIS-to-ledger interface | The current CIS works and is not near a support deadline |
| Billing first | New CIS against the old ledger | The CIS-to-ledger interface built twice | The old CIS is the system that is failing |
| Billing with ERP | Both in one window | Two cutovers in one test year | A hard external deadline forces both at once |
| Billing after ERP | ERP, then CIS later | Carrying the old CIS three years longer | The ledger and FERC accounts must be stable first |
Consider a 400,000-customer investor-owned electric utility replacing a 20-year-old CIS and an SAP ECC-based ERP. Suppose the ERP is quoted at $25 million over 36 months and the CIS at $45 million over 30 months. Running both at once compresses the calendar but doubles the change load on the same finance and customer teams and puts two cutovers in one test year. That raises the odds of joining the 31% of implementations that a survey reported by ERP Today found ran longer than expected.
The rule we recommend is ERP first unless the old CIS is the system that is failing, and never two cutovers within six months of each other. Settle this before the budget ask, because it decides how many years of dual-run and carrying cost the program funds.
Freeze the FERC chart of accounts before build, with the regulatory accountant in the room
None of the utility ERP vendors ships a FERC module. Regulatory accounting is a chart of accounts design problem, and in our judgment it is the part of a utility program most often rebuilt after go-live. The FERC Uniform System of Accounts uses three-digit accounts that sort cost into demand, energy, and customer components for cost of service and rate design.
Those three-digit accounts have to live inside a corporate chart of accounts string that PowerPlan, a regulatory accounting software firm, says runs 15 to 20 or more digits. Get that mapping wrong and every downstream report inherits the error. Reclassifying after go-live is the most expensive rework a utility program carries.
The design phase is 10 to 20 percent of implementation hours, according to ERP Research's cost breakdown, and the regulatory accountant belongs there, not in testing. Budget that person's time as a named line, not as a review at the end.
Confirm early who has to comply, because it is broader than many CIOs assume. Major electric utilities file FERC Form 1 once they pass thresholds such as 1 million MWh of sales for three consecutive years. Municipal systems are exempt but usually use the USOA anyway, because bond covenants and state law depend on it, and cooperatives file the analogous RUS Forms 7 and 12.
Set the go-live date against the rate-case test year, or plan a 182.3 deferral
For an investor-owned utility, the ERP program is recovered through a rate case, so when it goes live decides when it earns a return. A program that goes live inside the test year carries into the next filing. One that slips past the test year waits for the case after that. The CFO needs this answer before the budget is set, not after the SI publishes its resource plan.
Timing is tight. A general rate case in Indiana typically takes 10 to 12 months from filing to order, and a utility there cannot refile for 15 months. Rate activity is heavy: June 2025 alone saw 17 initial rate increase requests totaling $3.846 billion, according to S&P Global.
Take the same 400,000-customer utility. Suppose its last order came in March 2026 and its state uses a historical test year, with a plan to file in June 2028 on a calendar 2027 test year. An ERP go-live in October 2027 lands inside the test year, so the capitalized build and the first subscription payments both enter the filing. Slip go-live to February 2028 and the same spending falls outside the test year. Recovery on a $25 million program then begins two and a half to three and a half years later, with carrying cost at the authorized return the whole time.
When the date cannot land inside a test year, the escape is FERC account 182.3, Other Regulatory Assets. Major costs can be deferred there when it is probable a regulator will include them in rates later, under ASC 980 for FASB filers or GASB 62 for public power. Decide which of the two you are planning for, set the go-live from the test year backward, and write the chosen path into the budget narrative.
Budget asset-data cleanup as its own line, not a contingency
The asset register is usually the largest data object in a utility, and the capital program runs through it. Legacy asset records, some still on paper, often need heavy work before migration. If that work is folded into the SI's fixed price or hidden in contingency, it is the first thing cut when the schedule slips, and the cost reappears after go-live as crews close work orders against stale records.
Fund data cleanup as a named line in the budget, owned by the utility rather than the SI, and start it before build. The work does not need the new system to begin. A separate line also gives the board a number to hold, which a buried allowance never does.
Demo a storm day, not a planned meter exchange
Every vendor on a utility shortlist sells a field service product, so the differences only show up under load. Two tests separate them. The first is whether field service shares the asset register with asset management or merely syncs to it, because a crew closing a work order against a stale record recreates the cleanup you just paid for. The second is whether the product handles storm callout, mutual aid crews, contractor crews, and offline work.
Put a storm day in the demo script. Ask each vendor to run callout for 300 crews, mutual aid, and offline work-order close against the same asset record the finance team will capitalize. A planned meter exchange proves nothing a brochure does not already claim.
Build the shortlist by segment, not by analyst ranking
The US utility market is three markets, and the segment changes the shortlist more than vendor features do. Investor-owned utilities file FERC Form 1, run rate cases, and buy at tier 1 scale. Just over 2,000 community-owned utilities serve about 15 percent of electric customers, and NRECA counts 830 distribution cooperatives and 60 generation and transmission cooperatives serving an estimated 42 million people.
| Segment | Roughly how many | Typical CIS pattern | Shortlist to start from |
|---|---|---|---|
| Investor-owned utility | Form 1 filers above 1 million MWh | Its own CIS program, often replaced separately | SAP S/4HANA, Oracle; IFS Cloud where assets dominate |
| Municipal or public power | Just over 2,000 community-owned utilities | Billing inside a city or public-sector suite | Infor CloudSuite, Dynamics 365 with a partner CIS, Oracle |
| Cooperative or small municipal | 830 distribution cooperatives | Billing from a cooperative-specific vendor, left alone | Dynamics 365, Business Central; SAP or Oracle for large G&Ts |
An analyst ranking built for the average utility will push a small cooperative toward a tier 1 suite it cannot staff, or an investor-owned utility toward a product with no FERC story. Start from the segment, then let the CIS decision from the first question narrow the list. ERP Research's published cost bands run from $150,000 for a small cooperative to more than $5 million for an investor-owned utility, with no stated method, so plan at the top of each band and treat the numbers as a sanity check rather than a budget.