Every figure on this page links to the organization that produced it, and each was re-checked against that source before publishing. How founders start, fund and sell, and where video fits: our own count of video on the homepages of 608 Y Combinator startups; business applications, startups and survival rates from the Census Bureau and BLS; who founders are; how they pay for a new business, from Census, Kauffman and Federal Reserve surveys; venture, crowdfunding and accelerator figures from the SEC, PitchBook-NVCA, Crunchbase and YC; pitch videos and pitch decks; what B2B buyers and the public say about founder and executive content; and where five popular startup statistics came from.
On November 1, 2026, 39.8% of the 608 readable homepages of Y Combinator's Spring, Summer and Fall 2026 startups had a video player, but only 21.5% had a video a visitor would click to play; 26.5% ran a muted background loop (our count). (Our count)See details
More of those startups ask for a meeting than show the product: 43.1% had a demo, sales or calendar call to action, 54.3% named the founders or linked a team page, and 7.1% linked to a YouTube channel (our count). (Our count)See details
534,947 business applications were filed in the U.S. in September 2026; Census projects 5.4 employer businesses per 100 of them within a year, by our calculation. (U.S. Census Bureau)See details
77.9% of new U.S. establishments survive their first year, 51.4% five years and 34.7% ten; "90% fail" is close only after three decades (87.4% of the 1994 cohort had closed by 2025). (U.S. Bureau of Labor Statistics)See details
64.5% of U.S. employer firms were started with the owners' savings and 0.5% with venture capital (Census, 2017). (U.S. Census Bureau)See details
Y Combinator asks for a 1-minute founder video; in 1,139 accelerator application videos from 2010-2019 the median ran 68 seconds. (Y Combinator, NBER)See details
67% of people in 28 countries say the leader they would trust speaks to the public without anyone editing their words, but what employees say shapes opinion of a CEO more (51%) than the CEO's own social posts (19%) (Edelman, 2026). (Edelman)See details
Bain's "founder-led companies outperform 3.1 times" covers S&P 500 companies in 1990-2014; its December 2025 update puts it at 2.1 times over the past decade, and none of it is about startups. (Bain & Company)See details
How many startups put video on their homepage?
About two in five, but most of it is background motion rather than a video anyone presses play on. On November 1, 2026 we opened the homepages of every company in Y Combinator's Spring, Summer and Fall 2026 batches, 619 companies listed on YC's public directory, in a headless browser. 608 were readable. 39.8% of the readable homepages had a video player, but only 21.5% had a video a visitor would click to play; 26.5% ran a muted background loop, and for 45.9% of the homepages with video, a background loop was all there was. Nobody else publishes this count; the method and every company's result are below.
On the homepage (YC Spring-Fall 2026 batches, Nov 1, 2026, n = 608 readable)
Startups host their own video. Self-hosted files carried the video on 91.3% of the homepages that had one; other hosts such as Mux and Cloudflare Stream appeared on 2.1%. Asking for a meeting is more common than showing the product: a demo, sales or calendar call to action appears on more homepages than a click-to-play video, and homepages with a book-a-demo button were no more likely to have a video (38.8%) than those without.
By YC industry label (groups of 30+ readable homepages)
Video player
Book-a-demo call to action
B2B (n = 326)
37.1%
42.3%
Industrials (n = 127)
51.2%
9.4%
Healthcare (n = 50)
40.0%
32.0%
Fintech (n = 49)
24.5%
32.7%
Consumer (n = 36)
50.0%
8.3%
By batch, 45.0% of Spring 2026 homepages had a video player, 35.4% of Summer 2026 and 39.8% of Fall 2026. On YC's own directory, 2.4% of the 619 listings showed a public application or Demo Day video.
How we counted: one visit per homepage at desktop size, no clicks, robots.txt honored. 11 homepages could not be read: none showed a bot wall, 2 returned an error or blank page, 5 were one-screen placeholders, 2 had no website listed, 1 was disallowed by robots.txt and 1 had an unreadable robots.txt. A person checked 30 random homepages against the automated flags and agreed on the video and book-a-demo flags on 100.0% of them; the split between background and click-to-play video matched on 86.7% of the 15 sampled pages with a player. A video further down the site, or one that loads after a click, is not counted. Every company's result is in homepages.csv.
