Vendor presentation deck: the anatomy of the slides that carry the argument

A vendor deck is the one document in a property file written for a room rather than a reader. It is presented, questioned and judged in forty minutes, and the judgement it receives is rarely the one its author is measuring: the buyers in the room are not deciding whether the deck is attractive. They are deciding whether the numbers on it can be traced, because everything else they can see for themselves when they visit.

That makes the deck a particular kind of document. It is an argument made of slides, and each slide either advances the argument or answers a question the buyer is about to ask. Anything that does neither is decoration, and decoration in this document is not neutral: it takes attention from the evidence, and it invites questions the file cannot answer, which is the opposite of what the seller needs in the room.

This piece is about the anatomy — which slides do the work, in what order, and on what evidence — and about the audit that separates them. It is deliberately not about templates. A template is a set of slides someone else needed; the anatomy is the set this transaction needs, and the two coincide only by accident.

What it is not, and will not become here, is a claim about outcomes. This publication has no measurement of instruction rates, of the share of decks that produce a buyer, or of what any particular deck achieved, and it will not invent one. The claim made here is narrower and it can be checked by anyone holding a deck: every sentence on it either rests on something in the file or it does not.

Anatomy, not template worship

The template failure is easy to describe. An office buys a deck design, fills fourteen slides because fourteen is the number the design has, and the filler is not random — it is whatever the last deck in the office contained. The result reads professionally and argues nothing, because the slides were chosen for the design and the evidence was fitted to them afterwards.

The anatomy runs the other way. Start with the argument the deck has to make, then list the questions a buyer will ask while that argument is being made, then build one slide per question and delete the rest. That procedure produces different decks for different assets, which is the point: a development site, a let office floor and a single apartment do not need the same eight slides, and no design should force them to.

There is a second reason to distrust the template, and it is the one that matters commercially. A slide that the file cannot support is a sentence the seller will be held to. A deck that says the market is moving, that the area is improving, that the asset will achieve a particular level of income, has made statements that the other side’s adviser will ask about, and a statement made in a presentation is not retracted by calling it marketing. The discipline of the anatomy is also the discipline of not making claims that have nowhere to stand.

The classes, and the test that sorts them

Every slide falls into one of three classes, and the test is a question, not a taste.

The first class carries the argument: what is offered, on what terms, why the price is where it is, what conditions attach. Remove one of these and the buyer cannot form a view, so each is load-bearing.

The second class carries evidence the buyer will check: the comparable set, the rate per unit of area, the tenure and its unexpired term, the condition record and the index of the file. These slides exist to be interrogated, which means they have to survive interrogation — a table that cannot be reproduced from the file is worse than no table, because it has invited the question.

The third class only decorates: the lifestyle collage, the aerial image that shows nothing, the projected-return slide, the market-growth chart, the leadership photograph. It is worth being exact about what these do, because the usual defence is that they are harmless. They are not. They consume the part of the presentation where attention is highest, they imply promises nobody has agreed to make, and they give the buyer a claim to test that the seller did not need to offer.

A fourth kind of slide appears in every deck and belongs to none of the three: the process slide, which answers a question rather than advancing an argument. It is in the table below anyway, because a room that likes the asset and does not know how to bid for it leaves without bidding, and that is a cost the first three classes cannot recover.

Table 1. The anatomy of a vendor deck: twelve slides in the order that makes an argument, the question each answers, the evidence it must rest on, and what its absence or its inflation produces. Source: constructed for this piece; the classes are the test, not a required sequence.

The order is not a rule, but each of the first ten constrains the next, and that is why it is worth observing. Slide 5 cannot be written before slide 3, because a guide placed in a range is an assertion about the set. The process slide cannot be written before the file index, because a process with dates implies documents that have to exist, and a process written over an empty index has been invented. Written in that order the deck becomes hard to inflate, which is the only property of a deck that can be tested before it is presented.

The argument the deck has to carry

Five slides carry the argument, and each is a sentence a buyer could repeat afterwards.

The offer states what is for sale at what guide, and it states both in the first minute. Decks that withhold the price until the process slide are making a negotiating decision and calling it a narrative decision; the effect on the room is twenty minutes spent guessing, and a buyer who guesses low does not revise upwards because a slide later in the file says otherwise.

The property describes the asset before it prices it: type, extent with the basis of measurement named, tenure with its unexpired term, condition with the basis on which it was recorded, current use and permitted use where they differ. Two failures recur here and both are cheap to avoid — an area with no basis, so the rate per square metre cannot be compared with anything, and a tenure stated as “leasehold” without the term, so a range built on long leaseholds is applied to an asset with a decade left.

The guide sits inside the evidence rather than beside it, and it is shown as arithmetic. Not “priced to sell” and not “offers in the region of”, but a figure, the adjusted range it sits in, and the position it holds in that range expressed as a percentage. A buyer can argue with that; a buyer cannot argue with an adjective.

The conditions slide is the one sellers’ offices most often drop, and it is the one that prevents the argument at the end. What is included in the price, what is excluded, what is subject to consent, what the seller will and will not do before completion, and what remains unknown in the file. A deck that states its own gaps is trusted on its other slides; a deck that shows none is assumed to be hiding one.

