Many public agencies do not need a full replacement when they ask when to replace a government ERP. Keep versus replace, for a city or a county, is a decision. It is not a diagnostic you outsource to a procurement. Going to market is one option. It is not the default, and it is not how you find out whether the system you already run still has life.
This page is a decision framework for CIOs, finance directors, and board members. It is not a sales pitch for a new system. Avèro’s fee is the same whether the findings support keep, modernize, or replace.
What replace actually costs
A mid-size public-agency ERP replacement is not a software purchase. The published range to take into a board packet: a full replacement typically runs several million dollars and three to five years, counting license, implementation, backfill staffing, and change orders. The figure in a vendor proposal is usually the license and the implementation. Staff time, parallel running, data conversion, retraining, and slower work while the organization relearns its processes sit outside that quote.
Modernization of a stable core is a different shape of work. If the core is stable, the better path is usually to clean up the data, redesign the processes, and add purpose-built tools and responsible AI around the system you have, which avoids a multimillion-dollar trip to market. Modernization of an existing platform is usually a fraction of replacement cost. That is a comparison of paths, not a savings guarantee and not a go-live date.
Avèro’s advisory fees are published on the services page. Readiness and selection generally runs $150K to $1.5M depending on scope, over three to eighteen months. Contract review runs $10K to $35K. Those figures are the cost of advice, not the cost of software, and they do not change based on which logo the findings point to.
Signals the current system still has life
Keep or modernize is the better path when the evidence looks like this.
Core finance, HR, and payroll are stable enough for day-to-day operations. The books close, people get paid, and the statutory reports go out. Stability is not delight. It is enough to refuse a panic replacement.
The pain is mostly process, data quality, reporting, or integrations, not a dead platform. A chart of accounts copied from the last system, approval paths that were never mapped, modules licensed and never turned on, and spreadsheets beside the system of record are operational problems. They follow you into a new platform if nobody names them first. The test in Going to Market Is a Decision, Not a Diagnostic is specific: for each top complaint, can you name the technical constraint that makes it impossible? If the answer is a shrug or a sales sheet, it is a configuration or process problem.
The vendor roadmap still covers mandated reporting and statutory needs. An end-of-support letter is a commercial event, not proof that the platform cannot meet the next audit. A roadmap that still covers what the law requires is a reason to keep the decision on your clock. How to read that letter is in Manufactured Urgency.
Leadership does not yet share a picture of where the community is going. A new system will codify today’s mess if the board, the chief executive, finance, and the departments have not agreed on the outcomes that matter. Buying software does not settle that argument.
Signals replacement is actually justified
Replacement is justified when the evidence looks like this. These are reasons to go to market, not reasons to skip the assessment.
The platform cannot meet statutory or audit requirements without unsustainable workarounds. If a required report exists only as a side process that breaks when one person is out, the core is no longer carrying the agency. When the cost of the workaround is the audit, the finding is replace, not train harder.
The vendor has declared end of life, or the stack is unsupported, and there is no credible path to stay. That means you have looked, including at third-party support where it exists, and the remaining options do not cover security, statutory reporting, or a workforce that can still run the system. A date in a letter is not that finding.
The data model or the architecture blocks any responsible AI or analytics path the agency needs. That path needs trusted data, named owners, and processes written down before anyone automates them. If the core cannot produce that, replacement is how you get a data model the next decade can use. The Government AI Navigator maps products in public. It is not a reason to buy one first.
The cost of keeping exceeds a disciplined selection. Customizations only one person understands, dual systems that were never retired, and a failed project that left both platforms in production can cost more than a clean selection. Put both costs on one page, with the same assumptions, before an RFP. GoTriangle commissioned that sequence before a procurement: needs, a requirements matrix written from the work, and both paths priced. That published account is the sequence, not an outcome. The Port of Portland has trusted Avèro with other enterprise system work.
Three questions before an RFP
Going to market is a decision, not a diagnostic. Before a requirement is written into an RFP, a board can usually settle keep, modernize, or replace with three questions. They are already published in the Going to Market field brief, which we email on request. It is not a public download.