How many new businesses are started each year in the U.S.?
About half a million business applications are filed every month. The Census Bureau counted 534,947 applications for an employer identification number in September 2026, seasonally adjusted (U.S. Census Bureau), up from 531,728 in August (U.S. Census Bureau). In the same series, September 2019 had 290,721 (U.S. Census Bureau) and September 2005 had 183,543 (U.S. Census Bureau), so September 2025's 512,569 was 1.76 times the pre-pandemic September, by our calculation from the same series.
Most applications never become an employer. Census rates 147,331 of September 2026's applications as high-propensity, meaning likely to hire (U.S. Census Bureau), 27.5% of the total. It projects that 29,022 new employer businesses will form from that month's applications within four quarters (U.S. Census Bureau) and 39,703 within eight (U.S. Census Bureau). That is 5.4 employer businesses per 100 applications within a year, by our calculation.
Business applications, U.S., seasonally adjusted (Census BFS)
Counted as firms that actually hire, the number is flatter. The Business Dynamics Statistics show 508,777 startup employer firms (firm age 0) in 2023 (U.S. Census Bureau), against 521,312 in 2022 (U.S. Census Bureau), 453,027 in 2019 (U.S. Census Bureau), 518,727 in 2005 (U.S. Census Bureau) and 488,103 in 1978 (U.S. Census Bureau). Those 2023 startups employed 2,440,753 people (U.S. Census Bureau) and were 9.1% of the 5,593,727 U.S. employer firms that year (U.S. Census Bureau), by our calculation.
The Kauffman Foundation, which measures new entrepreneurs from the Current Population Survey, put the 2025 rate at 0.36% of U.S. adults starting a business each month (Ewing Marion Kauffman Foundation), and 83.3% of them started out of opportunity rather than necessity (Ewing Marion Kauffman Foundation). Startups created 5.3 jobs per 1,000 people in their first year in 2025 (Ewing Marion Kauffman Foundation).
Most businesses never hire at all. Census counted 31,985,874 nonemployer businesses in 2024 (U.S. Census Bureau), up from 27,104,006 in 2019 (U.S. Census Bureau). The SBA Office of Advocacy counts 36,207,130 small businesses, 82.3% of them without employees (SBA Office of Advocacy).
Florida, where Mark Studios is based, had 58,694 business applications in September 2026 (U.S. Census Bureau), 3,427,341 nonemployer businesses in 2024 (U.S. Census Bureau), and 83,028 private-sector establishments that opened in the year to March 2025 (U.S. Bureau of Labor Statistics).
How many new businesses survive?
About four in five new U.S. establishments are still open a year later, about half after five years and about a third after ten. Of private-sector establishments that opened in the year to March 2024, 77.9% were still operating in March 2025 (U.S. Bureau of Labor Statistics). For the cohort that opened in the year to March 2020, 51.4% were open five years later (U.S. Bureau of Labor Statistics), and of the year-to-March-2015 cohort, 34.7% were open ten years later (U.S. Bureau of Labor Statistics). BLS counts establishments (single locations, including new branches of existing firms), not startups or firms.
Establishments still operating (BLS Business Employment Dynamics, U.S. private sector)
The pattern has barely moved in three decades. Averaged over the 1994-2022 cohorts, the SBA Office of Advocacy puts five-year survival at 49.2% (SBA Office of Advocacy) and ten-year survival at 33.9% (SBA Office of Advocacy). Only over a working lifetime does closure approach the popular "90%": 12.6% of establishments that opened in the year to March 1994 were still operating 31 years later, in March 2025 (U.S. Bureau of Labor Statistics), so 87.4% had closed, by our calculation.
