Every figure on this page links to the organization that produced it, and each was re-checked against that source before publishing. How federal agencies use video on their homepages, in our own read of every site in USA.gov's agency index on November 1, 2026; what government video costs and who makes it, from USAspending.gov, GAO and the Bureau of Labor Statistics; how many local governments stream their meetings and what state law requires; how federal video measures up to Section 508, from GSA's reports to Congress; how Americans get government information, from Pew Research Center and GSA's Digital Analytics Program; how much video police and courts produce, from the Bureau of Justice Statistics; who sells video to governments, from filings and FedRAMP; and where the popular public-sector video numbers came from.
Only 13.8% of readable federal agency homepages showed a video player on November 1, 2026, but 70.4% linked to a YouTube channel, in our own read of USA.gov's agency index. (Our read)See details
A federal homepage was 5.1 times as likely to link to YouTube as to play video itself, and only 4.2% linked to TikTok. (Our read)See details
45% of federal agencies' top-viewed public videos fully conformed to Section 508 in fiscal 2025, up from 34% in 2024 and 29% in 2023, though GSA says reporting changes may explain part of the rise; the most common defect was missing audio description controls. (GSA)See details
Federal agencies obligated $24.0 million on film and video production contracts in fiscal 2025, 46% less in real terms than in 2011-2013, in our USAspending count. (Our study)See details
U.S. local governments, excluding schools and hospitals, employed 90 film and video editors in May 2025, against 16,820 public relations specialists (BLS). (BLS)See details
23% of New York's cities, towns and villages streamed meetings in February 2023, and 13% of those under 5,000 residents. (NY Coalition for Open Government)See details
Texas requires video of meetings in cities of 50,000+ people, posted within seven days; Indiana has required a livestream or a recording kept online for 90 days since July 2025. (Texas Legislature)See details
62% of local police departments used body cameras in 2020; in 2016, 77% of agencies without them cited the cost of storing video. (BJS)See details
62% of clerks in a Granicus survey said they do no editing on public meeting videos. (Granicus)See details
The '95% retained from video' line in public-sector guides comes from an informal survey of about 200 business buyers, by Insivia's own account. (Insivia)See details
How do federal agencies use video on their homepages?
Mostly by pointing elsewhere. On November 1, 2026 we opened every website in USA.gov's A-Z federal agency index, 430 distinct URLs from 597 index entries, in a headless browser; 334 answered with a readable page and 92 refused an ordinary browser with a 403 or a challenge page.
Only 13.8% of the readable federal agency homepages showed a video player or a labelled click-to-play video, and 12.9% counting only visible player elements of at least 150 by 80 pixels. Just 3.9% played a muted autoplay background loop. For comparison, the same rule found video on most of the largest U.S. colleges' homepages the same day (higher education video statistics).
Links to social video are the norm instead. 70.4% of the readable homepages linked to a YouTube channel, 77.8% to X, 71.9% to Facebook, 57.8% to LinkedIn, 52.1% to Instagram and only 4.2% to TikTok. 25.4% linked to the agency's own video, multimedia or livestream page. Counting any of the three (a player, a YouTube channel link or a video page link), 74.9% of homepages pointed visitors to video. A federal homepage was 5.1 times as likely to link to a YouTube channel as to play a video itself.
When a federal homepage does carry video, it is usually YouTube: 54.3% of the 46 homepages with video embedded YouTube, and 41.3% served a plain video file from the agency's own site.
Other domains: .org, .com, .edu, .us, .net (22; small group)
18.2%
The military is missing from these shares: 30 of 42 .mil homepages in the index refused an ordinary headless browser, so read the figures as civilian agencies. The per-agency results are in homepages.csv and the totals in crawl-summary.json. A hand check of 30 randomly drawn classifications found all 30 right.
What does government video cost, and who makes it?
Federal agencies obligated $24.0 million on contracts coded film and video production in fiscal 2025, in our federal video production contracts study of USAspending.gov records. Adjusted for inflation, that spending averaged 46% less in fiscal 2023-2025 than in 2011-2013, and fiscal 2025 had 56 new awards of $25,000 or more, the fewest since 2008. The Defense Department placed 41.4% of fiscal 2021-2025 video production contract dollars, and small businesses won 79.7% of them.
