Case Study 1 — Dana's Findings Paragraph: One Reorder, No New Words

A worked scenario showing the given-new contract repair a real-feeling paragraph end to end. The numbers and names are a composite, fictional but realistic.

The situation

Dana Whitfield, a data scientist, has finished her customer-churn analysis and is drafting the "Findings" section of a memo for Renée Okafor, the VP of Marketing. Dana knows her statistics cold. She also knows—from Chapter 2—that Renée is a decider, not a peer, so the memo leads with the recommendation elsewhere. But this particular paragraph, the one explaining why customers leave, keeps coming back from her teammate marked "a bit hard to follow." Dana is frustrated. Every sentence is true and clear. What's the problem?

Here is the paragraph her teammate flagged:

A 34% higher churn rate was observed among customers who filed more than two support tickets in their first 30 days. Onboarding friction is the most likely explanation for this elevated churn. A setup wizard that confuses new users generates the bulk of those early tickets. The single highest-impact intervention, according to our model, would be a redesign of that wizard.

Read it aloud. You stumble at every period. But you can't point to a broken sentence, because there isn't one.

The diagnosis

Dana's teammate was right but couldn't say why. Apply this chapter's diagnostic: are the sentences individually clear? Yes. Is the reading bumpy anyway? Yes. That combination means a cohesion problem—the failure is in the seams, not the sentences (§8.5). So Dana shouldn't rewrite the sentences. She should check the information order.

Look at where each sentence starts. Sentence one starts with "A 34% higher churn rate"—the new finding, fine for an opener. But sentence two starts with "Onboarding friction," a brand-new idea, and only reaches back to sentence one at its very end ("this elevated churn"). The reader meets "Onboarding friction" with nothing to attach it to; they hold it in suspension until the end of the sentence reveals the connection. Sentence three does it again—it opens with "A setup wizard," new, and connects backward ("those early tickets") only at the end. Every sentence is built new-first, given-last: exactly backward from what the reader expects (§8.3). That is the entire cause of the "hard to follow." Not the words. The order.

The repair

Dana reorders the information inside each sentence so the given (familiar) comes first and the new lands at the end—the stress position. She changes facts in none of them.

Customers who filed more than two support tickets in their first 30 days churned at a 34% higher rate. That early-ticket burden most likely reflects onboarding friction. Most of those early tickets, in turn, trace to one source: a setup wizard that confuses new users. Redesigning that wizard is, by our model, the single highest-impact intervention available.

Now trace the handoffs:

• Sentence 1 ends on "34% higher rate" (the finding).
• Sentence 2 opens with "That early-ticket burden"—reaching back to the support tickets just mentioned—and ends on "onboarding friction" (new).
• Sentence 3 opens with "Most of those early tickets"—reaching back to onboarding/tickets—and ends on "a setup wizard that confuses new users" (new).
• Sentence 4 opens with "Redesigning that wizard"—the wizard just named, now given—and ends on "the single highest-impact intervention" (new, and the point that matters to Renée).

Each sentence's new becomes the next sentence's given. That is known-new chaining (§8.3), and it's why the second version pulls the reader forward where the first jolted them at every seam.

What changed, and what didn't

Count the differences. Dana did not add a single fact. She did not add transition words—there's no "however," no "therefore"; the connections are carried by the information order itself, which is the most invisible (and theme-7) way to do it. She did not make the sentences notably shorter or longer. She reordered the information within each sentence. That's the whole intervention.

This is the lesson Dana most needed, and the one most writers resist: when prose feels choppy, the instinct is to rewrite the sentences to be "smoother," but the sentences were never the problem. Reorder before you reword. Nine times in ten, flow appears without your touching the words, because flow was never in the words—it was in the seams.

The payoff and the connection

Renée reads the revised paragraph once, at full speed, and gets it: too many early tickets → onboarding friction → a confusing wizard → fix the wizard. The causal chain is now felt rather than reconstructed. That is what a finding-explaining paragraph is supposed to do—carry the reader frictionlessly from symptom to cause to action.

There's a forward thread here too. In Chapter 27, Dana's full churn memo gets the three-version treatment (methodology-first → findings-first → recommendation-first)—a document-structure problem (Chapter 4's territory). This case is the same memo viewed through a microscope: even once the document is ordered correctly, each individual paragraph still has to flow, sentence to sentence, or the well-structured memo reads like gravel. Structure (Ch 4) orders the sections; cohesion (Ch 8) orders the sentences inside them. You need both. A perfectly structured memo made of choppy paragraphs still fails the reader—just at a finer grain.

Try it yourself

Here is one more sentence Dana could add to the paragraph, written new-first. Reorder it so it opens given and ends new, then check against the paragraph's flow:

"A six-week build estimate was given by engineering for the wizard redesign."