Five Marketing Statistics You Have Quoted. Four Do Not Hold Up

Why this is worth twenty minutes of your time

You have quoted at least one of these. In a pitch, in a board deck, in a proposal, probably last month.

Here is the risk. Every one of these numbers is one search away from being checked, and the person most likely to check is the sceptical finance director you most need to convince. Getting caught repeating a dead statistic does not just cost you that number. It puts a question mark over everything else you said.

So we traced five of the most repeated figures in marketing back to where they came from. Four do not survive. One does, and it is instructive that it does.

1. "Humans now have an eight second attention span, shorter than a goldfish"

Verdict: dead. Do not use it.

This is the most repeated statistic in the entire marketing industry and it is not real.

It is usually credited to a 2015 Microsoft study. That study exists, and it never states an eight second figure, and it does not mention goldfish. The number traces instead to a website called Statistic Brain, which attributed it to sources that could not be produced on request. A BBC journalist chased it and could not substantiate it. Attention researchers say there is no such measure, because attention span is task dependent rather than a single number you can put on a species.

There is also no evidence for the goldfish half. Fish sustain attention for far longer than nine seconds.

The Well by Northwell and Temple Law's evidence review both walk through the collapse in detail. LinkedIn's own marketing blog has published a piece calling it out, which is unusual, because the myth is commercially convenient for anyone selling short form video.

What is true instead: people abandon things that are not worth their time, fast, and they will watch a 45 minute documentary that earns it. That is a quality argument, not a biology one.

2. "It costs five times more to acquire a customer than to retain one"

Verdict: never validated. Use with a caveat or not at all.

The trail runs back to work by the Technical Assistance Research Project in Washington in the late 1980s, and to a Harvard Business Review article that restated it as a calculation built on assumptions.

That is the problem. It was a reasonable thought experiment, not a measured result, and it has been quoted as though it were the latter for over thirty years. Ipsos published a paper arguing the assumptions underneath it do not hold: that retained customers reliably spend more over time, pay full margin, and cost less to serve.

You will also see the same claim quoted as five times, seven times, and twenty five times, by people citing each other. When a number has a range of 5x to 25x depending on who is repeating it, it is not a measurement.

What is true instead: retention is usually cheaper than acquisition in your business. Work out your own figure from your own numbers. It takes an afternoon and it is defensible, which is the point of retention versus acquisition.

3. "It takes seven touches before someone buys"

Verdict: folklore. Charming, and not research.

The Rule of Seven is generally traced to the film industry in the 1930s, when studio marketers reportedly concluded that a cinemagoer needed to see a message about seven times before acting. There is no peer reviewed basis for seven as a universal number, and no study establishes it across categories, price points, or channels.

Real research on advertising frequency does exist. It says something more careful: that effectiveness changes with repetition, that the curve differs by category and by how familiar the brand already is, and that the optimum is not a fixed integer. None of that supports a magic seven.

What is true instead: most buyers are not in market when you reach them, so repeated exposure over time is genuinely how you get remembered. That is the 95:5 rule, and it is better evidenced than the seven ever was.

4. "Content marketing generates three times as many leads as outbound and costs 62% less"

Verdict: a vendor benchmark from over a decade ago, repeated as if it were current science.

This one is quoted more than any other statistic in content marketing. It comes from a Demand Metric infographic. Demand Metric is a research and advisory firm, so this is not fabricated, but there are three problems with how it gets used.

It is old. It is a vendor benchmark rather than independent peer reviewed research. And the methodology behind it is not something most people quoting it have ever seen, because almost everybody is citing a blog post that cited another blog post.

Notice also who benefits. A statistic that says content marketing is three times better and 62% cheaper is repeated most enthusiastically by people selling content marketing. We sell content marketing. That is exactly why we are not going to lean on it.

What is true instead: content compounds and paid media stops the day you stop paying. You can demonstrate that from your own analytics without borrowing anyone's infographic, as in does blogging still work.

5. "People form an opinion of your website in about 50 milliseconds"

Verdict: this one survives.

Here is the contrast, and the reason this article is not just contrarianism.

Gitte Lindgaard and colleagues published "Attention web designers: You have 50 milliseconds to make a good first impression!" in the peer reviewed journal Behaviour and Information Technology in 2006. Participants were shown web pages for extremely short exposures and their visual appeal judgements correlated strongly with judgements made with far more time.

Why does this one hold up when the others do not? Because there is a named author, a named journal, a described method, and a result other researchers have been able to engage with. That is the whole difference.

What it actually means: first impressions of visual appeal form nearly instantly. It does not mean somebody decides whether to buy from you in 50 milliseconds, and it is often stretched into that. Even a good statistic gets abused.

How to check any statistic in 90 seconds

This is the part worth keeping. Four questions, in order.

1. Who is the named source? Not the blog you read it on. Follow the link. Then follow that link. If the chain ends at another blog post, stop using the number.
2. Is there a method? A sample size, a date, a description of what was measured. No method, no statistic.
3. How old is it? Anything about digital behaviour from before 2020 describes a different internet.
4. Who benefits from it being true? A number that flatters the product of whoever is quoting it deserves a harder look. That includes ours.

If a claim fails any of those, you have two honest options. Say it with the caveat attached, or find your own number. Your own numbers are usually more persuasive anyway, because they are about the reader's situation rather than an industry average.

Why we bothered

Every figure on this site links to a source you can click, and where sources disagree we give the range and say they disagree. That is slower to write and it is the whole point, for the reason set out in say the number or say nothing.

If your marketing rests on numbers you have never traced, that is not a credibility problem yet. It becomes one the first time a buyer checks.

Frequently asked

Is it wrong to ever quote an industry statistic? No. Quote it, name the source, and note when it is old or contested. The caveat costs you a clause and buys you the room.

What if my own data is too small to be meaningful? Say so. "Across our last forty projects" is a real, checkable, honest claim, and it beats a borrowed average from an unknown sample.

Are there other zombie stats? Plenty. The 4% who complain, the seconds of page load that cost a fixed percentage of conversions, the share of communication that is body language. Apply the four questions.

Can Fixora audit the claims on our site? Yes. Every statistic traced, sourced, caveated or removed, and replaced with numbers from your own data where we can. Send us the domain and get the findings within 48 hours.