Advertising and public relations contracts are far larger. Agencies obligated $1.70 billion on them in fiscal 2025, the Defense Department paid for 64.6% of fiscal 2021-2025 dollars, and Omnicom Group companies received 25.7% (federal advertising contracts study). That matches the older pattern: the Government Accountability Office found advertising and PR contract obligations averaged close to $1 billion a year over fiscal 2006-2015, with Defense responsible for over 60 percent (GAO, 2016). An earlier GAO review counted $1.62 billion in media contracts at seven departments from fiscal 2003 to mid-2005, $1.4 billion of it to advertising agencies (GAO, 2006). The Congressional Research Service put fiscal 2013 advertising contracts at $892.5 million (CRS); it counted more contract codes than our study's fiscal 2013 figure.
Government is a small customer for the agencies themselves. In the 2022 Economic Census, the federal government supplied 2.7% and state and local governments 4.5% of U.S. public relations agencies' revenue, and 1.0% and 2.4% of advertising agencies' revenue (U.S. Census Bureau).
In-house staff
The federal civilian workforce had 726 audiovisual production specialists in August 2026, down 22.4% since December 2024, and public affairs staff fell 26.8% over the same months (federal video workforce study, from OPM data). Outside Defense, the median federal audiovisual production specialist earned $129,904 a year. GAO counted federal public affairs staff peaking at 5,238 in fiscal 2011, with combined salaries near $500 million and a median of about $90,000 in fiscal 2014 (GAO).
The Bureau of Labor Statistics counts government video jobs by occupation (May 2025, excluding schools and hospitals):
Occupation
Federal executive branch
State government
Local government
Public relations specialists, jobs
6,570
8,030
16,820
Median annual wage
$115,210
$76,130
$77,870
Producers and directors, jobs
820
510
930
Median annual wage
$113,070
$75,120
$74,790
Camera operators, jobs
230
50
500
Median annual wage
$85,340
$65,310
$52,810
Audio and video technicians, jobs
not published
260
2,430
Median annual wage
not published
$62,750
$61,820
Film and video editors, jobs
not published
not published
90
Median annual wage
not published
not published
$60,300
Source: BLS Occupational Employment and Wage Statistics, industry-specific estimates. BLS publishes no estimate where a cell is too small or unreliable. Ninety salaried film and video editors across all local governments is a sign of how much of that work goes to contractors and freelancers rather than staff.
Contractors charge by the hour on federal schedules: video editors on GSA Multiple Award Schedule contracts have a median ceiling rate of $97.47 an hour, and motion graphics designers and animators $124.43 (GSA video production rates study). Ceiling rates are maximums, not prices paid.
How many local governments stream or record their meetings?
No national count exists; the best figures are state surveys. In February 2023, 23% of New York's cities, towns and villages outside New York City were streaming their meetings, although 581 had streamed at some point since the pandemic (New York Coalition for Open Government). Size decides it: only 13% of New York places under 5,000 residents streamed, and 21% of towns. In Michigan in fall 2012, 14% of local governments streamed or posted meeting video, 36% of cities, and 51% of cities used cable broadcasts or streaming (University of Michigan CLOSUP). In a 2010 survey of U.S. cities of 25,000 to 250,000 people, 40% webcast video of key meetings and 37.5% used both television and webcasts (Science, Technology, and Environmental Policy Lab); the share of mid-size city departments using YouTube-type tools rose to 46% by 2012 (Feeney, University of Illinois Chicago).
There is a lot to cover. The 2022 Census of Governments counted 90,837 local governments, and the 2025 count was 91,438: 3,031 counties, 19,489 municipalities, 16,184 townships, 12,535 independent school districts and 40,199 special districts (U.S. Census Bureau).
Few residents attend either way: in 2025, 13% of U.S. adults said they had attended a local government meeting such as a city council or school board meeting in the past year (Pew Research Center), and 14% in a second 2025 Pew panel reading (Pew). One online audience can outnumber the room many times over: the Buffalo Common Council had 18,000 people watch one meeting online (NYCOG). The cost can be small: one New York school board streamed by paying a music teacher a $2,798 yearly stipend in 2023 (NYCOG).
