Government Permitting Software: What Cities Actually Paid, and What the Audits Found

Last updated: October 9, 2026

Across 52 permitting software contracts recorded in 2025 and 2026, cities and counties paid a median of $99,847. A quarter paid under $27,000. A quarter paid over $300,000. San Diego approved $10.9 million for a permitting system and a November 2018 performance audit found it was projected to reach $17.7 million. Seattle spent $12.3 million and saw permit throughput fall 75 percent in the first month. Here is what the software costs, and what the contracts do not tell you.

What is government permitting software?

Government permitting software is the system a city or county uses to handle permits. It takes in the application, routes it for review, tracks the inspection and issues the permit. It replaces a counter, a paper file and usually a spreadsheet.

Most products come as a set of modules. A buyer rarely takes all of them at once.

• Permits. Building, electrical, plumbing, mechanical, right of way.
• Licensing. Business licences, contractor registration, short term rentals.
• Inspections. Scheduling, field results, re-inspection fees.
• Plan review. Electronic plan review, markup, version control.
• Code enforcement. Cases, violations, abatement.
• Citizen portal. Online application, payment and status.

The number of modules drives the price more than anything else. A city buying permits alone is not the same buyer as a city buying the full land suite. The contract records show that gap plainly.

What cities actually paid

The median was $99,847. The middle half of contracts ran between $27,000 and $300,000. The range is wide for a reason. It holds a $2,000 small town subscription and a $7.6 million state platform.

Source: Civic IQ contract records, 52 permitting software awards with a stated value and a named agency, 45 agencies across 22 states. Fifty of the 52 are dated 2026. Outsourced building-services contracts, third-party inspection agreements and multi-vendor budget lines were excluded.

Here are named awards at each level.

Source: Civic IQ contract records. Values as stated in the awarding or budget document.

The most useful pair in that table is Columbia County, Georgia. It paid $296,076 to build the system and $220,754 a year to keep it running. The yearly cost is about three quarters of the build cost. It repeats every year. That ratio is what you model. It is also what a demo never shows.

Nobody publishes a price

That is not a figure of speech. RFP.wiki reviewed the category leader and states plainly: “No public list price or SKU card was verified.” The model is a yearly subscription, priced per seat, with volume discounts and a negotiated quote.

Every vendor page ranking for this term sells the product. None of them says what it costs. That is why the contract records matter. They are the only public record of the real number.

Renewals cost more than new purchases

This is the finding that surprised us most. Split the 52 contracts into renewals and new builds, and the renewals are the bigger number.

Source: Civic IQ contract records, 2025 to 2026. Two awards could not be classified either way.

The reason is not that vendors punish loyalty. It is that renewals sit with the jurisdictions that already run a large system, while new buys include every small town signing its first $2,000 subscription. Still, the direction matters. Nothing in this data supports the idea that the cost curve bends down after go-live.

Paducah, Kentucky shows what an active renewal looks like. The city renewed for three years at $438,671 a year, and in the same action moved five full permitting and licensing users to seven read-only licences. That is a city managing its seat count deliberately, which is the main lever a buyer still controls after signing.

Poway, California shows the other common renewal event: moving an existing on-premises system to a subscription. It awarded a five year agreement to shift its ERP and permitting applications to a hosted licence model. The migration itself is the purchase.

And budgets are climbing. Johns Creek, Georgia lists AI plan review at $30,000 inside a software budget line that rose to $1,019,250 for FY2027 from $836,394 the year before, an increase of about 22 percent.

The real cost is implementation, not licence

Two city audits, both several years old now, record the same failure. The software was not the problem. The build was. And the build is priced separately.

San Diego: $10.9 million approved, $17.7 million projected

San Diego’s City Council approved a permitting system purchase on September 22, 2015 for a total not to exceed $10,910,974.38. Year one was capped at $7,043,291.28, with $966,920.60 for the remaining four years.

The City Auditor published audit OCA-19-011 on November 16, 2018. It found the approved $10.9 million budget was projected to reach $17.7 million. Go-live moved from May 2017 to February 2020. That is about three years late, and near the end of a five year contract. The vendor had received $10.8 million by then. Change orders 5 through 14 added $821,024.

The audit states the cause directly. City management “skipped fundamental steps early in the implementation to speed up the process.” That left the project leaning on one person. He was the system architect, acting as both project manager and technical lead. On April 11, 2018 he retired without notice, and the blueprints were unfinished. The vendor’s own project manager left soon after. The new team ran nine change orders to rescope the work.

