3. Writing a Positioning Claim You Can Trace Back to a Line
Choosing one situation and declining the others
Ledgerlane's hero section is accurate and aimed at nobody. The missing piece is a decision no evidence file can make: which situation the message speaks to, and what it stops speaking to.
Three independent lines in the evidence file describe the same trigger — months of nothing, then a penalty notice or an unexpectedly large tax bill, then someone looking for help that week. One line stands alone: a contractor stung by a relative's neat books. The chapter explains why a message built on a single line fails unreadably, and sets a promotion rule — two independent sources to reach emerging support, three to carry a message of its own.
The competitive set is five options: spreadsheet or glovebox, the once-a-year accountant, a spouse, self-managed software at ten to thirty-five pounds a month, and doing nothing. The incumbent is the year-end accountant plus inertia, which makes the difference a matter of when rather than how good — the fourth piece of positioning most drafts skip.
Boundaries are named: accountant and bookkeeper are unprotected titles, chartered designations are not, and anyone providing bookkeeping or tax services as a business must register for money-laundering supervision before trading.
The benefit is written for a problem-aware buyer in the vocabulary of the lines themselves. Exclusions work as a filter that improves who books. Proof splits into three piles, with the Self Assessment and CIS penalty timetable — one hundred pounds at a day late, ten pounds daily from three months, VAT registration at ninety thousand, Making Tax Digital from April 2026 — counted as proof of understanding the customer's year, not a claim about Ledgerlane. Client counts, testimonials and tax-saved figures stay out.
The draft came back with three faults: an uncited benefit sentence about confidence, the single-source line promoted into proof, and a citation claiming monthly visibility the lines never mention. All three, and the revision, are shown.
Getting more out of the approved copy
Adapting the hero section to other channels under claim parity: format and pacing change, every number, timeframe, qualifier and exclusion holds. Channel limits set the shape — email preview truncation, the See more cut, two hundred and eighty characters on X. Ida is asked for a claim breakdown before the variant. When a mandatory qualifier will not fit, link to full terms or pick a different claim.
Ledgerlane has a hero section saved in the project, and every number in it can be traced to a line in the product-facts sheet. The price is right. The two-week onboarding says "from signing." The exclusions are on the page. As a piece of factual writing it holds up.
It is also aimed at nobody in particular.
That is not a tone problem. It is a decision that has not been made yet. The evidence file has twelve labelled lines in it, grouped into named situations, each group tied back to the lines that produced it. Those groups tell you which situations contractors are actually in. They do not tell you which situation the message should speak to. Nothing in the file can tell you that, because it is not a fact about contractors. It is a choice about who Ledgerlane is going to talk to first, and what it is willing to stop talking to in order to do it.
This is the part Ida cannot do for you. She can assemble, cite and flag. Ask her to choose, and she will pick whatever the grouped lines make loudest, which is not the same as whatever is smartest to sell against. So this chapter does two things at once. It makes the choice, out loud, with reasons. And it turns that choice into a document you can hand to any future request: an audience-and-offer brief, six named parts, each part carrying the evidence line numbers that support it. Everything in the course so far has been about traceability of claims about the product. This extends it to claims about the customer.
Choosing one situation and declining the others
Start with the audience part, because the other five all lean on it.
There is already an audience paragraph in the source packet: a contractor earning between sixty and two hundred thousand a year, no bookkeeper, prompted to act by a penalty or a large year-end bill. That paragraph is fine for telling Ida who she is writing for. It is not an audience situation. It is a description of a person. A situation is a moment — something that just happened, that the person is currently inside, and that they would recognise if you described it back to them.
The evidence file has one situation carried by three independent lines. Lines one, two and five all describe the same shape: nothing happened for months, then a letter or a bill arrived, and the person started looking for help that week. Different sources, different authors, same trigger. That group was marked strong support for exactly that reason.
The file also has a thin one, line twelve on its own, where someone's sister-in-law was doing her own books neatly and that stung. That is a real line. Someone actually wrote it. It might even be a strong motivator for some buyers. But it is one source, and one source is a hypothesis.
Why does that matter enough to govern a whole message? Because a message aimed at a thin idea fails in a way you cannot read. If Ledgerlane builds its page around the embarrassment of falling behind a relative, and the page converts badly, you have learned almost nothing. Maybe the idea was wrong. Maybe it was right and the wording was bad. Maybe it was right for a tenth of the market and you needed the other nine. With one supporting line you cannot separate those. Aim at the three-line group instead and a bad result means something: the situation is real and documented, so the failure is more likely in the offer, the wording or the channel, and those are things you can change and retest.
The thin idea does not get thrown away. It gets a promotion path. Line twelve moves up to emerging support the moment a second independent source says something like it — a different forum, a different first call, a review. At three independent sources it can carry a message of its own. Writing that rule down is the point. Without it, thin ideas get promoted by whoever finds them charming, which on current evidence is usually the assistant.
