The County That Went from Zero to Fully Accessible in 90 Days

Prepared by Convene Research and Development

Executive Summary

The program reduced decision latency via short runbooks and micro‑drills, paired with contract language that turned vague promises into measurable service levels (latency, accuracy, completeness).

How the Team Moved Fast

Residents reported clearer audio, fewer caption delays, and immediate access to complete archives. The county prioritized outcomes that residents can perceive rather than internal checklists.

What Changed for Residents

This case study details how a mid-size county transformed public meeting accessibility from near-zero capabilities to a sustainable, resident-centered program in 90 days. The program aligned governance, technology, and operations to deliver intelligible live streams, reliable captions and interpretation, and complete archival bundles posted on schedule.

The approach emphasized measurable outcomes residents can feel—clarity, continuity, inclusivity, and reliability—backed by simple governance (roles and drills), procurement clauses that preserve portability, and a publication workflow that produces accessible artifacts by default.

How the Team Moved Fast

Residents reported clearer audio, fewer caption delays, and immediate access to complete archives. The county prioritized outcomes that residents can perceive rather than internal checklists.

What Changed for Residents

1. Context and Initial Conditions

A walk‑through and live‑meeting shadow produced objective artifacts—encoder logs, caption screenshots, and link audits—so leaders could see bottlenecks instead of debating anecdotes.

Evidence‑Based Baseline

Headcount and budget could not expand mid‑year. The team explicitly deprioritized cosmetic upgrades in favor of intelligibility, accessibility, and publication integrity.

Constraints and Non‑Goals

Prior to the project, the county streamed intermittently, provided post-event captions when time allowed, and lacked a canonical repository for complete meeting artifacts. Public complaints centered on intelligibility, the absence of Spanish materials for budget hearings, and broken archive links.

Capacity was limited: one AV technician, a part-time webmaster, and a clerk’s office focused on agendas and minutes. Any solution had to improve reliability without expanding headcount—hence the emphasis on checklists, pinned configurations, and a practical division of duties.

Table 1. Baseline inventory and pain points

Area	Observed State	Pain Point	Owner
Live audio/video	Single encoder; no standby	Outages during peak meetings	AV/IT
Captioning	Manual, post-event only	Late or missing captions	Accessibility/Clerk
Interpretation	Ad hoc, phone bridge	Echo; inconsistent quality	AV/Accessibility
Publication	Recording only; filesplit	Broken links; no captions/transcripts	Records/Web
Governance	Shared logins; no drills	Slow response; unclear roles	Clerk/IT

2. Program Goals and Resident-Facing Outcomes

High‑salience meetings (budget, redistricting) received stricter targets and additional monitoring, while routine briefings followed a lighter checklist to conserve capacity.

Tiering by Meeting Salience

Success meant intelligible streams, inclusive access during the meeting, and trustworthy archives afterward. The team published these definitions to set expectations and guide trade‑offs.

Defining Success in Plain Language

The county adopted outcomes that residents would recognize immediately: intelligible live audio, captions with low latency, interpretation where required, and a complete bundle published to a single page with stable links. These goals framed the technical choices and the operational cadence.

Table 2. Resident-facing outcomes and verification

Outcome	Target	How Verified	Cadence	Owner
Intelligibility	No clipping; SNR > 20 dB	Rehearsal clip; operator meter	Per meeting	AV
Caption latency	≤2.0 s (Tier A)	Dashboard readout	Per meeting	Accessibility
Interpreter uptime	≥99% (Tier A)	Encoder/ISO logs	Per meeting	Accessibility/AV
Archive completeness	100% within SLA	Checklist + link audit	Weekly	Records/Web
Corrections transparency	Public, dated notes	Corrections page	Continuous	Clerk/Records

3. Governance and Roles

A one‑page scorecard tracked SLOs, and a change log recorded version pins, network changes, and vendor updates to keep context available for audits.

Scorecards and Change Logs

Each role had a deputy empowered to act, eliminating single‑point staff risk during vacations and illness. Deputies trained using the same micro‑drills to maintain parity.