The closing slide, if there is one, is a process instruction rather than a summary: who to contact, what to send, by when. Everything after that is the buyer’s own internal meeting, and no slide in the deck changes it.

The evidence a buyer actually checks

Four slides will be checked, and the checks are specific. Knowing which they are is most of the preparation.

The comparable set is checked against the buyer’s own knowledge first. Every buyer in a specialist segment knows two or three transactions the vendor does not, and the question in the room is not whether the table is right but whether it is honest: excluded lines, dated sources and stated adjustments are what distinguish a set that has been argued over from one that has been assembled.

The rate per square metre is checked by division. If the deck states a guide of 1,848,000 for 420 square metres, the buyer’s calculator produces 4,400.00, and the next question is what the comparables’ rates were on the same basis. Where the areas were measured on different bases — gross external against net internal, or a marketing area against a measured one — the two rates are not comparable and the division the buyer performs is not the division the deck performed. Stating the basis on the slide prevents a conversation the vendor cannot win.

Tenure is checked by a professional reader and it takes about ninety seconds: term, unexpired term, review basis, service charge mechanism, restrictions on use and on assignment. This is the slide most often reduced to one line, and the one where a single missing phrase causes a price to be repriced.

The index of the file is checked last and it is the slide that produces confidence rather than information. A buyer who can see that the survey, the planning file, the title and the abstract exist, and who is told where each one is, stops wondering what is being withheld. Its absence produces the request that stalls transactions: “can you send everything you have”, three days after the room.

The slides that only decorate, and what they cost

The decorative slide is not a crime against design; it is a cost that is easy to see once it is counted. Three kinds recur.

The lifestyle collage shows cafés, parks and a gym within the postcode. The buyer will visit and see them. Its cost is not the slide but the space it takes where the comparable set might have gone, and the signal it sends: a deck that spends three slides on the neighbourhood has perhaps not got three slides of evidence.

The projected return is the expensive one. A slide showing an income projection, a rental estimate or a resale expectation is a statement about the future placed in a document the seller issued, and it will be asked about — by the buyer’s adviser, by a lender, and, if the transaction goes wrong, by somebody reconstructing what was represented. The projection is not merely unsupported; it is a commitment dressed as an illustration, and no seller needs to make it in order to sell a property.

The market-growth chart is the same problem with a more respectable appearance. A chart of prices in a region, without a source, a date and a definition of the index, is decoration with an argument’s clothing: it says the buyer should expect a trend the file has not established, and it invites the question of why the asset needs a trend to be sold.

There is also a class of slide that is neither argument nor decoration but an unforced error: the one that shows something the vendor has not verified. A photograph of a view that belongs to a neighbouring building, an area quoted from a marketing brochure rather than a measurement, a rent copied from a listing. Each is checkable, each takes one question to expose, and each is unnecessary — the deck needed none of them.

The audit, worked

The audit is a table, and it takes about half an hour per deck. Every slide gets classified, every claim on it gets counted, and each claim is marked as resting on a source in the file, as arithmetic on such a source, or as having no source at all. The counts are the finding.

The example below is a constructed fourteen-slide deck for a constructed asset: 420 square metres, a guide of 1,848,000, a comparable set of five transactions and one asking price. No address, party or market is described, and the figures are arranged so the arithmetic can be followed rather than so any asset can be described.

Table 2. A claim audit of a constructed fourteen-slide vendor deck. Each claim is counted once: sourced, computed from a sourced figure, or unsourced. Amounts in constructed currency units. Source: constructed for this piece.

The counts are the argument. Thirty claims on fourteen slides, of which twenty rest on a document in the file, two are arithmetic on those documents, and eight stand on nothing. The eight are not evenly distributed: with one exception, they sit on the slides whose purpose was persuasion rather than description, which is why most of them are removed rather than repaired. The unsourced share is 26.7 per cent of every claim the deck makes.

Six slides go — 1, 5, 10, 11, 13 and 14 — carrying seven of the thirty claims; the address on the cover is a duplicate of the one on slide 3 and goes with it. Two of the three claims on slide 2 go as well, because neither rests on anything the file holds. The remaining unsourced claim, the one on slide 4, is repaired rather than cut: the distance it asserts is replaced with a figure from the file’s own record, which is the only place in the audit where an unsourced claim could be saved at all. What remains is eight slides — four carrying the argument, three carrying evidence, one carrying the process — and twenty-one claims, of which nineteen rest on a document and two are arithmetic on it. Nine of the thirty claims are gone, 30 per cent of the deck’s assertions, and six of fourteen slides, 42.9 per cent of its pages. The share of slides doing argument or evidence work rises from 50 per cent, seven of fourteen, to 87.5 per cent, seven of eight, because the decorative 43 per cent is simply absent.