The audit also records a rate comparison worth reading twice. The remaining design work came to 7,305 working hours. City management reported the vendor would charge about $5 million for it. Another services firm would come to $1.1 million for the same effort. That is the audit’s own comparison, for one scope of work at one city. It is not a market rate. It is still a 4.5 times gap on labour for identical hours.

Seattle: $12.3 million, and a 75 percent drop in throughput

Seattle spent $12.3 million to buy and launch its construction permit system, which went live on April 30, 2018. The Seattle Times reported that applications completing initial review fell 75 percent between April and May, from 266 to 66. The servers were overwhelmed at launch. Part of the reason was that large numbers of builders tried to create new passwords at the same time.

The department recovered fairly quickly. It processed 229 applications in June. The department director, Nathan Torgelson, said publicly on July 2 that “the recent launch did not meet our expectations for effectiveness and service.”

Both cities still run the same vendor’s software today. Neither rollout ended the relationship. That is the point. These systems are hard to leave once your data is inside them.

New York City: what a clean audit still finds

The NYC Comptroller published audit SI17-075A on June 26, 2017 covering a city agency’s permitting deployment. It contains no dollar figures. It found weak input checks on fields including dates and employer ID numbers. It found that access was not always switched off for former employees. And it found password expiry rules were not enforced.

The user survey is the part buyers should read. 42 percent of respondents said the system requires repetitive data entry. 22 percent said it is not easy to use. 49 percent wanted changes made. A system can be live, paid for, and still cost staff time every single day.

What the audits have in common

Three findings repeat across all three. None of them is about how good the software is.

• Scope was defined late or by one person. San Diego lost its blueprints when one employee retired.
• Change orders carried the overrun. Nine of them in San Diego after the project manager left.
• Go-live is not done. Seattle’s throughput collapsed on a system that had already been paid for.

Date these properly. The San Diego audit is from 2018. Seattle is from 2018 and New York from 2017. They are not current performance reports. They are the clearest public record of how these projects go wrong. The pattern has not changed.

The vendor landscape

No vendor owns this market. In 52 awards, the biggest single supplier accounts for nine. The category is still contested. That is unusual in government software, and it is good news for buyers.

Source: Civic IQ contract records, 2025 to 2026. A contract is counted for a vendor only where that vendor is the named subject of the award. Multi-vendor budget lines were excluded. Where n is 2, no median is shown. Granicus, CivicPlus, GeoCivix and Citizenserve appear on the search results page for this term but did not appear as the named subject of any priced award in this sample.

The spread by vendor is really a spread by buyer size. Cloudpermit and GovWell show up at $3,950 and $27,000 in small towns. Clariti and OpenGov show up at $2 million and above in cities running a full land programme. A vendor comparison that ignores city size compares nothing.

For a direct feature and pricing comparison of the two incumbents, see our head to head comparison of Tyler EnerGov and Accela, and for the challenger side, how GovWell compares against Accela.

Who owns these companies, and why it matters

Who owns the vendor belongs in your renewal model. Accela took a strategic growth investment from Francisco Partners on September 6, 2023. Berkshire Partners remains, in the firm’s words, “a significant investor with an equal equity holding.” Terms were not disclosed. Two private equity owners usually means firm pricing at renewal.

OpenGov went a different way. Cox Enterprises acquired majority ownership at a $1.8 billion valuation, announced in February 2024. OpenGov reported serving about 1,900 state agencies and cities. A corporate owner and a private equity owner do not act the same way at renewal. A five year model should account for which one sits behind the vendor.

What is changing: AI plan review and permit deadlines

Two things are changing what cities buy. One is a technology that moved from demo to line item. The other is a legal deadline that makes slow permitting costly.

AI plan review is now a priced purchase

It is no longer a pilot idea. Johns Creek, Georgia bought Tyler AI plan review for $30,000 in the 2026 records. That is a small number, and that is the point. The entry price for automated plan checking is now about what one staff month costs.

Seattle ran the most rigorous public test of the idea. The city’s Innovation and Performance team, working with its construction department, tested an AI plan review tool from March 2025 to October 2025. It published its findings in June 2026. The results were specific:

• 87 percent accurate on checking whether an application was complete.
• 92 percent accurate on design compliance checks.
• Twenty former applicants and more than a dozen staff were interviewed.

Seattle’s conclusion is the useful part. It recommended a live pilot for pre-screening whether an application is complete, while it looks for funding. It also warned that accuracy alone does not save applicants time. Coverage does. A tool that catches most problems but misses some still costs the applicant a whole extra review cycle.