Counted as firms rather than locations, Kauffman puts the one-year survival of new U.S. firms at 77.9% in 2025 (Ewing Marion Kauffman Foundation), and Census recorded 106,493 one-year-old employer firms shutting down in 2023 (U.S. Census Bureau). Openings and closings run close together: about 1.3 million establishments opened for the first time in 2023 and about 1.2 million closed (SBA Office of Advocacy), and startups were 14.2% of establishments in 2023, up from 12.5% in 2019 (SBA Office of Advocacy). In the year to March 2025, 980,333 private-sector establishments opened (U.S. Bureau of Labor Statistics), and openings added 1.5 million jobs in the first quarter of 2026 (U.S. Bureau of Labor Statistics).
Venture-backed startups are measured differently, by what investors get back. Cambridge Associates' data on 27,259 venture-backed startups from 1990 to 2010 put the share returning 1x or less at no more than 60% in any year since 2001 (Fortune), with a peak of 79% in 2000 (Fortune). The page's last section traces the "90%" to its source.
Who starts businesses in the U.S.?
Founders are older than the startup image suggests. In the Census Bureau's Annual Business Survey for 2023, 23.1% of owners of U.S. employer firms were 65 or over (U.S. Census Bureau) and only 5.3% were 25 to 34 (U.S. Census Bureau). Across 2.7 million founders who started U.S. companies in 2007-2014 that went on to hire, the mean age at founding was 41.9 (National Bureau of Economic Research). Kauffman's 2025 rate of new entrepreneurs was lowest for adults aged 20-34, at 0.26% a month, against 0.43% for both the 35-44 and 45-54 groups (Ewing Marion Kauffman Foundation).
Who else: Kauffman's monthly rate of new entrepreneurship in 2025 was 0.60% for immigrants, about double the 0.30% for people born in the U.S. (Ewing Marion Kauffman Foundation), and 0.28% for women against 0.44% for men (Ewing Marion Kauffman Foundation). In the Census survey, 30.9% of employer-firm owners held a bachelor's degree as their highest level of education (U.S. Census Bureau), and for 48.7% the business was their first (U.S. Census Bureau).
"Very important" reason for owning the business (owners of U.S. employer firms, 2023)
Mostly with their own money. In the Census Bureau's Annual Business Survey for 2017, the latest that asked, 64.5% of U.S. employer firms were started or bought with the owners' personal or family savings (U.S. Census Bureau), 16.6% with a bank business loan (U.S. Census Bureau) and 0.5% with venture capital (U.S. Census Bureau); 10.4% said they needed no start-up capital (U.S. Census Bureau). The 2014 Annual Survey of Entrepreneurs gave the same picture: 0.6% of employer-firm founders started with venture capital, against 10.3% who used personal credit cards (Ewing Marion Kauffman Foundation). From that survey Kauffman concluded that at least 83% of new businesses that hire never access a bank loan or venture capital at the start (Ewing Marion Kauffman Foundation).
The Kauffman Firm Survey of about 5,000 startups found 4.4% of startup funding came from venture capitalists and 5.8% from angels (Ewing Marion Kauffman Foundation). Even among 479 fast-growing Inc. 5000 companies, 6.5% had raised venture capital (Ewing Marion Kauffman Foundation).
Young firms and credit
The Federal Reserve Banks' Small Business Credit Survey (fielded September-November 2023, a convenience sample of 10,990 firms) compared startups, firms in business 0-2 years, with older ones. 25% of startup employer firms made a profit in 2022 (Federal Reserve Banks), against 52% of employer firms 3 or more years old (Federal Reserve Banks). 72% of startup employers facing financial challenges dipped into the owners' personal funds (Federal Reserve Banks), against 48% of older employers (Federal Reserve Banks).
Small Business Credit Survey (2023 survey)
Startups (0-2 years)
Older firms (3+ years)
Employer firms that applied for financing in the prior 12 months
In the 2025 survey (fielded September-November 2025, 6,525 employer firms), 48% of employer firms 0-5 years old that applied for a loan, line of credit or cash advance were fully approved (Federal Reserve Banks), against 63% of firms 21 years or older (Federal Reserve Banks). Across all small employer firms, 60% applied for financing (Federal Reserve Banks) and 59% of those with debt had given a personal guarantee (Federal Reserve Banks).
How much money do startups raise, and from whom?
