Council Agenda Management Software Problems: The 5 That Break the Packet, and How to Avoid Them

Why does approval routing get scoped as a linear workflow?

Because that is what routing looks like from outside the clerk's office. An item is drafted, reviewed by a director, checked by the attorney, confirmed by finance, then agendised. Drawn on a whiteboard it is a straight line with four boxes, and it fits neatly into any workflow engine.

Actual routing is a decision tree, and the branches come from your own rules rather than from general practice. Routing depends on item type, on whether the dollar amount crosses the bid threshold, on whether the item touches a labour agreement, on whether it amends the municipal code, and on which body hears it first. Some items go to a committee and then to council. Some go to two committees in sequence. A joint meeting with a redevelopment agency or a special district follows a path nobody has written down. And when a director retires, their approvals move to an interim, which in a hard coded workflow is a change request with a cost and a wait.

Two things fix this. Make the routing rules data that the clerk can edit, not code a developer maintains, so a threshold change or a staffing change takes ten minutes rather than a support ticket. And give every item a state, a current holder and a due date computed backward from the statutory posting deadline, so the system tells finance on Monday that their fiscal impact is due Wednesday.

The clerk's view then shows the packet as a set of items with a count of what will miss, five days out instead of the night before. That single change turns the packet from a chase into a schedule.

What goes wrong when you migrate historical minutes and ordinances?

It is almost always the largest line in the project and it is almost always scoped as an afterthought. The history is scanned portable document files of wildly varying quality, some produced by a photocopier in the 1990s, and a meaningful share of it is the legal record of decisions still in force.

Two mistakes recur. The first is trying to fully structure everything, which means extracting item numbers, vote records, ordinance numbers and code references from thousands of scans and reviewing each one. That work is real and expensive, and most of it buys nothing, because nobody queries the vote breakdown on a 1987 consent item.

The second is the opposite: dumping everything into full text search and declaring victory. That fails the day a title company asks whether a code section was amended in 2013 and again in 2019, because full text search will find three documents and cannot tell you which superseded which.

Tier the corpus. Full structure for the recent period, typically the last ten to fifteen years, where ordinances get their relationships to code sections recorded and votes are queryable. Searchable archive for everything older, with good optical character recognition, correct dates and a scanned original that can be produced on request. Then handle the exception explicitly: any ordinance still in force, however old, gets structured regardless of its date, because that is the set people actually ask about. Price the two tiers separately so the trade off stays visible.

Why do video, streaming and codifier handoffs break after launch?

Because they are one way exports that nobody watches. The pattern is familiar: an adopted ordinance is exported as a document to the codification vendor, the meeting recording is pushed to a streaming provider, and the connection between the three systems is a file that somebody sends.

The failures are quiet. An ordinance is adopted with a floor amendment, the export carries the pre amendment text because the export was built from the staff report rather than from the adopted version, and the municipal code is now wrong. A meeting runs long, the recording is split into two files by the streaming platform, and the item timestamps written by your clerk console no longer line up with anything. A streaming provider changes an identifier format and the deep links on two years of public item pages stop resolving, which nobody notices until a resident complains.

Build the video index as a byproduct of running the meeting rather than as a separate task, timestamping each item when the clerk console advances to it, and store those marks against the meeting rather than against a provider specific media identifier so a re encode or a file split can be remapped. Treat the codifier handoff as a queue with acknowledgements rather than an export, and reconcile monthly: adopted ordinances in your system against sections updated in the code.

What happens when amendments and substitute motions are not covered?

The official text of what passed gets reconstructed afterwards from a video recording, by a deputy clerk, days later. That is the gap in most packaged systems and it is the one that carries genuine legal weight.

What happens in the room is fast and messy. Item 7 is continued to the next meeting. Item 9 is amended twice on the floor and the second amendment fails. A substitute motion passes four to three with one member abstaining because of a conflict. Item 12 is pulled from the consent calendar by a single member and heard separately. Ninety seconds of proceedings produce a vote record, an ordinance text and a minutes entry that all have to be right.