So the brief's first element reads roughly like this, and I will give it in the words that go in the file: the audience is a one- or two-van trade contractor who has just received a penalty notice or an unexpectedly large tax bill, has no bookkeeper, and is looking for help in the days immediately after. Supported by lines one, two and five. Not the demographic paragraph. The moment.
Now the positioning claim, and here the word earns explaining. Positioning is not a slogan and it is not a summary of your good points. It is a statement of where your offer sits in the buyer's head relative to the other things they could do. Geoffrey Moore and April Dunford both build it from the same handful of pieces: the target situation, the category the buyer files you under, what you do differently, what they would use if you did not exist, and why they should believe you. The one that gets skipped is the fourth. Most drafts describe the product against nothing at all.
So before writing the claim, get honest about what Ledgerlane is actually up against. For a one- or two-van trade contractor in the United Kingdom, the realistic set is five things.
The first is the spreadsheet, or the receipts in the glovebox. It costs nothing. It falls apart on faded till receipts, on deductions nobody claimed, and on having no idea whether a job made money.
The second is an accountant who appears once a year. Contractors mostly see this as unavoidable compliance rather than help, and it breaks in a specific way: the accounts get prepared nine or ten months after the year ended, so the tax number arrives far too late to prepare for, sometimes with a rush fee attached.
The third is the spouse or a family member. Free and trusted. It breaks on the technical corners — the Construction Industry Scheme, the domestic reverse charge on VAT — and on the fact that home admin gets done last, plus the friction of being told off about paperwork by someone you live with.
The fourth is self-managed software: Xero, QuickBooks, FreeAgent, roughly ten to thirty-five pounds a month. Cheap and genuinely capable. It breaks because nobody operates it. Receipts get captured sometimes, bank feeds go unreconciled, subcontractor costs land in the wrong place. Rubbish in, rubbish out. Subscriptions without an operator get abandoned.
The fifth is doing nothing until something forces it. Ten-hour days on site, then a letter from His Majesty's Revenue and Customs.
Look at those five and notice which ones pair up. What actually holds the business is not a rival bookkeeping firm. It is the year-end-only accountant plus inertia. The accountant covers the legal requirement. The inertia covers everything in between. That combination is the incumbent, and it is what Ledgerlane has to be chosen over.
That changes how the claim is written, in a way worth being precise about. If you think your competitor is another bookkeeper, you write about being better: more responsive, better software, nicer people. If you know your competitor is a once-a-year accountant plus nine months of silence, you write about when, not how good. The gap is time. Ledgerlane's whole difference is that the numbers exist monthly instead of retrospectively.
So the positioning claim, one sentence, naming who it is for, what it does, and what it stands instead of: for one- or two-van trade contractors who have just been stung by a penalty or a surprise tax bill, Ledgerlane keeps the books current every month and the tax position visible all year, instead of a year-end-only accountant and a glovebox full of receipts.
That sentence is deliberately unglamorous, and it does something a slogan cannot. It tells every future draft what to argue against. And it is written against alternatives that are documented, not invented. Nothing in it claims a rival firm is slow, because the files say nothing about any rival firm. If a claim about a named competitor ever goes in, it needs its own evidence, and none exists yet.
One boundary belongs here, because it bites on positioning language specifically. In the United Kingdom the words accountant and bookkeeper are not protected, so Ledgerlane may call itself a bookkeeping service. The chartered titles are protected — chartered accountant, chartered certified accountant, and the membership designations of bodies like the Association of Accounting Technicians or the Institute of Certified Bookkeepers. Copy must not state or imply any of those without an active licence. Separately, anyone providing bookkeeping or tax services as a business is an accountancy service provider under the money-laundering regulations and must be registered for supervision with the Revenue or a recognised professional body before trading. That is a fact about the business, not a line of copy, but it belongs in the brief as a constraint so no future draft implies credentials the business does not hold. And under the advertising code, claims have to be accurate and capable of being backed up, which rules out something like a penalty-free guarantee unless you spell out exactly what gets reimbursed and when.
Next element: the benefit. This is where most briefs quietly turn back into a feature list. Monthly bank reconciliation. Quarterly VAT preparation. Receipt capture by photo. All true, all in the product facts, and none of it is a benefit. A benefit is the change in the customer's week.
The person the brief is aimed at is problem-aware. Eugene Schwartz's old distinction still does the work here: this buyer knows they have a problem and does not yet know that a solution like yours is a category they could buy. They are not comparing bookkeeping providers. They are annoyed and slightly frightened. So the benefit has to name the symptom they would name themselves, not the mechanism that fixes it.