What state law requires
State
Rule
Source
Texas
Video and audio of regular meetings for cities of 50,000+ people, counties of 125,000+ and school districts of 10,000+ students; posted within seven days and kept online two years; a video-sharing or social networking site counts; junior colleges over 20,000 students must broadcast
School, community college and university boards post meeting video within seven days; other bodies may use sound, video or digital recording or written minutes
Indiana's rule is the newest, and nearly all of its 567 localities were complying by November 2025; Seven other states required meetings to be recorded and archived, according to the Indiana Capital Chronicle. In England, 95 per cent of councils answering a 2023 Local Government Association survey wanted the power to hold virtual meetings again (LGA).
Government-access television
Before streaming, meetings reached residents on cable. Subscribers to more than 1,500 U.S. cable systems had public, educational and governmental (PEG) access channels, and one estimate put the channels at more than 5,000 (Congressional Research Service). More than 100 access centers closed between 2005 and about 2011 as state franchising laws ended local cable contracts. In 2021, researchers citing the Alliance for Community Media counted about 3,000 channels, against about 2,500 PEG operations in the 1980s (Nieman Lab); the two counts measure different units, so they do not show a clean decline. Federal law caps franchise fees, which fund many of these stations, at 5 percent of cable revenue (47 U.S.C. 542).
Two stations show the shift. Vancouver and Clark County's CVTV in Washington produced 389.05 hours of live programs and 424.80 hours of original programming in 2022, and its 2023 operating budget was $762,084 (CVTV). Seattle Channel had nearly three million online views in 2024 (Seattle Channel) but faced a $1.6 million cut in 2025 as cable fees fell (Seattle City Council).
What does accessibility law require of government video, and do agencies comply?
Most federal video still fails the federal standard. In fiscal 2025, 45% of the top-viewed public videos that federal agencies tested fully conformed to Section 508, up from 34% in fiscal 2024 and 29% in fiscal 2023 (GSA, Governmentwide Section 508 Assessment). The tested pools differ each year, so the trend is indicative. 40 of the 60 agencies reporting in fiscal 2025 tested their top videos at all, and the most common defect was missing Audio Description Controls (36 CFR 1194 503.4.2).
The wider picture is similar. The fiscal 2025 assessment covered 212 agencies and components, with Sixty agencies submitting conformance data; the governmentwide conformance score was 1.96 out of 5, after falling from 1.79 to 1.74 in fiscal 2024. Only 27% of agencies required Section 508 training, fewer than 30% almost always checked deliverables for conformance, testing and remediation was the weakest factor at 2.00, and 24 of 60 agencies had no way to track test results. Agencies reported owning or operating about 1,683,529 public web pages in fiscal 2025, and only 38% of the electronic documents they tested conformed. The fiscal 2024 report drew on 245 reporting entities; 39 said they had no videos at all, up from 24, which GSA flagged, and the average entity had 5.5 full-time Section 508 staff.
The rules themselves are specific about video. Section 508 covers electronic and information technology regardless of the type of medium (29 U.S.C. 794d), and GSA has had to report compliance to Congress every year since the December 29, 2022 appropriations act. The Revised 508 Standards require hardware that plays video with synchronized audio to show closed captions (section 413.1) and software players to offer caption and audio description controls (section 503.4) (36 CFR 1194). They also cover internal content such as Educational or training materials, so an agency's internal training video must be accessible too. Compliance has been required since January 18, 2018 (Federal Register); unaltered older material that met the 2000 standards keeps a safe harbor from the same January 18, 2018 date. The Access Board put the yearly cost of the revision at 182.4 million dollars (regulatory impact analysis).
State and local governments fall under the Justice Department's ADA Title II web rule instead; its WCAG 2.1 AA requirement and its 2027 and 2028 compliance dates are covered on our video captions statistics page. The Plain Writing Act of 2010 gave agencies 1 year to write public documents plainly and 18 months to start annual compliance reports (Public Law 111-274).
