Writing for the Reader in Front of You

Opening Scene

The same city council vote gets filed to three different desks by the same reporter. For the front page, it’s framed around what it means for residents: taxes, services, timelines. For the business desk, it’s framed around which contractors and industries stand to gain or lose. For the politics desk, it’s framed around who voted which way and what it signals about the next election. None of the three versions changes a single fact. All three lead with a different sentence, because a different reader is going to open each one, with a different question already in their head before they start reading.

A data team producing one dashboard and calling it done for every audience is skipping this step entirely — writing one story and running it, unedited, on every desk in the building.

In Plain English

Framing for the audience means deliberately choosing which true, accurate finding to lead with — and how to contextualize it — based on what the specific reader in front of you actually needs to know or decide. The underlying data doesn’t change. The entry point, the emphasis, and the level of technical detail do. A finding that leads with “conversion rate dropped 4%” for an analyst might need to lead with “we’re on track to miss the quarterly target by three weeks” for an executive, because the executive’s question isn’t “what happened to the metric” — it’s “do I need to act, and when.”

The Old Way

Without deliberate audience framing, data communication tends to default to one of these:

One report for everyone — the same dashboard, deck, or summary distributed unchanged to executives, analysts, and frontline teams alike, on the assumption that the data speaks for itself regardless of who’s reading it.
Framing for the author’s own comfort zone — an analyst instinctively writing at the technical depth they themselves find natural, regardless of whether the actual reader has the background to use that depth productively.
Guessing at the audience after the fact — building the report first and only considering “who is this for” once someone in a meeting asks a question the report doesn’t answer, rather than designing the framing in from the start.

Each of these treats “the data” and “the story” as the same object, when in fact the same data can support several honest, differently-framed stories depending entirely on who’s asking.

What’s Changing (and Why AI Is the Reason)

AI can now generate multiple audience-specific framings from the same underlying dataset at a cost that used to require rewriting the whole report by hand. What once meant an analyst manually producing an executive summary, a technical appendix, and a frontline briefing separately can now start from a single model-assisted pass that adapts tone, depth, and emphasis per audience.
This raises the bar on knowing your audience precisely, because the tool can now execute a framing decision much faster than it used to be able to make one. A model can write confidently for “an executive audience,” but only a human who actually understands what a specific executive needs to decide can specify what that framing should really emphasize.
Personalized framing at this speed makes consistency across framings a new and real risk. When three different audience versions of a finding are generated quickly, it becomes easier for them to drift into subtly different — or even contradictory — claims unless someone deliberately checks that all three still say the same true thing, just differently.

The Metaphor, Fully Extended

Newsroom Element	Data Storytelling Concept
The same council vote filed differently to three desks	The same finding reframed for three different reader audiences
The front-page framing built around what residents need to know	Framing a data story around what the reader’s actual decision or concern is
One wire story run unedited on every desk in the building	A single dashboard or report distributed unchanged to every audience
A reporter writing at the depth they personally find comfortable	An analyst defaulting to their own technical comfort zone rather than the reader’s
A wire service drafting a front-page version and a business-desk version from one underlying report	AI generating multiple audience-specific framings from one underlying dataset

For Beginners: What to Actually Do

Before writing a report or building a dashboard, name the specific reader you’re writing for and the one decision or question they actually have in mind.
Practice rewriting the same finding’s opening sentence twice — once for a technical peer, once for someone with no data background — and notice what changes and what doesn’t.
Never assume “the data speaks for itself” is enough; the same accurate data supports multiple honest stories, and choosing among them is your job, not the data’s.
When you don’t know who the actual audience is, ask before you build the report rather than guessing and revising after a confused meeting.

For Practitioners and Leaders: The Deeper Layer

Build audience framing into your team’s report templates as an explicit, named step — “who is this for, and what do they need to decide” — rather than leaving it as an unstated instinct some analysts have and others don’t.
When using AI to generate multiple audience-specific versions of a report, review all versions together for consistency, since fast parallel generation makes silent drift between framings a real and easy-to-miss risk.
Recognize that framing skillfully for an audience is not the same as slanting or cherry-picking; the underlying claim must remain equally true and defensible across every version.
Treat this skill as a genuine specialization worth developing on your team, the same way a newsroom treats its beat reporters as distinct from its wire editors — knowing an audience is its own expertise.

Quick Recap

Audience framing means choosing which true finding to lead with, and how to contextualize it, based on what a specific reader actually needs — the underlying data doesn’t change.
Common failure modes are one report for everyone, framing for the author’s own comfort, and guessing at the audience only after the fact.
AI can now generate multiple audience-specific framings quickly, but knowing what each audience actually needs remains a human judgment.
Fast, parallel framing generation introduces a new risk: different audience versions of the same finding silently drifting apart unless someone checks them against each other.

Where This Fits in the Series

Article 2 established leading with the headline; this article covers choosing which headline, for which reader. Article 4 looks at the structural principle that applies once you know your lede and your audience: ordering the rest of the piece so the most important point always comes first.