Deputies and Coverage

An ICS-lite model named four accountable leads—Incident (Clerk/Deputy), Operations (AV), Communications, and Records—with designated deputies. Weekly change windows avoided updates during marquee meetings and established a predictable rhythm for improvements.

The model reduced decision latency during incidents and aligned ownership for resident-facing outcomes. A one-page monthly scorecard tracked progress toward targets and documented next actions when thresholds were missed.

Table 3. RACI for accessibility operations

Process	Requester	Clerk/PM	AV/IT	Accessibility	Records/Web	Legal/Comms
Intake and scoping	R	A	C	C	C	I
Caption engine configuration	I	C	C	A/R	I	I
Interpretation routing	I	C	A/R	A/R	I	I
Publication bundle	I	C	I	C	A/R	I
Corrections and errata	I	A	I	C	R	C

4. 90-Day Critical Path

Change‑freeze windows and version pins prevented regressions during high‑salience sessions. Drill timelines remained short to lower disruption risk.

Protecting Marquee Meetings

Identity and endpoints first reduced blast radius; then accessibility workflows stabilized; finally, publication and link audits ensured residents saw the benefits immediately.

Sequencing Rationale

Work proceeded in three phases: stabilize identity and endpoints (Days 1–30), institutionalize accessibility and publication (Days 31–60), and rehearse failover and finalize procurement clauses (Days 61–90). Each phase produced visible artifacts residents could see: a canonical meeting page, an accessibility corrections page, and a monthly scorecard.

Table 4. Milestones by phase with owners and artifacts

Phase	Milestone	Owner	Artifact/Proof
Days 1–30	SSO + MFA for admins; remove shared accounts	IT/Clerk	Access test; role roster
Days 1–30	Pinned caption engine; glossary baseline	Accessibility	Latency snapshot; glossary file
Days 31–60	Interpretation routing; ISO capture verified	AV/Accessibility	ISO tracks; operator checklist
Days 31–60	Canonical page + bundle format	Records/Web	Linked bundle; checksum log
Days 61–90	Standby encoder + LTE profile; simulcast	AV/IT	Drill timeline; encoder config export
Days 61–90	Procurement addendum with SLOs	Clerk/Procurement	Contract clause addendum

5. Technology Architecture and Configuration

Per‑user logins replaced shared accounts; updates occurred in maintenance windows; presets were laminated to reduce operator variance.

Configuration Hygiene

Standby encoders sat on separate power and VLANs; an LTE profile was pre‑tested monthly. The golden path diagram lived at the console for crisis clarity.

Golden Path and Independent Failure Domains

The design centered on a documented golden path from microphone to archive with independent failure domains for power, network, and provider. Standby encoders were placed on UPS and a separate VLAN; dual RTMP destinations provided platform redundancy; ISO recording preserved per-language quality for the archive.

Table 5. Minimal technical stack with configuration cues

Layer	Primary	Fallback	Configuration Cue
Capture	Goosenecks; lectern mic	Handheld wireless	Gain ledger; saved presets
Video	PTZ cameras with presets	Static wide camera	Laminated preset sheet
Encode/Record	Primary encoder + cloud record	Standby encoder + LTE profile	Dual RTMP; health alerts
Caption/Translate	Pinned engine + glossary	Alternate engine; human pass	Latency ≤2 s; QA sampling
Publication	Canonical page + bundle	Corrections note	Checksums; link audit
Identity	SSO + MFA; per-user roles	Break-glass with dual control	Quarterly access review

6. Accessibility Operations and QA

Random sampling of captions and translations produced targeted fixes—engine switches, glossary edits, or a brief human post‑edit pass for sensitive items.

Sampling and Corrective Actions

Quarterly glossary refreshes incorporated common resident terms and program names from community organizations to keep terminology consistent across languages.