Video is also a federal record. The National Archives lets agencies destroy routine audiovisual records once they are 2 years old (General Records Schedule 6.4), while permanent audiovisual records move to the Archives 5-10 years after creation (36 CFR 1237), including agency-produced public information films (NARA).
How do Americans get government information online?
Increasingly from social media and from government itself. In January 2024, 68% of U.S. adults said they often or sometimes get news about local government and politics, 54% from social media and 32% from local government websites; among adults under 30, 71% got it from social media (Pew Research Center). By December 2025, 40% said they get local news from local government agencies or officials at least sometimes, while only 21% followed local news very closely (Pew).
Trust is low and uneven. In September 2025, 17% of adults trusted the federal government to do what is right always or most of the time, down from 22% in 2024, and just 9% of Democrats did (Pew). A 2023 Pew survey found 61% viewed their local government favorably. Views of agencies differ widely: in Pew's survey of 5,062 adults in September and October 2026, eight of 12 agencies were viewed favorably on balance, from 73% for the National Park Service to 47% for the CDC and 37% for the IRS (Pew). Across 30 OECD countries, 39% had high or moderately high trust in their national government in November 2023, rising to 69% among people who feel they have a say (OECD).
Federal websites
GSA's Digital Analytics Program tracks 6,000+ public federal websites (Digital.gov). Mobile devices brought 58% of their visitors in 2023, up from 13% in 2013, while desktop fell to 40%; 47% arrived from organic search, and visitors downloaded close to 4 billion PDFs over the decade and spent about 7.7 billion days on the sites. More than 4,000 federal staff use the program, up from 62 participating agencies in 2019 (Performance.gov). In the 30 days to October 21, 2026, tracked sites logged 1,275,641,811 visits, led by tools.usps.com with 265,900,870; irs.gov had 40,430,256 (analytics.usa.gov). In the 30 days to February 4, 2025, 142,746,958 of 251 million users were on mobile (DAP). GSA's Site Scanning Program checks about 26,000 federal websites daily (GSA).
Video is a small part of that traffic. DAP's video report for the 30 days to late March 2025 put a short NASA mission clip first with 94,675 starts, and the top Defense Department file had 36,177 (analytics.usa.gov). The leading NASA clips look like autoplaying page loops, so starts are not deliberate views.
The rules push agencies toward this. The 21st Century IDEA of 2018 (115-336) requires new and redesigned federal websites to work on mobile, gave agencies 2 years to digitize public forms and required progress reports for 4 years (Public Law 115-336). OMB memo M-23-22 of September 2023 gave agencies 180 days to meet most of its website requirements, required them to find the pages carrying the top 80 percent of traffic, gave 30 days to name a digital experience lead (OMB) and made Digital Analytics Program participation mandatory for all executive agencies (Digital.gov).
Satisfaction with federal services fell to 68.6 on ACSI's 100-point scale in fiscal 2026 from 70.4 in fiscal 2025, in a study of 7,932 surveys (ACSI); the website satisfaction index went from 72 to 71, and website quality to 75. ACSI sells its research to agencies.
Outside Washington, few local governments even use .gov addresses: 12,636 .gov domains were held below the federal level, only 10.6% of eligible local governments had one, counties led at 78.5%, and only 31 school districts and 194 of 574 federally recognized tribes had registered (UC Berkeley CLTC, 2025). Households are reachable: 90% had broadband in 2021 and smartphones were in 90% of them (U.S. Census Bureau). OECD reports 47% of governments' key datasets are openly available (OECD).
Agencies that invest in video find big audiences. NASA's flagship accounts had 197 million followers across platforms in 2022, its broadcast of the DART asteroid impact was viewed more than 5 million times, and by June 2025 its footprint had grown to over 400 accounts on 15 platforms, which it began to cut back (NASA).
How much video do police and courts produce?