The city added a second warning, and it applies everywhere. The benefit depends on making your requirements clearer and less subjective first. Automation does not fix a vague code.

Treat vendor accuracy claims with care. The figures above come from a city that ran its own test and published the result. Some widely repeated numbers about time savings do not appear in Seattle’s report at all. They are not used here.

Electronic plan review without the AI layer is the more common purchase, and it is not cheap. Delray Beach, Florida moved to a plan review platform at $599,169.

Permit deadlines are becoming statutory

The reason a city buys faster permitting software is increasingly that state law requires faster permitting. Texas House Bill 14, effective September 1, 2023, requires a city to approve, conditionally approve or reject a permit application within 45 days. Miss that deadline by 15 days and the applicant may hire a qualified third party to finish the review or inspection.

That turns a service problem into a money problem. A city that misses the clock loses the review, and the fee with it. Any city under a legal deadline has a direct reason to buy software that tracks and reports cycle time. It also has a reason to make the vendor prove it during the bid, not after.

The practical step is narrow. Ask for cycle time reporting in the demo, against your own deadline, using your own application types.

What to ask before you sign

Every question below comes from something an audit found. None of it comes from a vendor checklist.

• Who owns the implementation blueprint, and what happens if they leave? San Diego lost three years when one person retired with the design in his head. Ask for the documentation standard in writing. Then ask when it is delivered.
• How are change orders priced? Nine change orders carried San Diego’s overrun. Ask for the hourly rate card and whether it is fixed for the term.
• Are professional services fixed price or time and materials? The San Diego audit recorded a 4.5 times gap between two firms for the same 7,305 hours. Rates are negotiable. They are also the largest variable in year one.
• What data migration is actually in scope? Migration is a common hidden cost. Ask how many years of records, in what condition, and who cleans them.
• How many integrations, and who builds each one? Mapping, payments, records and sign-on are usually separate work. Get the count and the owner into the contract.
• What do years two through five cost? Columbia County pays $220,754 a year to run a system that cost $296,076 to build. Model the yearly number, not the headline.
• What does go-live actually mean? Seattle was live and processing a quarter of its normal volume. Define acceptance by how much the system processes, not by the launch date.

One more, from the ownership section above. Ask who owns the vendor, and whether renewal increases are capped. It is a fair question, and the answer is public.

Frequently asked questions

What is permitting software?

The system a city or county uses to handle permits. It takes in the application, routes it for review, tracks the inspection and issues the permit. Most products come as modules: permits, licensing, inspections, plan review, code enforcement and a public portal. The number of modules drives the price more than anything else.

How much does government permitting software cost?

The median was $99,847 across 52 contracts recorded in 2025 and 2026. A quarter came in under $27,000. A quarter ran over $300,000. Small towns pay a few thousand a year. Large cities running a full land programme pay into the millions. No vendor publishes a list price.

Who are the top government permitting software providers?

In the contract records, Clariti and Tyler Technologies are each named in nine awards. OpenGov appears in five. Accela and CityView appear in three each. GovWell and Cloudpermit turn up in the smallest towns. No single vendor owns the market, which is unusual for government software.

What software does the government use for permits?

It varies by size. Large cities and counties tend to run Accela, Tyler EnerGov, OpenGov or Clariti. Mid-sized places often run CityView. Small towns increasingly buy Cloudpermit or GovWell for a few thousand dollars a year. Many agencies still run an old in-house system beside the new one.

How long does a permitting software implementation take?

Longer than the contract says. San Diego planned to go live in May 2017. A 2018 city audit records the date moving to February 2020, about three years late. Plan for phased delivery. Define acceptance by how many permits the system processes, not by a launch date.

Why do permitting implementations go over budget?

Because the licence is the small part. A 2018 San Diego audit found a $10.9 million approved budget projected to reach $17.7 million, after nine change orders. Implementation, data migration and integrations are quoted separately. The same audit recorded a 4.5 times gap between two firms for the same hours of work.

Related reading: Tyler EnerGov against Accela, head to head, GovWell measured against Accela, our comparison of procurement software for municipalities, the Euna Solutions contract and pricing picture, a guide to the permit software market, and what government CRM software costs.

Contract figures come from individual records surfaced by Civic IQ from council agendas, award recommendations and budget documents, pulled September 13, 2026. Audit figures come from the cited city audit reports and contemporaneous reporting, and are dated in the text. No vendor-published pricing, ROI or efficiency claim is used anywhere in this article. Civic IQ tracks 100,000+ agencies, $14T+ in tracked spend, 1.5M+ documents per month, all 50 states.