Venture money is huge in total and concentrated in very few companies. Crunchbase, which sells startup data, counted $425 billion of venture and growth funding into more than 24,000 private companies worldwide in 2025 (Crunchbase News), and $679 billion in the first three quarters of 2026 (Crunchbase News). Close to 60% of the 2025 total went to 629 companies raising rounds of $100 million or more (Crunchbase News). The PitchBook-NVCA Venture Monitor found that half of all U.S. venture dollars in 2025 went into 0.05% of deals (PitchBook and NVCA), and that megadeals took 87.5% of U.S. venture dollars in the first half of 2026 (PitchBook and NVCA).
AI takes most of it. AI startups raised 64% of global venture capital in Q3 2026, per Crunchbase (Crunchbase News), and through Q3 2026 U.S. deal value reached $515.8 billion, with AI taking 82.7% (PitchBook and NVCA). Seed money is a small slice: global seed funding was $13 billion in Q3 2026 (Crunchbase News), and North American seed funding was $20.4 billion in 2025, about 9% less than 2024 (Crunchbase News), across a year in which North American deal count fell to just under 10,500 rounds (Crunchbase News). PitchBook and NVCA count more than 60,000 private venture-backed companies in the U.S. (PitchBook and NVCA).
Private offerings and crowdfunding (SEC data)
Young companies dominate private offerings. The SEC's small business advocate reports, in its fiscal 2025 staff report, that 82% of operating companies raising money under Regulation D over the past three years were less than 3 years old (U.S. Securities and Exchange Commission), and that private companies raised 36% ($840 billion) of all capital raised by U.S. companies in the period it covers (U.S. Securities and Exchange Commission).
As of November 2026, Y Combinator invests $500,000 in every company it accepts (Y Combinator) and says that on average 40% of the companies in each batch are just an idea (Y Combinator), while 7% of recent batches had more than $50,000 in monthly revenue when accepted (Y Combinator). a16z speedrun says fewer than 0.4% of applicants are accepted (Andreessen Horowitz) and invests up to $1M in each (Andreessen Horowitz). Techstars says it has accelerated more than 10,900 founders since 2006 (Techstars).
How do pitch videos and pitch decks figure in fundraising?
The best-known founder video is a one-minute one. As of November 2026, Y Combinator's application asks for a founder video that should be 1 minute long and contain nothing but the founders talking (Y Combinator), and says a demo or promotional video belongs in a separate part of the application. Paul Graham's guide, hosted by YC, says YC is statistically much more likely to interview applicants who submit a video, without a figure (Y Combinator). a16z speedrun may ask shortlisted applicants for a short video pitch or a 15-minute interview (Andreessen Horowitz). Government grant pitches can be text only: NSF's SBIR/STTR Project Pitch (as captured in May 2026) is four text sections, with 3500 characters for the technology innovation (U.S. National Science Foundation), 3500 for the technical objectives (U.S. National Science Foundation), 1750 for the market (U.S. National Science Foundation) and 1750 for the company and team (U.S. National Science Foundation), and no video.
The one research study of accelerator pitch videos we found measured how founders deliver, not what they show. Allen Hu and Song Ma analysed 1,139 application pitch videos submitted to top accelerators in 2010-2019 (National Bureau of Economic Research). The median video ran 68 seconds (National Bureau of Economic Research) and the mean 83 seconds (National Bureau of Economic Research). A one-standard-deviation more passionate delivery went with a 35.2% higher probability of being chosen, from a low base (National Bureau of Economic Research). It is a correlation in one sample of applicants.
How long investors look at a pitch deck
Less than three minutes, by DocSend's figures. DocSend, a Dropbox product that sells deck sharing and analytics to founders, measures time spent on decks shared through its own service. Its 2015 study of about 200 decks that raised money found 3 minutes, 44 seconds per successful deck (DocSend), with 40 investor meetings for a typical successful seed round (DocSend). Its later figures are lower: an all-time low of 2 minutes 18 seconds in Q4 2022 (DocSend) and, a year later, an "all-time low" of 2 minutes 24 seconds in Q4 2023 (DocSend). The two releases disagree with each other, which suggests the measure was recalculated, so treat them as approximate. In April 2025 DocSend's research lead wrote that investors spend less than 3 minutes on a deck (DocSend). In 2020, decks that went on to raise money got over four minutes on average and decks that did not got 1 minute 30 seconds (DocSend). Between Q2 2021 and Q2 2023 founders sent 15.9% more decks while investor activity fell 11.5% (DocSend).