More than any other part of government. About 62% of U.S. local police departments used body-worn cameras in 2020, a 43% rise from 2016, and 79% of local officers worked in departments that used them (Bureau of Justice Statistics). Every department serving 1 million or more residents used them, and use in departments serving 250,000 to 499,999 residents reached 94%. Sheriffs' offices were at 68%. Local departments had one camera for every 1.7 officers, sheriffs' offices one for every 2.4 deputies.
Other video is common too: in 2020, 61.7% of local police departments used cameras in patrol cars, 67.6% of sheriffs' offices did, and 11.6% of local departments used cameras on drones (BJS).
The 2016 baseline shows how fast this grew. That year 47% of the 15,328 general-purpose law enforcement agencies had body-worn cameras, 7,259 agencies in all, against 69% with dashboard cameras (BJS). Local departments had about 90,000 body cameras in service, and about 80% of the largest departments had them (BJS). Of agencies with cameras, 60% of local police departments had fully deployed them, 86% had a formal policy, and 93% of policies that named events to record required officers to record traffic stops. Among agencies without cameras, 77% cited the cost of storing and disposing of video. Public demand for footage was modest: 63% of agencies with cameras averaged no public requests.
Federal money helped. The Justice Department's body-worn camera program gave over 41 million dollars in its first two years, after applicants requested over 85 million dollars for more than 86,000 cameras (Bureau of Justice Assistance); fiscal 2015 awards alone were over 18 million dollars to 70 sites (BJA). The fiscal 2024 program capped awards at 2 million dollars each (BJA).
Federal agents' own cameras have been policy, then not. A June 2021 Deputy Attorney General memo gave ATF, DEA, FBI and the Marshals 30 days to submit body camera policies (DOJ). Executive Order 14074 gave federal law enforcement agencies 90 days to issue camera policies in 2022 (Federal Register); Executive Order 14148 rescinded order 14074 on January 20, 2025 (Federal Register).
States set the release and retention rules. As of NCSL's April 2021 update, Eight states mandated body cameras statewide (NCSL); that count is a 2021 baseline. Illinois finished phasing in its mandate on January 1, 2025 and requires recordings to be kept 90 days (Illinois General Assembly). Since 2019, California has let agencies delay release of critical-incident video by up to 45 days (California Penal Code); Colorado required cameras by July 1, 2023 and release of recordings within 21 days of a complaint (Colorado SB20-217).
The evidence on effects is mixed. A 2020 Campbell systematic review of 30 studies found body cameras cut citizen complaints by about 17%, with no consistent effect on officers' use of force (Lum et al.).
The companies behind the cameras
Axon, the largest supplier, sold $397.0 million of body cameras and accessories in 2025, $265.9 million of in-car, fixed and drone camera systems, and $1.20 billion of software and services, mostly digital evidence storage (Axon 10-K). Its recurring cloud revenue reached $1.3 billion, and $2.31 billion of its revenue came from the United States. Motorola Solutions' Video business sold $2,120 million in 2025 (Motorola Solutions 10-K).
Courts
Federal courts still keep cameras out. Fourteen district courts took part in the 2011-2015 cameras pilot, recording 158 proceedings in 757 video files that were accessed 21,530 times in 2014 (Federal Judicial Center). An audio streaming pilot ended March 31, 2023 (U.S. Courts), and since Sept. 22, 2023 judges may allow live audio of civil and bankruptcy non-trial proceedings (U.S. Courts). The Supreme Court has recorded its arguments' audio since 1955, has released it within the same term since the 2010 Term, and allows no video (Supreme Court).
Who sells video technology to governments?
A few specialists dominate meeting video, while general video platforms treat government as one market among many and do not break it out. Granicus, a private company that sells meeting-video, email and website software to governments, says on its website in 2026 that it serves more than 7,000 public-sector organizations and delivers 12,000 government video streams a day (Granicus). It also claims 30 billion digital interactions a year, more than 19B messages sent by 250,000 administrators, and 2,500 hosted government websites; its GovDelivery email service reached over 100 million people in 2015 (Granicus). Swagit, a meeting-video service that films and streams meetings for local governments, is now part of Granicus (Granicus). In a 2023 Granicus case study, Burleson, Texas held 50 council and planning meetings a year on that service (Granicus).
