You cannot move someone without knowing where they are standing

Every piece of work is a delta

A brief that says “we want people to know we are reliable” describes a destination and no starting point. Depending on where the viewer actually stands, the same destination needs four completely different pieces.

• If they have never heard of you, reliability is meaningless noise; you first have to exist.
• If they think you are expensive, a reliability claim reads as a justification for the price and hardens the objection.
• If they were let down by you once, a reliability claim reads as a lie, and the piece has to acknowledge the failure before it can claim anything.
• If they already trust you and are just slow to act, reliability is a wasted beat; what they need is a deadline.

So write two sentences, not one.

Before: She thinks a salon-supply delivery will take a week and probably arrive wrong.

After: She thinks this particular supplier gets it right the same day, and that the claim is checkable.

The work is the distance between those two sentences. Everything that does not shorten that distance is decoration.

The commonest failure: stating the after

Most weak work announces the after state. “Same-day delivery. Guaranteed.” Printed large, set well, beautifully spaced. It fails because it treats a belief as something you can install by declaring it, and beliefs do not work that way. The viewer already holds a belief; yours arrives as a competing claim from an interested party, and the existing one has the advantage of experience behind it.

What actually shifts a belief is evidence the viewer can check without trusting you first. A photograph of the dispatch board at 4pm with the day’s dockets on it. The line “we missed 11 of 2,400 deliveries last quarter: here is what we did about it”. A named customer with a phone number. None of these are more persuasive because they are cleverer; they are more persuasive because they cost you something to publish and a false claim could be caught.

Belief has three components, and they fail differently

When a piece does not move anybody, the stall is usually in one of three places, and the fix for each is different:

1. Attention. They never took it in. Nothing about your argument is the problem. Fix the first beat, the size, the placement, the contrast.
2. Comprehension. They saw it and did not understand what was on offer. Fix the language, the hierarchy, the number of ideas competing.
3. Belief. They saw it, understood it, and did not believe it. No amount of better layout fixes this. You need evidence, a guarantee, a third party or an admission.

Teams spend most of their time on the second, because it is the most enjoyable to work on. In my experience the third is where the traffic is being lost, and it is the only one that requires going back to the client and asking for something they may not want to give.

How to find the before state cheaply

• The objection log. Ask whoever answers the phone what people say just before they decide not to buy. Sales staff can recite this list instantly; it is almost never in the brief.
• Search suggestions. Type the category into a search box and read the autocompletions: “is X safe”, “X vs Y price”, “X refund”. Those are unedited before-states, at scale, free.
• The review floor. Two-star reviews of competitors, again. “Good product but the sizing runs small” is a before-state you can defuse in one line of copy.

Write the sentence the viewer would say

In their words, not yours. Not “perceives the category as unreliable” but “last time this took nine days and nobody answered the phone”. The reason is practical: you can answer a sentence, and you cannot answer an abstraction. Almost every good headline in this kind of work is a direct reply to a sentence somebody actually said.

What this cannot do

A before-and-after pair makes the job tractable; it does not make it achievable in one piece. If the gap is large (from “never heard of you” to “bought”) no single flyer or reel closes it, and pretending otherwise is how a piece ends up carrying five messages and landing none. When the distance is too big, split it: name the one intermediate belief you can actually install this time, and say plainly which piece of work does the next step.

It also cannot tell you whether the after state is true. If the client’s deliveries do take nine days, your job is not to build a more persuasive picture of same-day delivery. It is to tell them that the piece will work once and then cost them more than it earned, because the fastest way to destroy a claim is to make a lot of people check it.