Crowdfunding: video is the norm
On Kickstarter, a video is close to standard. In Ethan Mollick's study of 48,526 U.S. Kickstarter projects from 2009 to July 2012, of which 48.1% reached their goal (Journal of Business Venturing), 86% had a pitch video (Journal of Business Venturing), and leaving one out cut the mean project's chance of success by 26%, holding other factors equal (Journal of Business Venturing). Indiegogo reported in April 2012 that 53% of campaigns reaching their goal had a video under three minutes (Indiegogo) and 87% one under five minutes (Indiegogo). Indiegogo's December 2011 figure, that campaigns with a pitch video raised 114% more money (Indiegogo), comes from its own campaigns with no published method.
Kickstarter's own stats page, captured on January 21, 2026, counted 687,200 launched projects (Kickstarter), 292,732 funded (Kickstarter) and $9,419,264,675 pledged (Kickstarter). Its lifetime success rate was 42.74% (Kickstarter), but 24.94% in Technology, where hardware and app startups launch (Kickstarter), against 44.63% in Design (Kickstarter). 9% of finished projects never got a single pledge (Kickstarter), while 79% of projects that passed 20% of their goal went on to be funded (Kickstarter).
Does founder and executive content help sell?
B2B buyers say it does, in surveys run by companies that sell into this market. In Edelman and LinkedIn's 2025 B2B Thought Leadership Impact Report, a survey of 1,934 U.S. business executives on LinkedIn fielded March 17 to April 3, 2025 (Edelman and LinkedIn), 56% of target decision-makers said reviewing an organization's thought leadership is one important way they vet vendors (Edelman and LinkedIn). 95% of "hidden" decision-makers, such as finance, legal and procurement staff who rarely meet sales, said strong thought leadership makes them more receptive to outreach (Edelman and LinkedIn), and 71% of them said they have little or no interaction with sales (Edelman and LinkedIn). 41% of target decision-makers said a C-suite executive had encouraged them to consider a vendor after engaging with its thought leadership (Edelman and LinkedIn), and 53% agreed that high-quality thought leadership makes how well known the vendor is matter much less (Edelman and LinkedIn). On style, 65% of hidden decision-makers preferred a human, less formal tone (Edelman and LinkedIn) and 57% quick takeaways over deep content (Edelman and LinkedIn).
The step from interest to purchase is smaller. In the 2024 edition (3,484 executives in seven countries, fielded late 2023) (Edelman and LinkedIn), 75% said a piece of thought leadership had led them to research a product they were not considering (Edelman and LinkedIn), and 23% of those went on to buy from or work with the organization (Edelman and LinkedIn).
What the public says about leaders who speak for themselves
The 2026 Edelman Trust Barometer surveyed 33,938 people in 28 countries from October 25 to November 16, 2025 (Edelman). 67% said the business leader they would trust speaks directly to the public without anyone editing or rewriting their words (Edelman). But what a CEO says in their own channels ranks below what others say about them:
Most influences opinion of a CEO (Edelman Trust Barometer 2026, 28 countries)
In the same survey, 70% said learning a CEO's personal values matters for trusting them (Edelman), and 63% said the same of obstacles the CEO overcame (Edelman). Consumers want less of it than marketers might expect: in Sprout Social's February 2026 survey of 2,250 social media users in the U.S., UK and Australia (Sprout Social), only 9% wanted content from brand executives (Sprout Social), against 40% for educational posts about products (Sprout Social) and 16% for content from front-line employees (Sprout Social). Sprout sells social media software.
