/* global window */
// Content object for the article renderer (blog/blog-article.jsx).
// Produced by the Writer→Editor→Humanize→Proofreader chain; draft in blog/pipeline/drafts/marlow-messaging.md
window.__ARTICLE__ = {
slug: "marlow-messaging",
laneLabel: "CASE STUDY",
kicker: "MESSAGING",
readMins: 10,
dateLabel: "Jul 2025",
title: "How a DTC brand killed three messaging directions before spending a dollar on the wrong one",
deck: "A direct-to-consumer brand ran four value-prop directions through real customer reactions before launch. Three died on contact. One tested strongly. The discipline that saved the launch: pre-test the message, don't launch and pray.",
tags: ["brand", "messaging", "synthetic"],
toc: [
{ id: "setup", num: "01 · THE SETUP", title: "58 buyers, two versions" },
{ id: "why-test", num: "02 · THE CASE", title: "Test before you spend" },
{ id: "the-dead", num: "03 · WHAT DIED", title: "The three that died" },
{ id: "survivor", num: "04 · THE WINNER", title: "The one that survived" },
{ id: "pattern", num: "05 · THE PATTERN", title: "What the survivor had" },
{ id: "closing", num: "06 · THE VERDICT", title: "The case for pre-testing" },
],
body: [
{ t: "p", html: `A direct-to-consumer brand came to us a few weeks out from relaunching its site, with a problem most teams don't admit they have. The new site was beautiful. The marketing lead loved it. The founder loved it. Nobody in the building could agree on what it should say. There were four candidate directions for the headline value prop, each with a champion, and the plan on the table was to pick the one the room liked best and ship it.` },
{ t: "p", html: `That is the moment most messaging dies. Not at launch, where it's expensive to fix, but in a conference room where the people who built the thing vote on what it means. So we did the unglamorous version instead: we put the actual directions in front of the actual buyers, recorded the reactions, and let the reactions decide. Three of the four directions died. One survived and tested strongly. The point of this piece is what killed the three, because the failures are more instructive than the winner.` },
{ t: "h2", id: "setup", num: "01 · THE SETUP", text: "58 buyers, two versions, one question" },
{ t: "p", html: `We recruited 58 usable participants out of more than 1,100 who entered the study, screened to people who actually buy in this category and are in-market right now. AI-moderated interviews, each reaction tied to a clip, a transcript line, and a behavioral signal. We split them: 31 saw the new direction, 27 saw the existing one, so we could measure the delta a message creates rather than just collecting opinions about a single page in a vacuum.` },
{ t: "p", html: `The cohort skewed the way the category does: majority employed, mostly under 35, most of them new to the product and feeling more excited than stressed about the decision ahead. That last part matters. These weren't skeptics we had to win over. They walked in warm. Which makes the directions that still died more damning, not less.` },
{ t: "pullquote", text: "Most messaging doesn't die at launch, where it's expensive to fix. It dies in a conference room, where the people who built the thing vote on what it means." },
{ t: "h2", id: "why-test", num: "02 · THE CASE", text: "Why you test the message before you spend on it" },
{ t: "p", html: `The case for pre-testing is boring and overwhelming, which is exactly why teams skip it. Message quality is not a rounding error on a launch. One analysis of message testing found that when marketing copy works, it is roughly twice as influential as design in converting customers (Wynter, 2024). You can pour the budget into a gorgeous site and still lose, because the words are doing more of the conversion work than the visuals everyone argued about.` },
{ t: "p", html: `And the words fail more often than teams expect. The pattern shows up every time someone runs the test: a target audience that can't follow the value prop the company is in love with. In one published message test, a B2B team discovered their own language "felt too complicated, inhuman, and sales-like" to the people it was written for, something no internal review had flagged (Maze, 2024). The room is the worst possible judge of its own messaging, because the room already knows what the product does. The buyer doesn't. That gap is the whole game.` },
{ t: "p", html: `The cost of getting it wrong compounds after launch. Only about 60% of new consumer packaged goods survive two years on the shelf, and weak positioning is a reliable way to join the other 40% (GLG, 2024). A killed messaging direction costs you a research study. A launched one that misses costs you the launch.` },
{ t: "p", html: `The good news: you don't need a huge sample to kill a bad direction. Message testing reaches thematic saturation fast — the qualitative literature puts it around 9 to 17 participants, with practitioners often calling it at 12 or 13 (Wynter, 2024). By the time the tenth person stumbles on the same word, you don't need the eleventh to confirm it's dead.` },
