Report · Updated
The Most Misquoted Marketing Statistics
We traced 264 of the statistics marketers repeat most back to the documents they came from. 127 of 264 (48%) don’t hold up as usually quoted. Below: the most famous offenders, what people say, what the source actually says, and a line you can use instead.
- Verified 137
- Outdated 55
- Misquoted 39
- Unsourced 33
How the headline number is defined
48.1% = 127 ÷ 264. The denominator is every stat check in our registry: a widely repeated marketing statistic we traced to its origin (264 as of 2026-10-17). The numerator is the stat checks whose verdict is misquoted (39), outdated (55) or unsourced (33), i.e. every verdict other than verified (137). The 216 reference facts we took directly from primary sources (Pew, Census, IAB/PwC, FTC…) are excluded, because we didn’t pick them as claims to test.
Stricter cut: 72 of 264 (27%) are misquoted or have no primary source at all. Not a random sample: we chose stats that are quoted a lot, some because they looked suspicious, so read this as “the famous ones”, not “all marketing statistics”.
How stats go wrong
The most common kinds of drift
Among the 127 stats that don’t hold up, how many show each problem (one stat can show several). Counted from the drift notes on each claim page.
The list
25 famous stats that don’t say what the slide says
Each entry links the full trail and the primary source we opened. Verdicts describe how a number is sourced and quoted, not whether the idea behind it is true.
-
1MisquotedConfidence: high
Influencer marketing returns $5.78 for every $1 spent
What people say
"Influencer marketing ROI is $5.78 per $1" (as revenue or ROI)
What the source actually says
Average earned media value per $1 spent has increased to $5.78
What changed: Metric: earned media value (cost of equivalent ad exposure) quoted as revenue/ROI
Say this instead: Respondents to Influencer Marketing Hub's 2020 survey reported an average $5.78 in earned media value — not sales — per $1 spent in 2019.
Full source trail Primary source: Influencer Marketing Hub — Benchmark Report 2020
-
2MisquotedConfidence: high
92% of consumers trust influencers, UGC or word of mouth more than ads
What people say
"92% of consumers trust influencers (or UGC) over ads" — Nielsen
What the source actually says
“recommendations from people I know” scored highest on trust, with 92 percent of consumers trusting this source completely or somewhat.
What changed: Subject: 'people I know' (friends and family) rewritten as influencers or UGC
Say this instead: In Nielsen's 2012 global online survey, 92% said they trust recommendations from people they know completely or somewhat; ads on social networks were trusted by 36%.
Full source trail Primary source: Nielsen: Trust in Advertising, Paid, Owned and Earned (2012)
-
3UnsourcedConfidence: high
Human attention span is 8 seconds, shorter than a goldfish's 9
What people say
"Microsoft study: attention spans shorter than a goldfish"
What the source actually says
the figure that everyone picked up on - about our shrinking attention spans - did not actually come from Microsoft's research.
What changed: Attribution: Microsoft → Statistic Brain → nothing
Say this instead: There is no measured 'average attention span'; attention depends on the task and the content (BBC More or Less, 2017).
Full source trail Best source we found (no primary exists): BBC — Busting the attention span myth (2017)
-
4UnsourcedConfidence: high
The average person sees 4,000–10,000 ads a day
What people say
"We see up to 10,000 ads a day"
What the source actually says
"Our intention was not to endorse any particular number," says Smith ... These numbers are unsubstantiated, and the citations never get more specific than 'experts agree'
What changed: Number: no study ever measured thousands
Say this instead: Estimates of daily ad exposure vary widely; counted studies found roughly 76–362 exposures a day, far fewer of them noticed.
Full source trail Best source we found (no primary exists): The Drum — How many ads do we really see in a day? (2023)
-
5OutdatedConfidence: high
Email marketing returns $36 for every $1 spent
What people say
"Email ROI is 36:1" (as measured and current)
What the source actually says
These insights are derived from Litmus’ 2020 State of Email Survey of more than 2,000 marketing professionals worldwide.
What changed: Method: self-reported survey average, not measured ROI
Say this instead: Marketers surveyed by Litmus in 2020 self-reported an average email ROI of about 36:1; in 2025, 35% of marketing leaders reported $10–36 per $1.