What the platforms report
LinkedIn said creator-led ads on its platform grew 63% year over year in the quarter to September 30, 2026, and paid video 25% (LinkedIn). It has given two different figures for founders: 60% growth in members adding "founder" to their profile (September 2026) (LinkedIn) and the number of founders "up 70%" (October 2026) (LinkedIn), neither with a period or definition. Meta says more than 35 million U.S. businesses use Facebook and Instagram every month (Meta). In YouTube's 2025 U.S. Impact Report by Oxford Economics, 76% of small and medium-sized businesses with a YouTube channel said it helped them reach new customers (YouTube). TikTok's commissioned Oxford Economics study counted 7.5 million U.S. businesses on TikTok in early 2025 (TikTok).
Small businesses, marketing and AI
Finding customers is the problem founders name most. In the Fed's 2025 Small Business Credit Survey, reaching customers and growing sales was the most commonly reported operational challenge, ahead of hiring (Federal Reserve Banks). In the U.S. Chamber of Commerce's 2026 technology survey, conducted by Teneo, marketing and promotions was the top use of technology, at 46% (U.S. Chamber of Commerce), 43% of small businesses used social media at work (U.S. Chamber of Commerce), and 22% used generative AI image or video tools (U.S. Chamber of Commerce). In the Fed survey, 46% of small employer firms used AI (Federal Reserve Banks), and 83% of those used it for writing or marketing (Federal Reserve Banks).
What the popular startup numbers really say
Startup blogs, pitch-deck tools and video vendors repeat a small set of numbers. We traced each one to the organization that produced it.
"90% of startups fail"
No dataset measures this. The startup version goes back at least to Startup Genome's premature-scaling report of August 2011, which says in its introduction that more than 90% of startups fail, with no source (Startup Genome); what that report measured was premature scaling in 3,200 startups. GeekWire repeated it within days as 90 percent (GeekWire), and it now circulates as "nine out of ten startups will fail" (Forbes contributor) and "the 90% of businesses that fail" (Failory), with no population or period. An older form, "90% of businesses fail within 12-18 months", was already credited to the SBA by 2007 (Hacker News, quoting a 2007 post); that year an SBA Office of Advocacy economist told Entrepreneur that no one had found any finding that 9 out of 10 new businesses fail in their first year (Entrepreneur). The even older "four out of five fail in five years" had no traceable source either (The Christian Science Monitor); MIT's David Birch traced it to a Dun & Bradstreet report from the early 1960s saying that four out of five failing businesses were small, which someone flipped (The Christian Science Monitor). Startup Genome's 2019 report adds a variant, that only 1 in 12 entrepreneurs succeed, without defining success (Startup Genome).
The venture figures measure something else. Shikhar Ghosh of Harvard Business School told The Wall Street Journal in 2012 that about three-quarters of U.S. venture-backed firms funded in 2004-2010 do not return investors' capital (The Wall Street Journal); counted as liquidations it was 30% to 40% (The Wall Street Journal), and only if failure means missing projections did it pass 95% (The Wall Street Journal). The NVCA's estimate at the time was 25% to 30% (The Wall Street Journal).
Use instead: BLS survival rates. 77.9% of new establishments survive a year, 51.4% five years and 34.7% ten; closure reaches the high 80s only after three decades (87.4% of the 1994 cohort by 2025).
"Founder-led companies grow 31% faster" or "outperform 3.1 times"
The 3.1 is Bain & Company's, and Bain described it two ways. Its 2016 brief says founder-led S&P 500 companies performed 3.1 times better than other S&P 500 companies from 1990 to 2014 (Bain & Company), 1.8 times excluding tech (Bain & Company); Bain's own HBR piece the same year describes 3.1 times for an index of Fortune 500 companies over 15 years (Bain & Company). The 31% is a different finding: a Purdue study that founder-CEO S&P 500 companies generate 31% more patents (Bain & Company), not more growth. Marketing blogs merge the two into "31% higher growth rates according to Bain" (Averi), which no Bain publication says.