Say it as relief from a specific recurring moment: no more Sunday evenings guessing what the tax bill is going to be, and no more digging faded fuel receipts off the van floor in January. That is aimed at the recognised symptom. And crucially, the wording comes from the lines. The evidence file is the vocabulary source — the phrases contractors actually used about January, about receipts, about not knowing. If a phrase in the benefit sentence does not appear in some form in a cited line, it came from marketing instinct, and marketing instinct is exactly what this brief exists to keep out.
Then scope. The product facts already list exclusions: no payroll, no invoice chasing, no incorporation advice. In the hero section those went in as a scope line, and the temptation is to treat them as small print you are obliged to carry.
They are doing more than that. Consider what happens without them. A contractor with two subbies on the Construction Industry Scheme reads a page about keeping trade books current, books a call, and spends twenty minutes discovering Ledgerlane does not run his payroll. He is annoyed and the slot is gone. Put the exclusion in the message and he either self-selects out before the call, or he books anyway knowing what he is buying. The exclusion line is not caution. It is a filter that raises the quality of what arrives. In the brief it goes in as its own element, with the product-facts lines behind it, and marked as text that must survive into every channel variant.
Then proof. Sort the candidates into three piles and only two of them get used.
The first pile is what the product facts establish. One van, one hundred and eighty pounds a month; two vans, two hundred and sixty. Onboarding about two weeks from signing. The specific cadence — receipts captured by photo, banks reconciled monthly, VAT prepared quarterly, year-end accounts tax-ready. For a problem-aware buyer, that cadence is proof in itself. It describes a machine that runs whether or not the contractor remembers, which is precisely the failure mode of the ten-pound-a-month app.
The second pile is what the evidence lines establish about the problem, not about Ledgerlane. This pile is easy to underrate. Naming the penalty timetable accurately is proof that you understand the customer's year. A Self Assessment return filed one day past the thirty-first of January draws an immediate hundred-pound penalty even if no tax is owed. Three months late and ten pounds a day starts accruing, up to nine hundred. At six months and again at twelve it becomes the greater of three hundred pounds or five percent of the tax. Late payment interest runs from the deadline at somewhere around seven and a half percent. Construction Industry Scheme returns are due monthly by the nineteenth, with a hundred pounds at one day late and two hundred at two months, which is brutal for someone with one or two subbies. Value Added Tax registration becomes mandatory once rolling twelve-month turnover reaches ninety thousand pounds. And from April two thousand twenty-six, Making Tax Digital for Income Tax Self Assessment starts requiring digital records and quarterly filing for sole traders over fifty thousand pounds, extending to over thirty thousand the year after — which is a real deadline arriving for exactly this audience's spreadsheet.
None of that is a claim about Ledgerlane. All of it is checkable, and getting it right signals competence more convincingly than an adjective.
The third pile is everything the offer cannot yet prove. Numbers of clients. Average tax saved. Testimonials. Time saved per month. Any suggestion that penalties will not happen. That pile stays out of the brief entirely — not parked, not softened, out. Write it down as a named exclusion list so that in three months, when someone is drafting an ad and wants a number, the brief says no and says why.
Last element: the call to action. Look again at where the trigger situation puts someone. A letter arrived on Tuesday. It is now Thursday. They are on a phone, on site, mildly panicking, and they are not shopping. Ask that person to book a paid consultation and they close the tab. High-commitment asks assume a buyer who has already decided the category is worth money, and this one has not.
What works is a low-friction diagnostic — something that gives them a partial answer to the thing that frightened them, without requiring a decision. A five-minute sole-trader tax checklist. An admin health check. A receipt audit. The ask matches their state: they want to know how bad it is. Give them that, and the conversation about a hundred and eighty pounds a month happens from a better place. In the brief, the call to action is written with a note explaining why it is low-commitment, because otherwise the first person who wants more leads will swap in a demo booking and never know what they changed.
Six elements. Now they become a document.
Getting the brief out of Ida and finding what is wrong with it
The request goes into the Ledgerlane Marketing project, where the product-facts sheet, the audience paragraph, the brand example and the labelled evidence file already live, and where the standing instructions already tell Ida to take factual claims only from uploaded files and to flag missing facts rather than invent them.
The request, in substance: draft a reusable audience-and-offer brief from the uploaded files, with six named sections — audience situation, positioning claim, benefit, scope boundaries, proof and call to action. After each element, cite the evidence line numbers or the product-facts entry that supports it. Where the files cannot support an element, write that it is unsupported and say what would be needed, instead of filling it. Do not cite a line for anything the line does not say. The competitive alternative set and the penalty timings are supplied in this request, not in the files, so mark those as supplied here rather than attributing them to the evidence file.
That last clause matters. Half the alternatives and all the penalty numbers came from research done outside the project. If they arrive without a source label, in a month nobody will know whether the hundred-pound figure came from a file or from Ida's guess.
