The Story a Stranger Can Retell

There is a number every storytelling talk opens with. A fact wrapped in a story, it says, is twenty-two times more memorable than the fact alone. It is quoted in agency decks, training courses and keynote slides, usually credited to a Stanford professor. I went looking for the study. There isn't one.

The claim that travelled because nobody checked it

Start with the number, because dismantling it is not a digression. It is the whole subject arriving early.

The twenty-two times figure is attributed, almost universally, to Jennifer Aaker at Stanford's Graduate School of Business. It does not appear in her published research. Trace it back through the citations and each one points at another article that points at another article, and the chain terminates in a talk in which the figure is stated as fact with no derivation and no study behind it. Nobody fabricated it in bad faith. It simply got repeated by people who assumed the person before them had checked.

I am opening with this for a reason beyond housekeeping. This chapter is about what survives being retold by people who did not originate it. The twenty-two times figure is the purest available demonstration — a claim so transmissible that it crossed an entire industry without its evidence, because it was short, causal, surprising and flattering to the person repeating it. It is a case study in the mechanism, wearing the costume of a finding about the mechanism.

Which leaves an obvious question. If the famous number is empty, what is the real one?

What a narrative actually moves, and by how much

There is good evidence. It is just much smaller, and much duller, than the version that travels.

The best available summary is a meta-analysis by Braddock and Dillard, published in Communication Monographs in 2016, pooling experiments that tested whether exposure to a narrative shifted what people believed, felt, intended or did. It found consistent, positive effects across all four. It also found them small.

Read the bottom row carefully, because it is the one that will be quoted back at you. Behaviours — people actually doing something differently — shows the largest coefficient on the chart and rests on five studies and 978 people. The three rows above it pool between 5,000 and 7,400. That is an order of magnitude difference in how much weight the number can carry, and it sits underneath the claim marketers most want to make.

I want to be precise about what this does not mean. A small average effect is not a useless effect — at scale, and repeated, three per cent is a great deal of money, and it is roughly the size of effect that most of marketing runs on. What it means is that anyone promising you a transformation from a better story is selling something the evidence does not contain.

Why it works when it works

The mechanism that has held up best is not about chemistry. It is about attention, and specifically about what your attention is not doing while it is inside a story.

Green and Brock called it transportation: the state of being absorbed into a narrative such that the immediate world recedes. Their argument, published in the Journal of Personality and Social Psychology in 2000, is that a transported reader generates fewer counter-arguments. Not because their critical faculties have been disabled, but because those faculties are occupied following the story instead of interrogating the claim.

Notice what that mechanism implies, because it is uncomfortable and it belongs in a study that keeps saying it would rather be trusted than admired. The thing that makes narrative persuasive is that it lowers the reader's guard. That is a capability, and like every capability in this study it is available to anyone, for any purpose. A story is exactly as good at transmitting something false as something true. Nothing in the mechanism checks.

Which is why this chapter sits where it does — after a position you can defend and a memory map you can verify, not before them. Transmissibility applied to an unexamined claim is not strategy. It is amplification.

The brain science you should stop quoting

Somewhere in the last fifteen years, storytelling acquired a neuroscience. It is worth going through it carefully, because a good part of it did not survive the decade and it is still being sold.

The most commercially successful version runs: stories cause the brain to release oxytocin, oxytocin produces trust and generosity, therefore stories make people trust and buy from brands. The chain is attractive because each link sounds mechanical. The trouble is the first link.

Take them in order. The foundational result that intranasal oxytocin increases trust has been cited thousands of times and has not replicated well; the Envelope Task paradigm, which reported one of the largest behavioural oxytocin effects ever published, failed replication twice in studies with enough power to have found the original effect. The wider human oxytocin literature is characterised by small samples, disputed statistics and probable publication bias. Oxytocin has also been shown to increase envy and in-group favouritism, which sits awkwardly beside the warm reading the marketing version needs. The talk that popularised the idea now carries a corrections page published by the platform that hosted it.

The speaker–listener coupling work is genuine imaging research, and the extension to “your story synchronises your brain with your customer's” goes considerably further than the studies do. And the Hero's Journey is not an empirical claim at all — it is a comparative-literature thesis whose universality is disputed inside its own field. It remains a serviceable compositional tool. It is not evidence, and a strategy deck that cites it as though it were is making a category error.