CivicPlus, which sells local-government websites and meeting software, said in October 2026 that it had more than 13,000 customers, up from over 10,000 in January 2025 (CivicPlus); its claim of 340M resident users is close to the whole U.S. population, so read it as reach, not use. OpenGov, which sells budgeting software, claims 2,000+ agency customers (OpenGov). BoxCast, which sells streaming to local governments, churches and schools, says its platform reached 42.7M viewers in 2025, a platform-wide figure (BoxCast).
Listed companies give harder numbers but no government split. Tyler Technologies, the largest public-sector software company, had $2.33 billion of revenue in 2025, 87% of it recurring (Tyler 10-K), and describes a market of about 36,000 cities and towns. Brightcove had 2,559 customers, government agencies among them, and $201.2 million of revenue in 2023 (Brightcove 10-K) before Bending Spoons took it private on February 4, 2025 (8-K). Kaltura had $180.9 million of revenue in 2025 and serves various government entities (Kaltura 10-K). Zoom reported $4,868.8 million in fiscal 2026 and says it is FedRAMP authorized (Zoom 10-K).
Federal agencies can only use cloud video tools that pass FedRAMP. On GSA's FedRAMP Marketplace:
Live records read November 1, 2026; counts change as agencies add authorizations.
What vendors' own surveys say
Vendor surveys are the only recent data on government communicators, so read them as vendor data. In a 2024 Granicus survey of clerks, 62% said they do no editing on public meeting videos (Granicus); 56% had been in their role under five years and 25% were re-examining or delaying projects. Its 2026 State of Digital Government report draws on nearly 1,300 respondents (Granicus), and its 2026 public communications report on more than 550 government professionals, of whom 68% were optimistic about AI (Granicus), with a 15% yearly rise in communicators using AI tools. In its civic engagement survey, 81.1% of practitioners cited limited resources (Granicus), and in its 2026 accessibility research 69 percent of agencies sat in the two earliest stages of its accessibility model (Granicus). Granicus publishes sample sizes but not field dates. It also says its meeting tools raise public engagement by as much as 250% (Granicus), a marketing claim with no published method.
What the popular public-sector video numbers really say
Local-government communications blogs and vendor posts repeat a small set of video statistics. We traced each to where it started.
"Viewers retain 95% of a message in video, 10% in text"
An informal agency survey, not a study. A UK public-sector video agency tells public bodies they can expect viewers to retain 95% (Zero Above). The line belongs to Insivia, a marketing agency, which now says on its own site that the 95% figure came from a small, informal survey of around 200 business buyers in the late 2000s, not a study (Insivia). The earliest dated copy is Insivia's own 2013 list, with the 95% line and no source (Insivia). Nothing in it concerns residents. Use instead: nothing; no study of government audiences measures retention this way.
"A million minutes of video every second" and "82% of traffic"
Old Cisco forecasts, not measurements. A town communicator told local-government peers on ELGL in 2019 that a million minutes of video would cross the internet every second (ELGL). Both figures are Cisco forecasts, not measurements: Cisco's 2017 white paper projected a million minutes a second by 2021 (Cisco). Cisco gave its 82% two different bases: 82 percent of consumer internet traffic in the white paper, and 82 percent of all IP traffic in its press release (Cisco). In 2018 it moved the 82 percent to 2022 (Cisco), now four years past. Use instead: a current measurement of your own audience, or say plainly that it was a forecast.
"One third of all online activity is watching video"
No original found. Two ELGL posts repeat it: a 2016 vendor guest post (1/3) (ELGL) and the 2019 Collierville column (One third) (ELGL). We found no original; the oldest copy is Insivia's unsourced 2013 list (1/3) (Insivia). Unsupported.
"Video in email raises click-through 200-300%"
Overstated: the 2009 source said two to three times. The 2016 ELGL post, by the head of a firm selling video to cities, promises a 200-300% increase (ELGL). The source is Forrester's May 2009 report, built on interviews with marketing executives, which said video emails got two to three times the click-through (Forrester). Two to three times is an increase of 100-200%, not 200-300%, and the figure describes 2009 marketing email.
