A tracker records where a bill is. It does not hold the work behind it.
The record is distributed across tools,and reconciling it is manual.
Bills sit in a spreadsheet, notes in email, documents on a shared drive, and the current analysis in a file on one person's machine. Each holds part of the answer.
Reconciling them is manual work performed under time pressure, and the cost is not visible until someone has to reconstruct a file that another person built.
Where context is lost
A manual tracker is accurate only on the day it is updated.
A shared spreadsheet depends on one person to maintain it, and updates compete with everything else that person is responsible for.
Between updates the tracker reports a position the legislature has already moved past, and the team plans against it without knowing that it has.
Status, chamber, stage, and last action update from the source on every row, so the list does not depend on anyone maintaining it.
Relevance is not confined to the bills already tracked.
A bill outside the tracked set can be amended in committee to reach a program the team is responsible for.
Tracking by number cannot surface it, so the first notice comes from outside the organization, usually from someone asking what the position is.
Bills can be tracked by topic, agency, and keyword as well as by number, so a bill that becomes relevant is treated as relevant.
Context does not transfer with the assignment.
When responsibility for a file changes mid-session, the reasoning behind it sits in email threads and a folder structure one person understands.
Rebuilding that context takes a week the session does not allow, and decisions get remade because the record of the first one cannot be found.
Notes, documents, tasks, and comments are attached to the bill, so the context is present for whoever picks it up.
Fiscal analysis is assembled by re-keying.
Figures are drawn from the bill, the fiscal analysis, and the agency's own numbers, then entered into a report section by section.
Most of the preparation window goes to transcription rather than to analysis, and every manual step is a place for a figure to be entered wrongly.
Reports start from an outline the team reuses, with figures extracted from the bill and uploaded documents, and section status showing what remains.
Version history is reconstructed after the fact.
Questions arise about which draft was circulated and when, often once a position is already public.
The answer has to be assembled from sent mail and file names, and until it is, nobody can say with confidence what was represented to whom.
Version history and snapshots are kept on every analysis, and any previous state can be restored and compared.
See Billflo in action.