Glossary Cadence and Community Inputs

Accessibility was operationalized with measurable targets, a living glossary, and sampling-based QA. ASL picture-in-picture consistency and interpreter return audio were verified in daily preflights. Where residents reported issues, the county posted corrections notes with timestamps to normalize transparency.

Table 6. Accessibility KPIs and action thresholds

KPI	Target	How Measured	Action on Miss
Caption latency	≤2.0 seconds	Operator dashboard	Switch engine; check audio path
Caption accuracy (sampled)	≥95%	Reviewer rubric	Glossary update; post-edit pass
Interpreter uptime (Tier A)	≥99%	Encoder and ISO logs	Hot swap; verify mix-minus
ASL PiP visibility	≥95% of meeting	Operator checklist	Preset recall; PiP lock

7. Publication Workflow and Records Integrity

A simple script verified expected links post‑publish and flagged missing artifacts for same‑day remediation.

Automation and Link Integrity

Each meeting’s artifacts appeared on a single canonical page with stable URLs. Hashes, timestamps, and corrections notes preserved chain‑of‑custody and trust.

Canonical Page Discipline

The county created a canonical meeting page template that linked the recording, captions (WebVTT), transcript (HTML/PDF), agenda, minutes, and translations. Media files were hash-checked on upload; weekly link audits prevented silent regressions. A corrections page documented post-publication fixes with dates and reasons.

Table 7. Publication bundle and integrity checks

Artifact	Format/Standard	Integrity Check	Public Location
Recording	MP4 + checksum	Hash verify on upload	Meeting page (canonical URL)
Caption file	WebVTT/SRT	Validator + human spot	Meeting page (linked)
Transcript	Tagged PDF/HTML	Accessibility checker	Meeting page (linked)
Agenda/minutes	Tagged PDF/HTML	Link audit	Legislative portal
Translations	Tagged PDF/HTML	Glossary alignment	Meeting page (linked)

8. Risk Register and Controls

Trigger	Audience	Message Elements	Channel	Owner	SLA
Caption outage >60 s	Residents	Status, workaround, ETA, corrections note	Banner on meeting page	Clerk/Comms	2 min
Interpreter dropout	Residents	Switching interpreter; timestamp of restoration	Lower-third + page note	Accessibility	2 min
Broken archive link	Residents	Acknowledgment; fix window; alternative link	Meeting page + social	Records/Web	Same day
Platform disruption	Residents	Simulcast info; phone-in option; recap plan	Banner + email list	Comms/IT	5 min

Triggers and Evidence

The risk register prioritized what residents notice first: audio intelligibility, caption latency spikes, interpreter dropouts, and broken links. Each risk had a trigger threshold, a prescribed first action, and an evidence artifact.

Table 8. Risk register with triggers and mitigations

Risk	Trigger	Mitigation	Owner	Evidence
Caption latency spike	>2 s for 60 s	Switch engine; verify audio path	Accessibility	Dashboard snapshot
Encoder failure	>1% dropped frames	Lower bitrate; switch to standby	AV/IT	Drill note; encoder logs
Interpreter echo	Operator report or viewer complaint	Verify mix-minus; adjust returns	Accessibility/AV	ISO sample
Broken links in bundle	Weekly audit finds issues	Repair; post corrections note	Records	Link report; corrections page

9. Procurement and Contracting

Award decisions emphasized resident outcomes—latency, accuracy, completeness—supported by small bake‑offs with the county’s own audio and agenda content.

From Features to Outcomes

Contracts were amended to protect portability and security: per-user roles with MFA, exportable logs, open artifact formats, change-control windows around marquee meetings, and no-fee artifact export at exit. Vendors were evaluated with a bake-off using real agenda content and room audio.