None of this means storytelling does not work. Chapter four's meta-analysis stands. What it means is that the mechanism is psychological rather than pharmacological, that it is modest, and that the confident biological story told about it is the weakest part of the case rather than the strongest.

Six shapes, and what they do not license

There is one genuinely striking empirical result in this area, and it deserves better treatment than it usually gets.

Reagan and colleagues, publishing in EPJ Data Science in 2016, tracked emotional valence across the length of thousands of works of fiction and asked what shapes the resulting curves take. They selected 1,327 books from Project Gutenberg and classified arcs across a filtered subset commonly reported as 1,737 stories — the two numbers get used interchangeably in coverage, and they are not the same number. Four independent methods converged on the same answer: six recurring shapes account for most of it.

Now the caveats, which matter more than the result. The corpus is out-of-copyright English fiction, which is not a sample of stories in general and is certainly not a sample of commercial messages. “Success” in the paper is operationalised as download counts from a free public-domain library — a weak proxy that is not readership and is nowhere near persuasion. And most importantly, this is a decomposition of what exists, not a prescription. Nothing in the finding says a brand should adopt one of these shapes, and the industry habit of turning it into a template gets the logic backwards.

The counterweight

Every chapter of this study carries the same discipline, and this is where it bites hardest. The Ehrenberg-Bass position — set out in chapter two and not re-litigated here — holds that for most brands, most of the time, mental availability and distinctive assets do more commercial work than narrative depth. A brand nobody can retrieve does not benefit from having a beautiful story.

I think both are true, and I think the relationship between them is the useful part. This is my position rather than a finding, and I mark it as such: distinctiveness is the retrieval mechanism, which chapter four established is where choice actually happens. Narrative is the transmission mechanism — how the thing being made distinctive travels between people who have never met you. Story does not replace distinctiveness. It is how distinctiveness gets distributed by people other than you.

If you have to choose — and small companies usually do — chapter four comes first. Be retrievable, then be retellable. In that order.

What deforms, and what survives

Chapter four was about decay — the brand fading in a mind that is not being reminded. This chapter's risk is different and less discussed. A story that travels does not fade. It changes.

Anything retold is retold by someone reconstructing it, and reconstruction drops what it cannot hold. What survives five retellings is not what you wrote. It is the causal spine: who wanted what, what was in the way, what happened as a result. Ornament goes first — adjectives, qualifications, the carefully negotiated phrase your team spent a week on. Numbers usually go second, unless they are startling and round.

I want to label that clearly, because this study's rule is that borrowed authority gets named. The reconstruction research behind this is about eyewitness and narrative memory, not brands. The claim that your positioning deforms in a predictable direction — ornament first, structure last — is my extension from it, and it is an argument rather than a measured result.

But it produces a design instruction that is testable, and that instruction is the practical heart of this chapter: build the thing that survives, and stop polishing the thing that does not. If your positioning only works when delivered in your words, at your pace, with your slides, it is not positioning. It is a performance.

T3 — the Retelling Test

Every tool in this study so far has been something you fill in. This one is not. T3 is the first test in the study that other people have to pass on your behalf, which is exactly what makes it worth running.

Try this

Run T3 this week. It costs one afternoon and nothing else, and it is unusually unpleasant, which is a reasonable sign that it is measuring something.

Write your positioning as a single causal sentence — who wants what, what is in the way, what changes when it works. Strip every adjective. Read it once to five people who do not work with you and do not know your category well. Then say nothing for a day.

The next day, ask them to tell you what you do. Write down exactly what comes back, including the parts that are wrong. Score three things: did the structure survive, did the specific claim survive, and did your name survive attached to it. Then read the wrongness carefully, because the deformations are the most valuable output — they tell you which part of your story the world finds easiest to replace with something more familiar.

Most people running this discover the same thing. The structure survives, the specifics do not, and the name detaches. That is a brand doing transmission work for its whole category, which is generous of you and expensive.

So: a position gives you something worth retrieving. A memory map tells you what it must be retrieved from. A story is how it moves between people without you in the room. All three assume something that has not been examined yet — that the thing arriving at the other end still looks and feels like you when it gets there. Every retelling in T3 travelled as words. Almost nothing about a brand reaches anyone as words alone. It arrives as a surface: a colour before a sentence, a shape before a claim, a sound before a name, and a judgement formed in less time than it takes to read one. Which is where we go next.