Full source trail Primary source: Litmus — The ROI of Email Marketing
-
6OutdatedConfidence: high
Text messages have a 98% open rate
What people say
"SMS has a 98% open rate (vs 20% for email)"
What the source actually says
neither datapoints represent an accurate reflection of today's world of mobile messaging.
What changed: Metric: a 'read' estimate quoted as a measured open rate
Say this instead: Mobilesquared's 2010 estimate of 98% SMS read rates is withdrawn by its author; its 2022 data suggests about 55% are read.
Full source trail Primary source: Mobilesquared: SMS engagement rates dropping (2022)
-
7MisquotedConfidence: high
A 5% increase in customer retention increases profits by 25–95%
What people say
"Bain: +5% retention = +25–95% profit" (any business)
What the source actually says
By retaining a mere 5% more customers, e-companies can boost profits by 25–95%
What changed: Population: e-companies in 2000 → every business today
Say this instead: Bain's Reichheld estimated in 2000 that retaining 5% more customers could lift e-commerce profits by 25–95%; his 1990 HBR cases (bank, insurance, auto service) showed 25–85% gains in customer value. The effect varies by industry, so model it with your own retention and margin data.
Full source trail Primary source: Bain & Company — E-loyalty: Your secret weapon on the web (2000)
-
8UnsourcedConfidence: high
Acquiring a new customer costs 5–25x more than retaining one
What people say
"HBR: acquisition costs 5–25x more than retention"
What the source actually says
Depending on which study you believe, and what industry you’re in, acquiring a new customer is anywhere from five to 25 times more expensive
What changed: Attribution: an HBR blog line quoted as an HBR study
Say this instead: Acquisition is often more expensive than retention, but the ratio depends on the industry; measure your own CAC against retention cost.
Full source trail Primary source: HBR — The Value of Keeping the Right Customers (2014)
-
9UnsourcedConfidence: medium
Prospects need 7 touches before they buy (the Rule of 7)
What people say
"1930s Hollywood studios found 7 exposures are needed"
What the source actually says
We searched for the Hollywood 'seven posters' research and couldn't find a single source that names the studio, the year or the study.
What changed: Lant's rule was about prospects noticing a message within 18 months, not about buying
Say this instead: The 'Rule of 7' is a rule of thumb; advertising research on effective frequency has proposed 2–3 exposures, and results vary by brand and channel.
Full source trail Best source we found (no primary exists): SalesBlog — Myth: the Rule of 7 (2026)
-
10UnsourcedConfidence: high
Viewers retain 95% of a message in video vs 10% in text
What people say
"Studies show viewers retain 95% of a video message"
What the source actually says
It was directional. It was not a peer-reviewed neuroscience study.
What changed: Precision: a directional pattern became a 'scientific' fact
Say this instead: Don't cite 95%/10%. If needed, say video can improve recall for complex topics, and cite peer-reviewed work on dual coding.
Full source trail Primary source: Insivia — origin explanation
-
11UnsourcedConfidence: high
The brain processes visuals 60,000 times faster than text
What people say
"3M research: visuals are processed 60,000x faster than text"
What the source actually says
the 2001 3M document you get is not a research summary, not product of their R&D, but a sales brochure
What changed: Attribution: a projector-maker's brochure quoted as 3M research
Say this instead: Don't cite it. Images can be understood quickly, but no study supports '60,000 times'.
Full source trail Best source we found (no primary exists): CogDogBlog: 60,000 times, truthiness achieved (2014)
-
12MisquotedConfidence: high
Only 7% of communication is words (7-38-55 rule)
What people say
"93% of communication is nonverbal"
What the source actually says
was actually concerned with feelings and attitudes
What changed: Scope: findings about judging feelings from inconsistent cues generalized to all communication
Say this instead: Mehrabian's 1960s experiments concern how people judge feelings when words and tone conflict; he says the ratios don't apply to communication in general.
Full source trail Best source we found (no primary exists): Max Atkinson: Mehrabian's moans about the myth (2009, quoting Mehrabian)
-
13OutdatedConfidence: high
B2B buyers are 57% through the purchase decision before talking to sales
What people say
"Buyers are 57% (or 70%) done before contacting a vendor" (as current)
What the source actually says
the average customer had completed 57% of the purchase decision-making process prior to engaging a supplier sales rep directly.