Table 9. Procurement clauses aligned to accessibility outcomes

Area	Minimum Standard	Verification	Risk Mitigated
Identity & roles	Per-user SSO/MFA; no shared admins	Access test; audit log	Account takeover; weak attribution
Logging & exports	Exportable logs; immutable retention	Sample export; policy	Opaque incidents; audit gaps
Formats & portability	WebVTT/SRT; tagged HTML/PDF; no-fee export	Artifact samples; contract	Vendor lock-in; inaccessible archives
Change control	Freeze windows around marquee meetings	Change log; clause	Regression during high-salience events
DPA & data use	No training on county data; residency controls	Signed DPA; console settings	Privacy & compliance risk

10. Training Drills and Culture

Daily preflights took under five minutes and prevented most failures. Quarterly failover drills rehearsed LTE profiles and standby encoders without disrupting operations.

Micro‑Drills that Fit Real Calendars

Short, frequent micro-drills built operator confidence without burdening calendars. A quarterly live-switch drill rehearsed standby encoder and LTE failover, and the drill timeline was attached to the meeting record to normalize transparency and learning.

Table 10. Drill cadence and success criteria

Drill	Cadence	Pass Criterion	Artifact
Daily preflight	Before doors open	All checks green in ≤5 min	30 s rehearsal clip; dashboard glance
Quarterly failover	Once per quarter	Standby in ≤60 s; no audio loss	Drill timeline; operator notes
Monthly link audit	Once per month	Zero broken links in last month	Audit report; repair log

11. Budget and TCO

Measure	Target (Tier A)	Current	Trend	Narrative / Next Action
Caption latency	≤2.0 s	1.7 s	↘ improving	Pinned engine; glossary refresh scheduled
Caption accuracy (sample)	≥95%	94%	↗ rising	Add post-edit on sensitive items
Interpreter uptime	≥99%	99.2%	→ stable	Hot-swap verified in Q3 drill
Archive completeness	100% within SLA	100%	→ stable	Link audit weekly; zero broken links
Complaint volume	↓ MoM	-18%	↘ improving	Corrections page widely referenced

Variance Reduction Story

The county framed the budget narrative as variance reduction: fewer emergency purchases, stabilized caption spend via flat-rate tiers, and reduced staff rework after publication. A conservative five-year TCO reflected predictable operating costs and lower PR risk.

Table 11. TCO components and savings levers

Component	Driver	Savings Lever	Verification
Licenses/services	Minutes, languages, seats	Flat-rate tiers; version pinning	Invoices; change log
Staff time	Meetings × minutes	Checklists; automation	Timesheets; queue metrics
Storage/egress	Media + captions growth	Lifecycle tiers; CDN	Usage reports
Training/drills	Turnover; cadence	Micro-drills; runbooks	Drill logs

12. Results and Resident Impact

Early declines in complaint volume and duplicate records requests signaled improved trust even before final KPIs reached target levels.

Leading Indicators

Within 90 days, caption latency stabilized, interpretation handoffs were reliable, and publication bundles appeared complete and on schedule. Complaints declined, and duplicate records requests dropped as residents learned to rely on the canonical meeting page. Local media cited the corrections page as a trust signal when issues occurred.

13. Lessons Learned and Replicability

Short checklists, canonical publishing, and per‑user identity policies scaled across departments without new headcount.

What Scaled Well

Begin with identity and publication; drill small and often; and write procurement to lock in what worked. These steps scaled across departments without new headcount and can be replicated by similarly resourced counties.

14. Frequently Asked Questions

Deputies, universal checklists, and laminated presets kept service levels stable despite staff changes.

Staffing and Turnover

Do we need new hardware? Often not. Start with governance, checklists, and drills; add targeted spares where the register shows risk.

How do we keep up quality under turnover? Deputies, universal checklists, and laminated presets minimize the impact of staffing changes.

15. Endnotes

Endnotes should cite your county’s accessibility policies, continuity guidance, and records schedules for audiovisual materials so successors and auditors can retrace decisions quickly.

16. Bibliography

Award decisions emphasized resident outcomes—latency, accuracy, completeness—supported by small bake‑offs with the county’s own audio and agenda content.

Convene helps Government have one conversation in all languages.

Engage every resident with Convene Video Language Translation so everyone can understand, participate, and be heard.

Schedule your free demo today:

Tagged White Paper