"Investors spend 3 minutes 44 seconds on a pitch deck"
This is DocSend's 2015 study with HBS professor Tom Eisenmann of more than 200 seed and Series A decks that together raised $360 million, reported by TechCrunch on June 8, 2015 as 3 minutes and 44 seconds (TechCrunch). It measured successful decks shared through DocSend, which sells deck analytics to founders; the same study found the rounds took 40 investor meetings (TechCrunch). DocSend's blog was still citing it in 2019 as 3 minutes and 44 seconds (DocSend), and a 2026 page calls it "DocSend's 2026 research" at 3 minutes and 44 seconds (Presentations.AI). Use instead: DocSend's newer averages are lower: 2:51 in Q3 2020 (DocSend), 2 minutes, 44 seconds in Q3 2021 (DocSend), and in 2022 24% less time than in 2021 (TechCrunch), with failing decks dropped after 2 minutes and 13 seconds (TechCrunch). They are not like-for-like with 2015: the old figure covers only successful decks.
"Kickstarter projects with a video raise 115% more"
Kickstarter never said that. Its own blog reported in December 2009 that, of nearly 1,000 finished projects, those with a video succeeded 54% of the time (Kickstarter) against 39% without (Kickstarter), and its Kickstarter School page, captured in April 2012, said 50% (Kickstarter) against 30% (Kickstarter), with no sample or period. Those are success rates, not money raised. The "115% more" line credited to Kickstarter on a video-production company's page (Beverly Boy Productions) matches neither, and seems to borrow Indiegogo's 2011 "114% more" figure for its own campaigns. A "15% versus 37%" version (Stephen Follows) is not in the published Mollick paper. Use instead: Mollick's 2014 study of 48,526 U.S. Kickstarter projects, described above: 86% had a video, and leaving it out cut the chance of success by 26%. That is a correlation; it does not show video causes funding.
"The average successful founder is 45"
The 45 is the study authors' own Harvard Business Review headline (Harvard Business Review), and it is narrower than it reads. Using Census Bureau records, Pierre Azoulay, Benjamin Jones, J. Daniel Kim and Javier Miranda found a mean founding age of 45.0 for the fastest-growing 1 in 1,000 new U.S. firms (American Economic Review: Insights); across all 2.7 million founders of new employer firms in 2007-2014 it was 41.9. The young-founder image comes from curated lists: a mean of 29 for startup award winners and 31 for magazine lists (National Bureau of Economic Research).
We started from the questions people search for about founder and startup 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 263 figures from 44 publishers.
On November 1, 2026, each sourced figure on this page was re-fetched with the fetch cache cleared and matched against the recorded wording: Census Bureau Business Formation Statistics, Business Dynamics Statistics, Nonemployer Statistics and Annual Business Survey (some through the Census data API); BLS Business Employment Dynamics survival tables (through Internet Archive copies, because bls.gov refuses automated requests); the SBA Office of Advocacy; the Federal Reserve Banks' Small Business Credit Survey; Kauffman Foundation reports; SEC staff reports on private offerings and crowdfunding; PitchBook-NVCA, Crunchbase, Kickstarter, Indiegogo, DocSend, Y Combinator, a16z and NSF pages about their own data or programs; the Edelman-LinkedIn and Edelman Trust Barometer reports; LinkedIn, Meta, YouTube and TikTok on their own platforms; journal and NBER papers; and the pages behind each popular claim. The homepage figures are our own reading of the 608 readable homepages of the 619 companies in Y Combinator's Spring, Summer and Fall 2026 batches in a headless browser on November 1, 2026. We left out Gusto's founder survey (no field dates published), Carta and the full PitchBook-NVCA tables (blocked or behind a form), and figures the SEC and SBA repeat from other organizations.
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:
count(youtube_channel_link = yes) / readable = 43 / 608; a youtube.com link whose first path segment is a channel, handle or user (not watch/embed/shorts)
count(founder_team_text or team_link) / readable = 330 / 608; text 'meet the team/founders', 'our team/founders', 'founding team', 'founded by', 'co-founder', or a same-site link to /team, /about, /founders, /company, /our-story
count(yc_page_application_video_public or yc_page_demo_day_video_public) / frame = 15 / 619; the directory's own app_video_public and demo_day_video_public fields
This edition contains no Mark Studios customer data.
