Storytelling Theory: Narrative Inquiry as Organisational Research

Storytelling theory treats the stories told inside an organisation not as decoration around the real business but as the mechanism through which people make sense of what happens to them. On this account, an organisation is substantially constituted by its narratives, and studying them is studying the thing itself rather than commentary on it.

The position matters methodologically because it changes what counts as data. An interview transcript is normally mined for statements of fact or opinion. In narrative inquiry the shape of the account — what is told as a beginning, who is given agency, what is left unresolved — is the finding.

The three positions worth distinguishing

The field is usually introduced through three authors whose claims differ more than summaries suggest.

Gabriel treats stories as relatively well-formed narratives with plot, characters and a point, told for effect, and distinguishes them from reports and opinions. His interest is in poetic licence: the ways a story departs from what happened are precisely what reveal what the teller wants it to mean. He classifies by plot type — comic, tragic, epic, romantic — and reads the type as evidence of how the teller positions themselves.

Czarniawska treats organising itself as a narrative accomplishment. Organisations are, in her account, sites where narratives are produced, circulated and contested, and the researcher's job is closer to that of an anthropologist of texts than an interviewer collecting accounts.

Boje goes furthest and is the source of the term most associated with the field. He argues that the polished, retrospective story is the exception. What actually circulates is fragmentary, interrupted, incomplete: half a story told in a corridor, finished by someone who was not there. He calls these fragments antenarrative — both prior to narrative and a bet on a meaning not yet settled — and his objection to conventional narrative analysis is that by collecting whole stories it studies only the ones that have already been tidied.

What the method involves

Collection. Stories are rarely produced on request. Asking "tell me a story about this organisation" generally yields the official one. Naturalistic collection — sitting in on meetings, listening in shared spaces, noting what gets retold — produces material that the interview does not reach, at considerable cost in time and access.

Provoking without dictating. Where interviews are used, the productive questions ask for incidents rather than assessments: the day something went wrong, the thing new starters get told in their first week, the decision people still argue about. These invite narrative structure without specifying its content.

Analysis. Several approaches coexist. Structural analysis examines how the story is built — orientation, complication, resolution, coda — following Labov's scheme. Thematic analysis across stories looks for recurring plots. Deconstruction, which Boje sets out as an eight-step procedure, works on what a story suppresses: whose voice is absent, what the hierarchy of terms assumes, what the story would look like told from the other side.

Reporting. The finished work quotes at length, because a story summarised has had its structure removed, and the structure was the evidence.

What stories are evidence of

The most common error in student work is to treat a story as a report of events, and then to worry that it might be inaccurate.

A story is evidence of how the teller construes what happened, of what they expect the listener to accept, and of what the organisation permits to be said. Its factual accuracy is a separate question and frequently the less interesting one. A story about a founder that is demonstrably false is strong evidence about the organisation's self-image, and treating its falsity as a problem misunderstands what was collected.

This is also the answer to the objection that people embellish. Embellishment is data. The question is not whether the teller exaggerated but what the exaggeration was for.

Where it is applied

Organisational culture. Stories are the most accessible surface of culture, and a change of culture can be tracked through changes in which stories circulate and which stop being told.

Change and resistance. Managers narrate change as a journey with a destination; those affected frequently narrate it as something that was done to them. The two narratives coexist, and resistance is often better described as a competing story than as an attitude.

Identity. Both individual and organisational identity are argued to be narrative in form — a self is the story a person can tell about themselves that holds together — which makes narrative methods the natural approach to studying how people absorb a restructuring or a merger.

Knowledge transfer. The observation that expertise travels through war stories, not through documentation is one of the field's more durable practical findings, and it underlies a good deal of what is now called organisational learning.

A worked illustration of what the analysis produces

A researcher studying a hospital trust two years after a merger collects, among others, this account from a ward manager:

"They came round with the new structure on a Friday. I found out my ward had been split between two directorates from a laminated chart on the wall. Nobody had told me. I rang the programme office and got a voicemail. So I did what you do — I sorted it out with the other manager over coffee, and we've run it that way since. Officially it's still split."

Read as a report, this says a communication failure occurred and was worked around. Read as narrative, several things emerge.

The plot is tragic in Gabriel's classification: an undeserved misfortune inflicted by an impersonal agent. Agency is given to "they" and to the chart; the teller is positioned as someone things happen to, right up to the turn where they act.

The coda does the heaviest work. "Officially it's still split" is offered as the resolution and simultaneously reports that nothing was resolved. The organisation is described as having two states, the official and the actual, with the teller living competently in the second.

What is absent is any account of why the split was made. The story contains no version of the reorganisation's purpose, which suggests either that it was never communicated in a form that survived, or that it was not persuasive enough to be worth retelling.

Collect thirty such accounts and the recurring structure — impersonal agent, discovery by artefact and not by conversation, local repair, a coda acknowledging the official position as fiction — is a finding about the merger that no satisfaction survey would produce, because the survey asks whether communication was good and this material shows what people did about it not being.

The criticisms

Selection. The researcher chooses which stories to report, and a vivid story is more likely to be selected than a representative one. The safeguard is transparency about the corpus — how many stories, from whom, and how the reported ones were chosen — and it is frequently absent.

Interpretation without constraint. Deconstruction in particular can produce almost any reading, since what a text suppresses is limited only by what the analyst can imagine. The discipline has to come from grounding each claim in the text and from showing enough of it that a reader can disagree.

Generalisation. Stories are local by nature, and claims drawn from them apply to the setting studied. This is the standard limitation of interpretive work and is only a problem when the writer forgets it in the conclusion.

Access and time. The naturalistic collection that the theory calls for requires sustained presence, and most taught-programme dissertations cannot supply it. The honest compromise is to acknowledge that interview-elicited stories are a partial substitute, and to say what that costs — mainly the fragmentary, in-progress material Boje regards as most revealing.

Ethics, which are heavier here than in most methods

Narrative work raises consent and anonymity problems that a survey does not, and a methodology chapter that treats them as routine is a weak one.

A story is identifying in a way an attitude score is not. An account of a specific incident, told in the teller's own phrasing, can be traced to its source by anyone who was in the organisation at the time, however carefully names are removed. Anonymisation therefore has to alter detail, and altering detail damages exactly the material the analysis depends on. There is no way round this; the choice is which loss to accept, and the chapter should say which was chosen.

Consent is also less stable. A participant who agreed to an interview did not necessarily agree to have the structure of their account analysed for what it conceals, and deconstruction in particular reaches conclusions the teller would not recognise or welcome. Practice varies, and returning the analysis to participants is one answer, at the cost that participants may object to a reading that is nonetheless defensible.

Using it in a dissertation

The approach suits a question about meaning, identity or culture, and suits a question about frequency or effect very badly.

Three practical points raise the quality of a chapter. State which position is being taken, because Gabriel's whole stories and Boje's fragments call for different collection and different analysis, and a chapter citing both without choosing has not designed anything. Describe the corpus in numbers as well as in character, so the selection can be judged. And quote at length in the findings, accepting that fewer stories treated fully demonstrate more than many stories reduced to a sentence each.