"57 seconds on a text site, 6 minutes with video"
No original found. The same 2016 post says users spend 57 seconds on a text-only site (ELGL). We found no original. The nearest measurement, Wistia's read of its own top 100 pages from January 2021 to April 2022, found about 6 minutes on pages with video against 4.3 minutes without, about 1.4 times, not six (Wistia).
"20% of viewers leave in 10 seconds"
A 2010 benchmark of entertainment and ad clips. ELGL readers were told in 2016 that twenty percent of viewers click away in 10 seconds (ELGL). That comes from Visible Measures' September 2010 benchmark of over 40 million short entertainment and advertising clips: nearly 20% left within 10 seconds and 33% within 30 (Visible Measures). It says nothing about meeting video or public-service messages, and it is 16 years old.
"Video gets 27% more engagement on social networks"
A 2010 Facebook brand-post study, trimmed. Granicus told government readers in 2010 that video gets 27% more engagement on social networks (Granicus on GovLoop). The source was a Vitrue study of some 32,000 Facebook brand posts from May to August 2010 (Chief Marketer), which found video got 27% more engagement than text, but images beat video by 22%, the part left out. A real public-sector comparison: a 2016 University of Sydney study of 20 Australian public health agencies found video posts got 25 per cent more likes than photos, which its author put down to Facebook's algorithm (University of Sydney).
"Malmö boosted engagement 1,300%"
One page, not the site. A case-study site headlines that the City of Malmö, Sweden, boosted website engagement 1,300% (CaseStudies.com). Brightcove's own story says video raised engagement 1,300 percent "in one case", comparing one video's views with one text page (Brightcove), and the city's video manager describes 1,300 percent as a difference measured in Brightcove's analytics. It is one page, not the site.
"Microsoft found an 8-second attention span"
Not Microsoft research. The 2019 Collierville column opens with an 8 seconds attention span from "a Microsoft study" (ELGL). Microsoft Canada's 2015 report was a gamified online survey of 2,000 Canadians, and its 8-second and goldfish figures are credited on the page to Statistic brain, a statistics website, not to Microsoft's own research (Microsoft Canada).
We started from the questions people search for about government video statistics and looked for the organization that actually produced each answer. Every figure on this page sits in an evidence ledger with its source, the exact table or paragraph it came from, the measurement period, the population measured, and the date we checked it. This edition holds 328 figures from 65 publishers.
On November 1, 2026, each of the sourced figures on this page was re-fetched with the fetch cache cleared and matched against the recorded wording: GSA's Section 508 assessments, Digital Analytics Program reports and FedRAMP records, eCFR, the Federal Register and govinfo, Bureau of Justice Statistics and Bureau of Justice Assistance reports, BLS Occupational Employment and Wage Statistics (through the BLS API), the 2022 Economic Census and the Census of Governments, GAO and CRS reports, nine state legislatures' statute pages, the U.S. Courts and the Supreme Court, the 10-K filings of Axon, Motorola Solutions, Tyler Technologies, Brightcove, Kaltura and Zoom on SEC EDGAR, surveys by Pew Research Center, ACSI, OECD, the New York Coalition for Open Government and the University of Michigan, vendors' own pages, and the original pages behind each popular claim. gao.gov, congress.gov, oecd.org, justice.gov, civicplus.com and several state legislature sites refuse scripts, so those rows were read through Internet Archive captures or official mirrors. Our own figures come from a headless-browser read of the homepages in USA.gov's federal agency index on the same day; 92 of 430 refused the browser and are left out. Federal contract and workforce figures are carried from our four federal studies. We left out Granicus's full reports (behind a form), the Defense Media Activity budget books (403), a current national count of streamed local meetings (none exists) and a New York videoconference rule whose 2026 status we could not confirm.
We only cite the producer: government agencies, regulators, company filings and official posts, and the survey owner's own report. We do not cite statistics roundups, Statista-style aggregators, or salary sites without a public method. We did not bypass paywalls or sign-up forms; where a full report was gated, we used only the figures its publisher states publicly.
Figures marked "our calculation" are arithmetic on the cited numbers, and anyone can reproduce them:
This edition contains no Mark Studios customer data.
