A few weeks ago, I needed one simple statistic for a thought leadership piece: what percentage of customers believe it takes too long to reach a live agent? Nothing niche, nothing exotic. You’d think the answer would be easy to find.
Instead, I fell into the familiar rabbit hole, page after page of recycled numbers. Half the search results confidently stated the answer was “75%,” yet not a single one linked to a primary source. One blog traced back to a PDF called “50 customer service facts you won’t believe” with no citations at all. Others simply repeated the same percentage as if it were a universal truth, complete with neatly designed infographics that gave the illusion of credibility.
It’s becoming the norm. And with AI accelerating content production, it’s only getting worse.
This isn’t a rant about AI. It’s a rant about behaviour. Because the real problem is much older: writers and marketers are publishing faster than they’re thinking.
Let’s talk honestly about ghost stats, why they spread, and what we’re risking every time we repeat them.
The problem with “ghost stats”
AI is not the villain here. Laziness is. The temptation to grab the first stat that fits the narrative, even if you can’t trace it, is now baked into online publishing. Content velocity has replaced content integrity.
The pattern usually looks like this:
- Someone publishes a stat with no source.
- Another writer pulls it into their article.
- It gets scraped, summarised, and turned into a listicle.
- AI models train on those listicles.
- The stat appears in even more blogs, newsletters, and whitepapers.
- Eventually, it becomes “common knowledge.”
This is how you end up with numbers like “90% of people only read headlines” and “humans have an 8-second attention span”, both of which are fabricated but quoted thousands of times.
Ghost stats are convenient. They’re quick. They give your writing a veneer of authority. But they also rot the foundation of your argument. And in industries like CX, where decisions are tied to budgets, team structures, and customer experience strategy, this isn’t harmless.
CX leaders, marketers, and executives are using these numbers to prioritise projects, allocate spend, and evaluate platforms. If the data is wrong, the strategy will be wrong too.
Why ghost stats spread so easily
There are three big reasons this problem has exploded.
1. SEO culture rewards speed over accuracy
Google rewards freshness, volume, and internal consistency. It does not check whether the stats in your content exist. As long as enough sites repeat the same number, the algorithm interprets it as relevance, not truth.
Writers respond to the incentives they’re given. If your job is to publish four blogs a week, verifying a stat becomes a luxury.
2. Most people don’t know how to trace a stat
If you’ve never worked with research data, it’s not always obvious how to evaluate a source. A journey that should start with “What’s the original dataset?” instead stops at “This looks like a big brand; it must be fine.”
And so numbers drift away from their source until they lose all meaning.
3. AI summarises the internet — including its mistakes
Large language models don’t know which sources are reliable. They know which patterns are most common. If a ghost stat appears on enough pages, the model will confidently surface it.
AI doesn’t create fake statistics out of thin air. It just repeats what humans have already repeated too many times.
EEAT: Google’s polite way of telling you to grow up
Google’s EEAT framework, Experience, Expertise, Authoritativeness, and Trustworthiness, exists partly to counter this behaviour.
- Experience: Has the author actually worked with the topic?
A person with first-hand CX experience won’t blindly copy a stat that contradicts what they see on the ground. - Expertise: Is the writer qualified to interpret the data?
Someone who understands surveys and methodology won’t treat a 50-person poll as universal truth. - Authoritativeness: Does the source have a track record of credible research?
A stat from Zendesk, NICE, Salesforce, Ipsos or McKinsey is not the same as a stat from “bestcustomerservicefacts.org”. - Trustworthiness: Can readers verify the claims?
If a number appears with no link, no date, no study, and no sample size, it’s not trustworthy.
Ghost stats fail every principle. They fail Experience because no practitioner believes them. They fail Expertise because they’re used without context. They fail Authoritativeness because they come from nowhere. And they fail Trustworthiness because no reader can check them.
If you drop a stat into your content without a source, you fail EEAT immediately.
