Walk into any mid-sized company on any Monday morning and you'll find a senior analyst staring at a chart, saying some version of: "The chart is fine. They just don't understand what it's saying." It is almost always the same diagnosis, and it is almost always wrong. The chart isn't the problem. The story around the chart is missing, and the chart alone cannot carry the weight.
This is the quiet truth of analytical communication: the hard work is not making the chart. The hard work is deciding what the chart means, arranging the surrounding claims so the meaning lands, and making the reader care enough to act. This is storytelling. Analysts are often uncomfortable with the word — it feels unserious, adjacent to marketing — but the best analysts we've ever worked with were, without exception, people who could write.
Borrow from journalism, not from marketing
The storytelling traditions worth borrowing from are not the ones used to sell things. They are the ones used to explain things. Journalism, the investigative kind, has spent a century refining how to take a complicated situation and make a reader understand it without dumbing it down. Its conventions — the lede, the nut graf, the kicker — are not stylistic affectations. They are load-bearing.
The lede is the opening sentence or two, and its job is to tell the reader why this story is worth their time. In analytical writing, the lede is the thing you'd say if someone stopped you in the hallway and asked "so what did you find." Not "we looked at conversion by channel." That's a topic. A lede would be: "Paid search is quietly subsidizing the performance of every other channel in the marketing mix." That's a claim. That's a lede.
The nut graf is the paragraph that explains why the story matters and what the reader is about to learn. In analytical writing, it previews the structure: here are the three things we found, here's the action we think follows. Most analyst write-ups skip this entirely and jump into methodology. The reader is not yet ready for methodology. They don't know why they should care.
The kicker is the close. What the reader should carry away. In analytical writing, this is usually the recommendation, and it is usually weaker than it should be — hedged, caveated, softened into uselessness. Journalists teach us to land a kicker cleanly. Analysts can learn to do the same without abandoning intellectual honesty.
A good analysis makes a claim. The claim is load-bearing. Everything else in the write-up is in service of that claim, or in service of the honest qualifications to it.
The structural moves that make analytical writing work
1. Lead with the finding, not the method.
Analysts are trained, sometimes explicitly, to walk the reader through their reasoning. "We pulled the data from X, joined to Y, filtered to Z, computed A, and found B." This is the inverse of how busy readers process information. They want B first. If B is interesting, they will ask about A. Most of the time, they won't ask about A — they will accept your competence, act on B, and move on. That is fine. That is the correct outcome. Save your methodology for an appendix, or a follow-up question.
2. Anchor every number to a story.
A number on its own has no weight. "Churn is 7%" is a fact; the reader does not know whether this is good or bad, expected or surprising, stable or worsening. "Churn is 7%, up from 4% a year ago, and the increase is almost entirely driven by our mid-market segment" is a story. Every headline number in your write-up should come with the context that makes it meaningful — what it was, what it should be, what's driving the gap.
3. Use contrast to create stakes.
Stories need tension. In analytical work, the tension usually comes from contrast. The expected vs the observed. The company average vs the specific segment. Last period vs this period. The forecast vs the actual. Without contrast, a reader has nothing to hang their attention on. Find the contrast in your data and lead with it.
4. Show, but then also tell.
Analysts, especially experienced ones, lean on "show, don't tell" as a principle. Put the chart in, the reader will see what you saw. This underestimates the reader's attentional budget. Most of the time, you need to show and tell — include the chart, and then write one sentence naming what the reader should see in it. The chart's title can do some of this work (see our piece on dashboard titles), but a sentence of prose immediately after the chart, written at the right level of abstraction, is the highest-leverage piece of writing in most analytical reports.
5. Own the uncertainty, but don't hide behind it.
Analysts are professionally trained to qualify. Sample sizes, confidence intervals, data-quality caveats. All of these are valid and sometimes essential. But a write-up that is 70% caveat and 30% claim is not intellectual honesty — it is a refusal to commit. The reader is a decision-maker. They will act on something. If you refuse to say what to act on, you have abdicated the analyst's job and outsourced the judgment back to the person with less data. That is almost always a worse outcome.
The better move is to separate "what I believe" from "what I'm less sure about," clearly, and commit to the former. "I am highly confident the trend is real. I am less confident about the driver. My best guess is X, and here's what would change my mind." That reads like a professional. The alternative reads like a person trying not to be wrong.
6. End on an action, not a summary.
Most analytical write-ups close with a recap of what was covered. This is the weakest possible ending. The strongest ending is a recommendation: a specific action the reader should consider, a next investigation worth running, or a decision the data makes clearer. The reader's question when they finish reading is "what should I do with this?" Answer that question explicitly. If you can't, you don't yet have a complete analysis.
The deeper point: analysis is a persuasive act
There is an honest disagreement in the analytical profession about whether the role is to inform neutrally or to recommend persuasively. We take a strong view: the best analysts are persuasive, and they are persuasive with rigor. They make the case. They anticipate the objections. They land the argument. And they do all of this while remaining genuinely open to being wrong — updating their view if the data demands it, publicly revising if a premise turns out to be faulty.
The analyst-as-neutral-observer is, in our view, a fiction. Every choice of what to measure, what to compare against, what to show and what to leave out, is already a persuasive act. The question is whether it's done consciously and in good faith, or unconsciously and defensively. The first is a craft. The second is a cop-out.
Every analysis is already persuasive. The question is whether it persuades well or badly. Pretending otherwise only makes it worse.
A practical exercise
The next time you finish an analysis and sit down to write it up, try this. Before you write anything, open a blank document and write one sentence: "The most important thing I found is ______." That sentence is your lede. If you can't write it, you aren't ready to write the report — go back and figure out what your analysis actually found. If you can write it, you now have the spine of the entire write-up. Every chart, every section, every piece of methodology is there to either support or qualify that sentence.
This is a discipline you can practice. Most analysts have never been asked to. The ones who build the habit become, reliably, the most trusted voices in their organizations — not because they have better SQL, but because they can write a sentence that a CEO can act on. That is a rare and expensive skill. The good news is that it can be learned.
A final thought
The gap between a competent analyst and a great one is almost never the analysis. It's the communication. Two people can pull the same data and reach the same finding. One of them writes it up in a way the business reads, debates, and acts on. The other writes it up in a way that lives unread in a Confluence page. The first one gets promoted. The second one complains, with some justification, that leadership doesn't understand data.
The fix is not for leadership to understand data. The fix is for the analyst to learn to write. It is, by almost any measure, the highest-leverage skill an analyst can build. And — this is worth saying plainly — it is a skill. Not a personality trait. Not an aptitude. A craft, learnable, improvable with reps. Start today.