Limitations
No national count of streamed meetings exists. The local figures are state surveys (New York 2023, Michigan 2012) and a 2010 survey of mid-size cities.
Most data on government communicators is vendor data. Granicus and CivicPlus sell to governments and publish sample sizes but not field dates.
Our crawl misses most of the military. 30 of 42 .mil homepages and 62 other federal homepages refused an ordinary browser; one visit to one page per agency; video behind unlabelled controls is missed.
Section 508 video results cover tested videos only. Agencies chose and tested their own top videos, and the pool differs each year.
Police camera counts are from 2016 and 2020. BJS has not published a newer national count.
BLS counts only salaried jobs. Contractors and freelancers who make government video are not in the occupation counts.
How to cite this page
For a single figure, cite the original publisher named next to it, and link here if the verification trail is useful to your readers. Each figure has its own anchor, so you can link straight to it, for example https://www.markstudios.com/research/government-video-statistics#DRV-004. To cite the page as a whole:
Mark Studios. "Government Video Statistics." Last verified Nov 1, 2026. https://www.markstudios.com/research/government-video-statistics
APA
Mark Studios. (2026, November 1). Government video statistics. https://www.markstudios.com/research/government-video-statistics
MLA
“Government Video Statistics.” Mark Studios, 1 Nov. 2026, www.markstudios.com/research/government-video-statistics.
Chicago
Mark Studios. “Government Video Statistics.” Last modified November 1, 2026. https://www.markstudios.com/research/government-video-statistics.
BibTeX
@misc{markstudios_government_video_statistics_2026,
author = {{Mark Studios}},
title = {{Government Video Statistics}},
year = {2026},
month = nov,
url = {https://www.markstudios.com/research/government-video-statistics},
note = {Last verified 2026-11-01}
}
Our calculations and the evidence table are free to reuse with attribution and a link to this page. Each underlying figure remains the claim of its original publisher, and some publishers ask for a link when their figures are reused.
We re-verify every figure quarterly and when a major source publishes a new edition. The next scheduled review is February 2027. Corrections go to contact@markstudios.com.
Nov 1, 2026: First published with 328 figures from 65 publishers, each checked against its original source.
Nov 1, 2026: Three takeaways tightened: GSA's caution that reporting changes may explain part of the rise in Section 508 video conformance; the BLS local-government count excludes schools and hospitals; Indiana requires a livestream or a recording kept online for 90 days, not a livestream alone.
Frequently asked questions
How many federal agencies use video on their websites?
Few put it on the homepage. On November 1, 2026, 13.8% of the 334 readable homepages in USA.gov's federal agency index showed a video player in our read, while 70.4% linked to a YouTube channel and 25.4% to their own video page.
How much does the federal government spend on video production?
Federal agencies obligated $24.0 million on contracts coded film and video production in fiscal 2025, and $1.70 billion on advertising and public relations contracts, in our counts of USAspending.gov records.
Do government videos have to be captioned?
Yes. Section 508 requires federal video players to offer caption and audio description controls (section 503.4), and in fiscal 2025 only 45% of agencies' top-viewed public videos fully conformed.
Do cities have to livestream council meetings?
Only in some states. Texas requires video recordings for cities of 50,000 or more people, posted within seven days, and Indiana requires a livestream or a recording kept online for 90 days; most states have no streaming rule.
How many local governments stream their meetings?
No national count exists. In New York in February 2023, 23% of cities, towns and villages were streaming, although 581 had streamed at some point since the pandemic.
How many police departments use body cameras?
About 62% of U.S. local police departments and 68% of sheriffs' offices used body-worn cameras in 2020, the latest national count from the Bureau of Justice Statistics.
Who makes government video?
Mostly contractors. All U.S. local governments employed 90 salaried film and video editors in May 2025, and federal contract dollars coded video production went 79.7% to small businesses in fiscal 2021-2025.
Do people really remember 95% of a video message?
There is no study behind it. Insivia, the agency the 95% line is credited to, says it came from an informal survey of about 200 business buyers in the late 2000s.