{ t: "h2", id: "the-dead", num: "03 · WHAT DIED", text: "The three that died" },
{ t: "p", html: `Here is what the reactions actually killed, in order of how confidently they killed it.` },
{ t: "p", html: `Direction one: "the functional planning tool." This was the incumbent message — describe the product by what it does. Organize your stuff. Keep track. A capable utility. It was clear, it was accurate, and it was inert. 70% of people who saw this direction described the product in purely functional terms; they understood exactly what it did and felt nothing about it. The most telling number was the absence: zero people in this group reached for an emotionally positive word like "soothing" or "inspiring" to describe it. The direction communicated what without ever touching why anyone would care. It didn't offend anyone. It just didn't move.` },
{ t: "figure",
fig: { key: "transcript", props: { time: "GROUP B · OLD", speaker: "P-14 · functional-tool direction", children: `"I think it's decent. It looks a little scarce... and a little bit dull."` } },
ref: "FIG 01",
caption: "The incumbent message was accurate and inert — understood, never felt. Verbatim anonymized." },
{ t: "p", html: `Direction two: "the smarter, hyper-personalized AI." This was the favorite going in — lean into intelligence, promise a bespoke, tailored, reads-your-mind experience. It tested as a real driver, but only for a minority: 29% believed the product would deliver a genuinely personalized experience and lit up at the idea. The other 71% either didn't believe the promise or didn't want it that strong. Worse, leaning on "smarter" had a cost nobody anticipated. The more capable and polished the message made the product sound, the more scrutiny it invited — 52% started asking harder questions precisely because the thing looked impressive. A "smarter" tool carries higher stakes, so people evaluate it more rigorously, not less. The personalization direction wasn't wrong; it was a feature, oversold as the whole story.` },
{ t: "p", html: `Direction three: "trust us with your data." Nobody proposed this one out loud, but it was implicit in the AI-assistant framing, and the reactions dragged it into the light. A full 32% voiced explicit, unprompted concern about handing personal information to an AI. This wasn't a vocal fringe; it was a third of the room, and it was the single most consistent barrier in the study. Any direction that asked for trust before earning it ran straight into this wall. The fix wasn't a better privacy policy buried in a footer. It was making safety a visible part of the message instead of a thing the buyer had to go looking for — and the directions that ignored it died on contact with the most cautious segment.` },
{ t: "figure",
fig: { key: "insight", props: { accent: true, title: "The data-trust barrier killed any message that asked for trust it hadn't earned", clips: 19, participants: 58, segments: 4, confidence: 0.78, body: `32% voiced unprompted privacy concern — the single most consistent barrier in the study. Cuts across every direction that led with "smarter AI" before making safety visible.` } },
ref: "FIG 02",
caption: "Finding card: the trust barrier, scored across the full n=58 cohort." },
{ t: "p", html: `Three directions, three different failure modes. One was true but inert. One worked for a minority and backfired with the majority. One asked for trust the message hadn't earned. None of them would have announced themselves in a conference room, because in the room, everyone already trusts the product, already wants the personalization, and already knows what it does.` },
{ t: "h2", id: "survivor", num: "04 · THE WINNER", text: "The one that survived" },
{ t: "p", html: `The direction that tested strongly wasn't on anyone's shortlist as the lead. It came out of watching where the emotion actually spiked. When people reacted to the new site, the strongest, most consistent positive response wasn't to any single feature. It was to a feeling: the design reframed an overwhelming, stressful task as an exciting, manageable one. 77% praised the aesthetic, and they didn't praise it as decoration — they praised it as relief. The visuals were doing emotional work, taking an anxious buyer and making them feel calm and capable.` },
{ t: "p", html: `Underneath the relief sat a concrete promise people did believe: control. 71% valued the product's ability to act as an organized command center over the chaos — the single most stressful part of the job, by their own account. So the surviving message wasn't "it's beautiful" and it wasn't "it's smart." It was the bridge between the two: turn the thing you're dreading into a plan you can see. Inspiration that resolves into control. The aesthetic earns the attention; the organization keeps the promise.` },