Most systems record a vote on an item and a disposition. They do not model an amendment as a versioned change to specific text that carries forward if adopted and disappears if not, and they handle substitute motions poorly or not at all. So the adopted version is assembled by a human reading the recording, and if that human is wrong, the code is wrong.

Model the meeting as a sequence of events: motions, seconds, amendments attached to specific text, withdrawals, roll calls and dispositions, captured on a clerk console as they happen, with recusals flagged and their reasons recorded. The adopted ordinance then assembles itself from the amendments that carried. When the gavel falls the vote record is finished and the minutes draft is generated from the event log rather than from memory.

Should you build custom or configure what you already own?

If you run one council and one planning commission with fewer than roughly forty items a month and a charter that reads like everyone else's, configure CivicClerk or PrimeGov and stop there. They are priced for exactly that jurisdiction and include the hosting, the public portal and the support you would otherwise carry yourself. A build there is a poor use of general fund money and we will say so.

If you already run Granicus Legistar and it works, keep it. It has the deepest legislative model in this market and large cities and counties run on it for good reason. Municode Meetings is a reasonable answer where your codification already sits with the same provider. Before commissioning anything, check whether the pain is unconfigured product rather than product limits, because a great deal of clerk frustration is item templates and routing that were set up once during implementation and never revisited.

Build when two or more of these are true. You run many bodies with genuinely different rules and the product forces them into one shape. Your charter or enabling act creates a process the product cannot express, so staff maintain a parallel manual process. You are a county or large city where the annual licence has grown into six figures and the roadmap still does not reach your case. Or a failed posting or a wrong adopted text has already cost you an action.

How do hidden costs get into the quote?

Boards and commissions. A quote priced against the council is priced against the simplest body you have. A county with two dozen boards, each with its own quorum rules, posting requirements and public comment procedure, is a different project, and the second body is where the assumptions in the first one break.

Accessibility is the second, and it is now a legal obligation rather than a preference. The Department of Justice rule on web accessibility for state and local governments under Title II of the Americans with Disabilities Act sets Web Content Accessibility Guidelines 2.1 Level AA as the standard, with compliance dates that arrived in 2026 for larger jurisdictions and 2027 for smaller ones. Confirm your own date and scope with counsel. Agenda packets are usually the largest volume of public documents a city publishes and they are full of untagged files and scanned exhibits. Publishing structured item pages first and assembled documents second is the practical route, and checking uploads at submission so a department fixes its own unsearchable scan is far cheaper than the clerk discovering it during assembly.

Then joint meetings, multilingual posting requirements, records retention enforcement, and historical migration. For calibration, in Digital Heroes delivery experience a focused first release covering item submission, configurable approval routing, packet assembly and public posting runs $55,000 to $110,000 and ships in 10 to 14 weeks. Adding the clerk console, motion and vote capture, minutes generation, ordinance numbering and code relationships, video indexing and a searchable public portal takes it to $130,000 to $300,000 over 6 to 10 months.

What separates a build that works from one that fails here?

The ones that work run in parallel for one or two meeting cycles rather than cutting over cold. The parallel period is where undocumented practice surfaces, such as how your consent calendar handles a pulled item or how a joint meeting is numbered. Treat it as planned project time, not slippage.

They scope the first phase to the council and one commission, then add bodies. Every body you add teaches the routing engine something, and adding them one at a time is how the rules stay editable rather than accumulating exceptions in code.

They leave the streaming vendor in place and integrate, because replacing video infrastructure adds cost and risk to a project whose value is in the packet and the record.

They ask a developer to walk through a meeting on a whiteboard before contracting. Someone who has done this draws motions and amendments as ordered events and asks about substitute motions before you mention them. Someone who draws items with a status field has built a ticketing system.

And they settle ownership before kickoff. The jurisdiction should hold the repository, the cloud accounts and the right to hire another firm. At Digital Heroes the client owns the code from the first commit. Public records live in this system for decades, which makes vendor dependency a governance question rather than a procurement preference.