What changed: Date: 14 years old, quoted as current
Say this instead: CEB research from 2012 found B2B buyers were on average 57% through the purchase process before engaging a sales rep.
Full source trail Primary source: CEB/Google: The Digital Evolution in B2B Marketing (2012, PDF)
-
14UnsourcedConfidence: medium
Color increases brand recognition by 80%
What people say
"University of Loyola study: color increases brand recognition by up to 80%"
What the source actually says
Google couldn't find any paper by dr. Hoadley that had that quote in it.
What changed: Attribution: a printer-maker's sales leaflet quoted as a university study
Say this instead: Don't cite a number. If you need evidence on color, cite a named peer-reviewed study and what it measured.
Full source trail Best source we found (no primary exists): Insights4Print: the real contents of the 'Loyola study' (2019)
-
15UnsourcedConfidence: medium
Every 100ms of latency cost Amazon 1% of sales
What people say
"Amazon found every 100ms of latency costs 1% in sales"
What the source actually says
In A/B tests, we tried delaying the page in increments of 100 milliseconds and found that even very small delays would result in substantial and costly drops in revenue.
What changed: Precision: 'substantial drops' became an exact 1% per 100ms
Say this instead: An early-2000s internal Amazon A/B test reportedly found small page delays reduced revenue; no data was published.
Full source trail Best source we found (no primary exists): Niels-Ole Kühl: Amazon 100ms = 1% revenue? (2018 trace)
-
16OutdatedConfidence: high
88% of consumers trust online reviews as much as personal recommendations
What people say
"88% of consumers trust online reviews as much as personal recommendations" (as current)
What the source actually says
88% of consumers say they trust online reviews as much as personal recommendations (vs. 79% in 2013)
What changed: Date: 12 years old; BrightLocal's later editions report much lower figures (49% in 2026, with changed wording)
Say this instead: In BrightLocal's 2014 panel survey, 88% said they trust online reviews as much as personal recommendations; the same company's recent editions report around half.
Full source trail Primary source: BrightLocal: Local Consumer Review Survey 2014
-
17OutdatedConfidence: high
53% of mobile users abandon a site that takes over 3 seconds to load
What people say
"53% of mobile visitors leave if a page takes longer than 3 seconds" (as current)
What the source actually says
53% of visits are likely to be abandoned if pages take longer than 3 seconds to load
What changed: Wording: 'likely to be abandoned' (modeled) quoted as measured
Say this instead: Google's 2016 analysis of ~3,700 mobile sites estimated 53% of visits are likely abandoned when pages take over 3 seconds.
Full source trail Primary source: Google: The need for mobile speed (2016)
-
18MisquotedConfidence: high
80% of sales require 5 follow-ups
What people say
"80% of sales require 5 follow-up calls" (National Sales Executive Association)
What the source actually says
80% of sales are made on the fifth to twelfth contact
What changed: Metric: 'fifth to twelfth contact' became 'five follow-ups'
Say this instead: A 1942 survey of fewer than 40 sales executives reported most sales closing after several contacts; it is not modern evidence.
Full source trail Primary source: SMEI: Sales Statistics (archive research)
-
19UnsourcedConfidence: medium
93% of online experiences begin with a search engine
What people say
"93% of online experiences begin with a search engine"
What the source actually says
A study by Forrester from 2006 found that 93% of all internet traffic comes from a search engine.
What changed: Meaning shifts between 'experiences begin with' and 'traffic comes from'
Say this instead: Don't cite it; use current referral-traffic data with a named source.
Full source trail Best source we found (no primary exists): Daniel Bianchini: 10 stats to justify SEO (example of the attribution)
-
20UnsourcedConfidence: medium
Content marketing costs 62% less than traditional marketing and generates 3x as many leads
What people say
"Demand Metric study"
What the source actually says
This infographic provides statistics and best practices for content marketing.
What changed: Type: an infographic quoted as a study
Say this instead: Don't cite it. Use a current survey with a stated method (e.g. CMI/MarketingProfs annual B2B research) and name its sample.