{ t: "p", html: `The proof was in the delta. Against the functional incumbent, the new direction produced a 62% jump in positive aesthetic reactions and shifted how people described the product — from a passive "tool" to an active "experience," even a "personal consultant." The biggest swing came from exactly the people you'd most want to convert: among the most stressed buyers, mentions of "calm" or "confidence" jumped 37% versus the old message. The direction worked hardest on the segment with the most to gain.` },
{ t: "pullquote", text: "The surviving message wasn't “it's beautiful” and it wasn't “it's smart.” It was the bridge between the two: turn the thing you're dreading into a plan you can see." },
{ t: "h2", id: "pattern", num: "05 · THE PATTERN", text: "What the survivor had that the others didn't" },
{ t: "p", html: `Line the four up and the pattern is clean. The functional direction was accurate but emotionally void. The personalization direction was exciting to a minority and alienating to the rest. The implicit data-trust direction asked for something it hadn't earned. The survivor did three things none of the others managed at once:` },
{ t: "ul", items: [
`It led with the buyer's emotional state, not the product's feature list.`,
`It resolved that emotion into a concrete, believable benefit (control) rather than an oversold one (intelligence).`,
`It left room to address the trust barrier head-on rather than walking into it.`,
] },
{ t: "p", html: `This is what a message test buys you that a vote in a room never will. The room optimizes for the message the builders are proudest of. The test optimizes for the message the buyer can't argue with. Those are almost never the same sentence. In this study, the favorite going in (the "smarter AI" direction) tested third. The incumbent that felt safe tested dead last. The winner was a framing nobody had written down as the headline until the reactions pointed at it.` },
{ t: "p", html: `This pattern isn't unique to one category. Across our consumer work, the same shape recurs: in a separate study of a premium apparel brand, the message the team assumed would carry the brand (heritage and prestige) ran straight into a barrier they hadn't priced in — buyers admired the brand but found it "too fancy" to actually reach for day to day. The thing the company was proudest of was the thing keeping the product in the closet. Different category, identical lesson. The message that wins is rarely the one the building is in love with.` },
{ t: "h2", id: "closing", num: "06 · THE VERDICT", text: "Why this is the case for pre-testing" },
{ t: "p", html: `The discipline here is simple and almost nobody follows it: before you spend on a launch, make every messaging direction defend itself against real reactions, and tie each verdict to something you can replay. The clip where the buyer's face goes flat at the functional headline. The transcript line where "smarter" triggers a privacy question instead of excitement. The behavioral signal of the stressed buyer leaning in when the message names their stress back to them. The segment pattern that proves the personalization promise only lands for 29%. An honest confidence indicator on each direction, so you know which verdicts are solid and which are still thin.` },
{ t: "p", html: `That's the difference between "the team likes direction two" and "direction two converts a third of buyers and raises the guard of the other two-thirds, here's the clip, here's the 52%." One is a preference. The other is a verdict you can defend when the launch gets second-guessed — and launches always get second-guessed.` },
{ t: "p", html: `Three directions died in this study. Killing them cost a few weeks and one round of interviews. Launching any of them would have cost the relaunch. The brand shipped the survivor, the one with the evidence behind it, and never spent a dollar finding out the hard way that the other three didn't work.` },
{ t: "pullquote", text: "A killed messaging direction costs you a research study. A launched one that misses costs you the launch." },
{ t: "references", items: [
{ n: 1, html: `Wynter (2024). "How to conduct message testing: The definitive guide." Reports that when marketing messages work they are roughly 2x as influential as design in converting customers, and that message testing reaches saturation at roughly 9–17 participants (often called at 12–13). wynter.com` },
{ n: 2, html: `Maze (2024). "How to test marketing messaging before launch." A pre-launch message test surfaced that the team's own language "felt too complicated, inhuman, and sales-like" to its target audience — something no internal review had caught. maze.co` },
{ n: 3, html: `GLG (2024). "Use science to make your product launch successful: concept testing and message testing." Reports that only about 60% of new consumer packaged goods survive two years on the shelf, underscoring the cost of weak pre-launch positioning. glg.com` },
] },
],
related: [
{ href: "/blog/sample-report.html", title: "See a full NeroView report end to end — the challenger sportswear study, n=259.", meta: "14 min · Report" },
{ href: "/blog/synthetic-vs-human.html", title: "Synthetic vs. human research: when each one wins", meta: "7 min · Index" },
],
};