Limitations
Nobody counts founder video directly. Government data counts applications, firms and establishments, not what founders publish. Our homepage count covers one accelerator's three most recent batches, which skew toward software and AI, and only the homepage.
Survival is counted by location, not company. BLS survival rates follow establishments, which include new branches of existing firms; Kauffman and Census firm-level figures are close but not identical.
Several sources sell to founders. DocSend, Crunchbase, PitchBook, Sprout Social, Wyzowl and LinkedIn sell to startups or marketers, and Edelman sells communications. DocSend measures only decks shared through its own service.
Convenience samples. The Small Business Credit Survey is not a random sample, and the Edelman-LinkedIn respondents were recruited on LinkedIn.
Some figures are old. The latest Census data on start-up capital is for 2017, and the best-known crowdfunding video studies date from 2009-2014.
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/founder-video-statistics#DRV-006. To cite the page as a whole:
Mark Studios. "Founder and Startup Video Statistics." Last verified Nov 1, 2026. https://www.markstudios.com/research/founder-video-statistics
APA
Mark Studios. (2026, November 1). Founder and startup video statistics. https://www.markstudios.com/research/founder-video-statistics
MLA
“Founder and Startup Video Statistics.” Mark Studios, 1 Nov. 2026, www.markstudios.com/research/founder-video-statistics.
Chicago
Mark Studios. “Founder and Startup Video Statistics.” Last modified November 1, 2026. https://www.markstudios.com/research/founder-video-statistics.
BibTeX
@misc{markstudios_founder_video_statistics_2026,
author = {{Mark Studios}},
title = {{Founder and Startup Video Statistics}},
year = {2026},
month = nov,
url = {https://www.markstudios.com/research/founder-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.
PitchBook and National Venture Capital Association (PitchBook-NVCA Venture Monitor), Q4 2025 PitchBook-NVCA Venture Monitor (2026-01). Accessed 2026-11-01.
PitchBook and National Venture Capital Association (PitchBook-NVCA Venture Monitor), Q2 2026 PitchBook-NVCA Venture Monitor (2026-07). Accessed 2026-11-01.
PitchBook and National Venture Capital Association (PitchBook-NVCA Venture Monitor), Q3 2026 PitchBook-NVCA Venture Monitor (2026-10). Accessed 2026-11-01.
PitchBook and National Venture Capital Association (PitchBook-NVCA Venture Monitor), Q3 2025 PitchBook-NVCA Venture Monitor (2025-10). Accessed 2026-11-01.
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 263 figures from 44 publishers, each checked against its original source.
Frequently asked questions
How many startups put video on their homepage?
In our November 1, 2026 count of the 608 readable homepages of Y Combinator's Spring, Summer and Fall 2026 companies, 39.8% had a video player and 21.5% had a click-to-play video. The rest of the video was muted background motion.
How long should a startup pitch video be?
Y Combinator asks for a founder video of 1 minute with only the founders talking. In a study of 1,139 accelerator application videos from 2010-2019, the median ran 68 seconds and the mean 83 seconds.
Do 90% of startups fail?
No dataset shows that. BLS finds 77.9% of new U.S. establishments survive a year, 51.4% five years and 34.7% ten. The 90% appears without a source in Startup Genome's 2011 report.
How many new businesses are started each year in the U.S.?
About half a million business applications a month: 534,947 in September 2026. Fewer become employers: Census counted 508,777 startup employer firms in 2023.
How many startups get venture capital?
Very few. In Census's 2017 survey, 0.5% of U.S. employer firms were started or bought with venture capital, and 64.5% with the owners' savings.
How long do investors spend on a pitch deck?
Under three minutes, by DocSend's data: 2 minutes 24 seconds in Q4 2023. The popular 3 minutes 44 seconds is DocSend's 2015 study of about 200 successful decks.
Do Kickstarter projects with video do better?
In Mollick's study of 48,526 U.S. Kickstarter projects, 86% had a video, and leaving it out cut the chance of success by 26%. Kickstarter's own 2009 figures were 54% success with video and 39% without.