Sources & method — every claim, cited

Sources — the Living Brand System (F1) and the Retelling Test (T3) are Edward Salvatierra’s frameworks; the mechanisms beneath them are not. On the twenty-two times claim: attributed almost universally to Jennifer Aaker (Stanford GSB), it does not appear in her published research. Independent attempts to source it terminate in a talk in which the figure is asserted without derivation, and no study — hers, Bruner’s, or anyone’s — has been produced. It is printed here as an example of transmission and explicitly not as evidence. If someone produces the study, this chapter is wrong and will say so. On the size of narrative persuasion: Braddock & Dillard (2016), “Meta-analytic evidence for the persuasive effect of narratives on beliefs, attitudes, intentions, and behaviors,” Communication Monographs 83(4), 446–467, DOI 10.1080/03637751.2015.1128555 — beliefs r = .17 (k = 37, N = 7,376); attitudes r = .19 (k = 40, N = 7,132); intentions r = .17 (k = 28, N = 5,211); behaviours r = .23 (k = 5, N = 978). The behaviours cell is printed with its k and N beside it because it is both the largest coefficient and by far the thinnest evidence on the chart, and it is the one most often quoted alone. An r of about .19 is roughly three per cent of variance explained: real, consistent, and small. On the mechanism: Green & Brock (2000), “The role of transportation in the persuasiveness of public narratives,” JPSP 79(5), 701–721. Used as a mechanism claim only — it explains why narrative reduces counter-arguing, and carries no magnitude, so none is attached to it here. On the neuroscience that did not survive: the foundational finding that intranasal oxytocin increases trust (2005) has been cited thousands of times and has not replicated well; the Envelope Task paradigm, reporting one of the largest behavioural oxytocin effects published, failed replication twice in adequately powered studies (Lane et al., 2015, PLOS ONE 10(9): e0137000). The wider human oxytocin literature is limited by small samples, disputed statistical approaches, probable publication bias and HARKing; oxytocin has also been found to increase envy and in-group favouritism, which the marketing reading omits. TED publishes its own corrections-and-updates page for the talk that popularised the idea. This chapter prints the failure rather than the claim. Stated fairly, as in chapter four: a failed replication is evidence that support is weaker and the effect smaller and more conditional than advertised — not proof of falsehood. On neural coupling: Hasson and colleagues’ speaker–listener fMRI work is genuine, and the marketing extension to “storytelling synchronises your brain with your customer’s” travels considerably further than the studies support. Marked here as a real finding, over-extended, and not used as load-bearing. On the Hero’s Journey: Campbell (1949) is a comparative-literature thesis, not an empirical result, and its universality claim is contested within its own discipline. Useful as a compositional tool, worthless as evidence — the same treatment chapter two gave the “what people say about you when you’re not in the room” line, which has no traceable primary source either. On the six arcs: Reagan, Mitchell, Kiley, Danforth & Dodds (2016), “The emotional arcs of stories are dominated by six basic shapes,” EPJ Data Science, arXiv:1606.07772. Note the two numbers, which coverage uses interchangeably and which are not the same: 1,327 books were selected from Project Gutenberg, and arc classification is reported over a filtered subset of 1,737 stories. Three caveats printed with the result rather than after it: the corpus is out-of-copyright English fiction, not a sample of stories in general and certainly not of commercial messages; “success” is operationalised as download counts from a free public-domain library, which is not readership and is nowhere near persuasion; and the finding is a decomposition of what exists, not a prescription for what to write. The industry habit of converting it into a template inverts the logic. On the counterweight: Sharp, How Brands Grow (2010, Oxford University Press) and Romaniuk, Building Distinctive Brand Assets (2018) — for most brands most of the time, mental availability and distinctive assets do more commercial work than narrative depth. That dispute was set out in chapter two and is not re-litigated here. My framing of the relationship — distinctiveness as the retrieval mechanism, narrative as the transmission mechanism — is my position, labelled as such in the text, not a finding. On deformation: the reconstruction literature carried over from chapter four — Bartlett (1932, Remembering); Loftus & Palmer (1974) — concerns eyewitness and narrative memory, not brands. The claim that positioning deforms in a predictable direction, ornament first and causal structure last, is my extension from it and is marked in the text as an argument rather than a measured result. T3 is a practical criterion built on that argument; it has not been validated as an instrument, and I would rather say so than dress it up. And the standing limit: almost none of the researchers cited here studied brands. Where their findings are carried across into commercial practice, the carrying is mine, and it is marked in the text every time — because I would rather you trust the study than be impressed by it.