1. What is actually broken? The system, or how the organization uses it. Most “the system can’t do that” complaints mean the system was never set up to do that. Name the constraint, or admit you cannot.
2. Can the work be specified without naming software? If the requirements describe the incumbent, you will buy the same thing again. A list written from the work, not from the logo, can be tested against both futures: the current platform, tuned, and a replacement. Start at Zero is that method. It is how the assessment is run, not a product for sale.
3. What is the honest cost of each path? Replacement and optimization, on the same page, with the same assumptions. Include backfill, parallel operations, data conversion, and slower work while people relearn the processes. Include the cost of tuning, which is not free just because no contract is involved. The decision often makes itself once both numbers are visible.
If those three point toward a market, add who sits on our side of the table through selection and go-live. The integrator works for the integrator. The vendor works for the vendor. Someone in the room has to work only for the agency.
What keep looks like in practice
Keep is not “do nothing.” A stable core still needs a modernization path, or the same complaints return with a louder vendor attached.
Data governance comes first: name owners for the records finance, HR, and payroll depend on, and decide what is clean enough for a board report. Process redesign is next. Side spreadsheets are the process. Write them down, decide what the system should carry, and retire the rest. Then name outcomes worth measuring, such as cycle time, hours spent reconciling, and reports still assembled by hand. If you cannot say what better means, you cannot tell whether modernization worked.
Purpose-built tools and responsible AI sit around that stable core, not in place of it. Data and AI strategy is that sequence: use cases the data can support, and governance a council can adopt. It is not a reason to replace the core, and not a reason to adopt a tool because a demo went well.
Who should run the assessment
The assessment should be run by an advisor the agency pays, and nobody else. No licenses. No referral fees. No system-integrator role on the same engagement. Firms that sell software, implement it, or take a fee when a logo wins are paid more when the answer is replace. Avèro holds no reseller relationships with Tyler, Workday, SAP, Oracle, Infor, or any other vendor a public agency might evaluate. We do not implement software. We work only with public agencies.
The recommendation costs the same whether the finding is keep, modernize, or replace. That is the point of structural independence. An uncomfortable finding (the platform is fine, and the problem is how the organization uses it) can then go in the board packet.
When replacement is the path, the selection still has to be traceable. Caliber is how Avèro keeps requirements, scoring, and evidence in one record. It is not sold as a product. ERP advisory and selection covers readiness, requirements, the RFP, and vendor evaluation, on the buyer’s side through go-live.
If you are mid-decision, under board pressure, or holding a vendor letter, write to us. A sentence is enough. You will hear back within one working day. If we cannot help, we will say so. The office line is 865-415-3848.
AV, Avero Advisors
FAQ
Do we have to replace our government ERP?
Often, no. Replacement is one option, not the default. If core finance, HR, and payroll are stable, the better path is usually to clean up the data, redesign the processes, and add purpose-built tools and responsible AI around the system you have. If the platform cannot carry the agency forward, say so and run a selection. The assessment fee is the same either way.
How do we know the current system still has life?
Four signals. The core is stable enough for day-to-day finance, HR, and payroll. The pain is mostly process, data quality, reporting, or integrations, not a dead platform. The vendor roadmap still covers mandated reporting and statutory needs. And leadership does not yet share a picture of where the community is going, so a new system would codify today’s mess.
What does modernization cost relative to replacement?
A full ERP replacement for a mid-size public agency typically runs several million dollars and three to five years, counting license, implementation, backfill staffing, and change orders. Modernization of an existing platform is usually a fraction of that replacement cost. Avèro’s advisory fees are published separately on the services page: readiness and selection generally runs $150K to $1.5M depending on scope, and contract review runs $10K to $35K. Those are advisory fees, not software prices, and they are not a savings guarantee.
Who should advise us if we might go to RFP?
An advisor paid only by the agency, with no licenses to sell, no referral fees, and no system-integrator role. Avèro works exclusively with public agencies. When replacement is the path, Caliber keeps requirements, scoring, and evidence traceable. The recommendation costs the same whether the finding is keep, modernize, or replace.