Full source trail Primary source: Demand Metric — Infographic: Content Marketing
-
21MisquotedConfidence: high
Businesses make $2 (or $8) for every $1 spent on Google Ads
What people say
"Google Ads returns $2 / $8 per $1" (as measured ROI)
What the source actually says
we assume that for every $1 businesses spend on Google Ads, they generate an average of $2 in profit.
What changed: Type: a seller's modelled assumption quoted as advertiser-measured ROI
Say this instead: Google's economic-impact model assumes $2 in profit per $1 of Google Ads, and $8 when organic search clicks are included.
Full source trail Primary source: Google Economic Impact — Methodology
-
22OutdatedConfidence: medium
Companies that blog get 55% more website visitors
What people say
"Companies that blog get 55% more visitors, 97% more links, 434% more indexed pages"
What the source actually says
that's the average increase in visitors to small business websites as a result of blogging, according to the study.
What changed: Causality: a correlation among one vendor's customers quoted as an effect
Say this instead: A 2009 HubSpot comparison of its small-business customers found those with blogs averaged 55% more visitors.
Full source trail Best source we found (no primary exists): Search Engine Watch: HubSpot blogging study (2009)
-
23OutdatedConfidence: high
Inbound leads cost 61% less than outbound leads
What people say
"Inbound leads cost 61% less than outbound" (as current)
What the source actually says
the 2009 survey showed that inbound marketing dominated organizations experienced a 61% lower cost per lead
What changed: Date: 14+ years old
Say this instead: HubSpot's 2009-2012 surveys of marketers reported roughly 60% lower cost per lead for inbound-dominated companies.
Full source trail Primary source: HubSpot: Inbound leads cost 62% less (2011 report post)
-
24MisquotedConfidence: high
61% of consumers trust influencer recommendations vs 38% for brands
What people say
"61% trust influencers, only 38% trust brands"
What the source actually says
61% are likely to trust recommendations from a friend, family member or influencer on social platforms.
What changed: Subject: influencers were bundled with friends and family in one answer
Say this instead: A May 2020 Matter Communications survey of 1,000 US adults found 61% likely to trust recommendations from friends, family or influencers on social, versus 38% from brands.
Full source trail Primary source: Matter Communications — press release, May 26 2020
-
25UnsourcedConfidence: high
People remember 10% of what they read, 20% of what they hear…
What people say
"Dale's Cone: we remember 10% of what we read, 90% of what we do"
What the source actually says
D. G. Treichler didn't cite any research
What changed: Attribution: numbers from a trade-magazine article credited to Dale or to 'research'
Say this instead: Don't cite these percentages; Dale's cone has no numbers and no research supports them.
Full source trail Best source we found (no primary exists): Work-Learning Research: People remember 10%, 20%... Oh really? (2006)
Browse all 480 entries in the registry · Download all 127 as CSV
Is one of these in your deck?
Check a stat free in seconds, or get the whole asset traced in 48 hours.
The free checker matches your text against all 480 entries. Stat Check traces every number in a deck, post or page, up to 15 stats, with a safe rewrite for each.
Methodology
- We normalize each statistic as it is usually quoted: number, unit, population, date and attribution.
- We follow citations hop by hop to the organization that collected the data, and open the primary document.
- We compare the quote with the source on number, subject, metric, population and date, and note every mismatch (“drift”).
- We assign a verdict (verified, misquoted, outdated, unsourced) and a confidence level, and write a safe rephrasing.
Research is AI-assisted; every verdict rests on the documents listed on its claim page. Full detail: methodology.
About these numbers
- All figures on this page are computed from the live registry at build time (2026-10-17). When a verdict changes, this page changes with it.
- The list above is an editorial pick of the most famous stats that don’t hold up; its order is editorial, not a measured citation count.
- Drift groups map our drift labels as follows: Too old or superseded: Date, Age, Superseded, Tense; Credited to the wrong source: Attribution, Source, Origin; Source type or method misrepresented: Type, Method, Sample, Measurement; Scope stretched beyond the data: Scope, Population, Generalization, Context; Metric or meaning changed: Metric, Definition, Subject, Concept, Meaning, Wording, Comparison, Direction, Currency, Detail; Number or precision changed: Precision, Number; Correlation sold as cause: Causality, Cause.
- Data license: CC BY 4.0. Quote it with a link to this page. Found an error? Tell us.