The legal bit (which most writers forget)
No, there is no global law forcing you to link to every statistic. But that doesn’t mean you’re off the hook.
UK — ASA rules
Marketers must not mislead by using claims they cannot substantiate. The ASA has upheld cases where brands used percentages without being able to produce the underlying evidence. Even B2B brands aren’t exempt.
US — FTC rules
Advertising claims must be “truthful, not misleading, and backed by evidence.” If you quote a number to influence a purchase decision, you need to be able to prove it.
EU — Unfair Commercial Practices Directive
Any communication that misleads consumers, including unsubstantiated data, can be challenged.
You won’t get fined for forgetting to hyperlink a stat. But if your organisation is using ghost stats in marketing, sales, or thought leadership that influences buying decisions, you’re opening the door to real risk.
The irony is that regulators don’t even need to intervene. Your audience will. If a CX leader reads your content and the numbers look suspicious, they won’t bother reporting you. They’ll simply stop trusting you.
And trust, once gone, is almost impossible to rebuild.
But let’s talk about the deeper issue: the credibility premium
People assume the cost of ghost stats is small: maybe a misinformed blog, maybe a slightly off webinar talking point. But the real cost is far higher.
Content is a trust exercise. Every fact, every example, every stat is a signal of your seriousness. When the foundation feels shaky, the entire argument collapses.
A few things that happen when you quote unverified numbers:
- Decision-makers assume your research is shallow
If the stats are sloppy, the thinking might be too. - Your brand loses authority
Your competitors who cite credible studies look instantly more trustworthy. - Journalists won’t quote you
Real media outlets won’t touch an article with unverifiable numbers. - Thought leadership becomes thought recycling
You can’t lead if you’re repeating fiction. - Your team copies your bad habits
If leaders include ghost stats, junior writers feel permission to do the same.
It’s a cultural issue, and culture spreads faster than facts.
How to raise the bar for content (and avoid ghost stats entirely)
Here’s what credible, trustworthy writing looks like today.
1. Trace every stat to the primary source
If you can’t find the original PDF, study, or dataset, don’t use the number. Ever.
A proper stat should tell you:
- Who was surveyed
- When the study was conducted
- How many respondents there were
- What market or region they came from
- What question they were asked
Without those details, the percentage is meaningless.
2. Don’t trust listicles, quote roundups, or “research summaries”
If a blog uses phrases like “According to research…” without naming the research, assume the stat is fake.
3. Context is just as important as the number
A study of 2000 UK consumers from 2018 is not the same as a study of 5,000 global decision-makers from 2024. But online, these get mixed together until they sound interchangeable.
Give readers context so they can judge relevance for themselves.
4. Use stats as evidence, not decoration
If your argument doesn’t work without the stat, your argument isn’t strong enough.
5. If the real data doesn’t exist, say so
Honesty builds more trust than bravado. A line like:
“Despite the number widely quoted online, no reputable study has actually measured this. Most available figures trace back to unsourced blogs.”
…is far more credible than a fake percentage.
6. Develop your own stats
This is an opportunity most brands waste.
Run your own survey. Interview your own customers. Analyse your own platform data.
Nothing builds authority faster than original insight.
7. Build an internal “approved stats library”
If multiple writers contribute to your output, give them a vetted list of credible numbers from real sources and real dates. It stops ghost stats at the source.
AI didn’t create this problem — it exposed it
AI forces a race to the top or a race to the bottom. You either become more credible, more transparent, and more thoughtful, or you become faster at repeating misinformation.
The real competitive advantage now is rigour.
The brands that check their sources, cite their facts, and refuse to use ghost stats will stand out instantly.
AI is a brilliant tool. It can summarise, transform, and speed up the grunt work. But it can’t replace the one thing writers can’t outsource: judgement.
Because when everyone else is repeating stats that don’t exist, the people who take five minutes to verify a number will look like leaders.
Trust is the currency of content.
And nothing burns trust faster than quoting numbers you can’t prove.