The first draft came back well-formed, correctly sectioned, mostly cited. It also had three things wrong with it, and they are the three worth learning to spot.
The benefit sentence read, roughly: "Ledgerlane gives trade contractors back their evenings and the confidence that their finances are under control." Fluent. Warm. Cited to nothing. Search the evidence lines for "confidence" and it is not there. Search for "under control" and it is not there. That sentence was not drawn from anything; it was produced by the model's sense of how bookkeeping is normally advertised. An uncited element is the easiest failure to catch and the easiest to leave in, because it reads better than the sourced version.
The second problem was quieter. Under proof, a bullet asserted that contractors feel judged when their books are behind — sourced to line twelve. Line twelve is the sister-in-law line. One source, marked thin. It had been promoted from hypothesis to proof by being placed under a heading called proof. The citation was honest; the pile was wrong.
The third was a citation claiming more than its line supports. The audience section said contractors in this situation want monthly visibility into their tax position, citing lines one, two and five. Go back and read those three. They describe a penalty or a bill arriving and the person looking for help. Not one of them mentions wanting monthly visibility. That is a reasonable inference — it may well be right — but stated as audience fact with three line numbers after it, it borrows their authority. Six months later nobody rereads the lines. They read the citation and believe it.
So the inspection is three questions, run element by element. Does the cited line actually say what the element claims? Is any line being cited for more than it supports? Is anything stated with no citation at all?
One revision pass fixed all three. The benefit sentence was rewritten from the wording the lines themselves use, about January, about receipts, about not knowing the number, and cited to the lines carrying that wording. The sister-in-law point moved out of proof and into a short section at the end of the brief called untested hypotheses, with its support tier and the promotion rule attached: two independent sources to reach emerging, three to carry a message. And the audience line was split. The observed part — the penalty-or-bill trigger — kept lines one, two and five. The inferred part — that they want ongoing visibility — was restated as an inference, labelled as such, with those lines named as the basis rather than the proof.
That is the whole discipline, and it is the same one from the evidence file, applied one level up. A line supports what it says and nothing further.
Then the brief gets saved into the project workspace, alongside the hero section and its request, with the request that produced it stored next to it. That pairing is not bookkeeping for its own sake. Future work starts from the brief, not from the chat that made it. If the brief ever looks wrong, you want to be able to see what was asked for, not reconstruct it from memory of a conversation.
What that document now buys you is narrower and more useful than a message. It is a message with its reasons attached. You can hand it to a customer situation and see whether it survives contact: take one contractor who just got a penalty notice, read the positioning claim and the benefit back to him in his own week's terms, and watch whether he recognises himself or shrugs. If he shrugs, the brief tells you which element to suspect and which lines to go back and reread. That is the difference between a message you like and a message you can argue with.
Which leaves the approved hero section sitting in the project doing one job in one place, and a fast way to get more out of it.
Reuse it as a channel variant, and treat the adaptation as translation rather than rewriting. The rule is claim parity: change the format, the length and the pacing as much as the channel demands, and hold every factual claim, number, timeframe, qualifier and exclusion exactly as approved.
The destinations set the shape. An email opener wants one to three sentences, somewhere around thirty-five to sixty-five words, and the preview text truncates somewhere between thirty-five and ninety characters, so the core context has to land in the first line. A LinkedIn post allows three thousand characters, but the hook before the "See more" cut is what governs it: roughly a hundred and forty characters on mobile, a little over two hundred on desktop. A post on X gives non-subscribers two hundred and eighty characters for the whole thing — proposition, qualification and link.
The prompt to Ida is worth writing as a standing one. Give her the role of a compliance-focused marketing editor. Paste the approved hero copy as the source asset. Then four preservation rules: keep every factual claim, number, timeframe, scope boundary and qualifying clause intact while changing tone and format freely; do not inflate capabilities or turn a conditional into an unconditional promise; add no metrics, testimonials or features not in the source; obey the channel's limit. And ask for the output in two parts — first a claim breakdown listing every explicit claim, condition and boundary she found in the source, then the variant. The breakdown is the useful half. It shows you what she thinks she is preserving before she preserves it.
Then check the variant against the original, line for line, never against whether it reads better. Three things to compare. The core claim: two weeks from signing must still say from signing. The scope limitation: no payroll, no invoice chasing, no incorporation advice, still present. The qualifiers: about two weeks has not become two weeks flat, and the price has not lost the per-van condition.
And one rule for when it will not fit. If two hundred and eighty characters cannot hold a mandatory qualifier, do not trim the qualifier. Either link straight to the full terms, or swap that claim for a different one that is self-contained and complete on its own. A shortened claim is not a shorter claim. It is a different one.