The arithmetic in the evidence slides is the part that has to add up in front of the room. The five adjusted rates are 4,646.00, 4,343.00, 4,743.00, 4,343.00 and 4,173.00 a square metre, and the asking price cross-check is 4,284.00. On 420 square metres the range is 1,752,660 at 4,173.00 to 1,992,060 at 4,743.00 — a width of 239,400. The median rate of 4,343.00 produces 1,824,060, and the plain mean of 4,449.60 produces 1,868,832. The guide of 1,848,000 is 4,400.00 a square metre; it sits 39.8 per cent of the way up the range, 23,940 above the median figure (1.3 per cent) and 20,832 below the mean figure (1.1 per cent). Those four numbers are what slide 5 ought to say, in that order, and every one of them can be reproduced from the table on slide 3.

What the audit does not establish

The first limitation is the important one: the audit tests traceability, not persuasion. It can tell an office that every claim on a deck rests on something in the file, and it cannot tell anybody that the deck will produce a buyer, an offer at the guide, or a faster process. Those are outcomes of the market, the asset and the room, and this publication has no measurements of any of them, so it makes no claim about them.

The second is that a deck cannot repair what is underneath it. Eight clean slides built on an unexamined comparable set are exactly as confident as eight clean slides built on a sound one, and the audit passes both. The comparable table is checked before the deck is written, which is the order the family of property documents runs in for this reason.

The third is that the anatomy is a convention, not a standard. A deck written for a tender, a joint-venture partner or an institutional buyer is a different document with a different first slide, and the classification holds while the sequence does not. An office that adopted this order as a template would have replaced one template with another, which is the failure the piece opened by naming.

The fourth is arithmetic honesty about the counts themselves. The thirty claims and the eight unsourced ones are one constructed deck’s, counted by one method, and the method has a judgement in it: a sentence like “the area is well served by transport” was counted as one claim, while “close to the station, the motorway and the airport” was counted as three. A different counter would produce a different total on the same slides. What would not change is the ranking of the slides, and it is the ranking, not the total, that the audit is for.

The decision at the end

The decision is whether the deck leaves the office with every claim traced, and it is taken slide by slide rather than in a single judgement.

For each slide, three questions in order. Does it carry part of the argument, or answer a question the buyer will ask? If neither, it goes, whatever it cost to design. If it does, does every claim on it rest on something in the file — a document, a measurement, a dated source — or on arithmetic over those? If not, the claim is cut or the slide is marked as a statement of the seller’s position rather than of fact. And does the slide say what it does not claim — the basis of the area, the basis of the condition, the jurisdiction of the process, the date of the evidence?

A deck that passes those three tests is short, unglamorous, and receives a different kind of attention in the room: buyers stop testing the deck and start testing the asset, which is the only thing the seller wants them to test.

Where this goes

• The pillar: x402.review/2026/10/31/property/ — the six documents and the boundary this piece works inside
• The role: studio.x402.press/roles/real_estate_assistant — the agent that assembles the deck and the table under it
• The service it buys: x402.press/services/deepseek/ — the canonical price
• The review: x402.review/2026/10/31/deepseek-review/ — the measured handshake
• The evidence a deck rests on: comparable-evidence-table — the eight fields that decide whether a table is usable
• The range the guide sits in: market-assessment-structure — the document that turns the set into a range

The cost of a deck

A deck is ten paid model calls: one to read the comparable file and the subject notes, five to draft the argument slides, one to draft the evidence slides from the table, one to write the conditions and the process, one to check every claim against the file, and one to render the file. At the measured DeepSeek rate on Base — $0.00700 a short-prompt call under exact — that is $0.07000. At the metered ceiling of $0.00955 the same ten calls read $0.09550.

The audit pass is two more calls, one to classify the slides and one to reconcile the claims, which puts the deck at twelve calls and $0.08400. Against one complete deal’s fifty calls at the measured rate — $0.35000 — the deck is 20 per cent of the total, or 24 per cent with the audit included.

The figure that circulates as DeepSeek’s price, $0.00875, is Qwen’s measured rate and not DeepSeek’s; at that rate the same fifty calls would read $0.43750, and the deck’s ten would read $0.08750 instead of $0.07000. The endpoint in front of the call is the price, which is why every figure here was read from the handshake rather than from a catalogue.

Method note

The price was read from the live 402 handshake at the DeepSeek endpoint on Base on 31 October 2026: $0.00700 for a short prompt under exact, $0.00955 for a long prompt under exact and as the upto ceiling at max_tokens: 1000. Every figure in Table 2 sums: 20 sourced plus 2 computed plus 8 unsourced is the 30 claims the fourteen slides carry, and the classes sum to fourteen slides — four argument, three evidence, one process and six decorative. The rate arithmetic divides and reconciles: 1,752,660 ÷ 420 is 4,173.00 and 1,992,060 ÷ 420 is 4,743.00, so the width of 239,400 is 420 × 570.00; 1,848,000 ÷ 420 is 4,400.00; and 95,340 ÷ 239,400 is 39.8 per cent, where 95,340 is 1,848,000 less 1,752,660. The deck, the asset, the comparable set and the counts are constructed so the arithmetic can be followed; no address, party, transaction or market is described, and none is real. Nothing in this piece is a forecast, a valuation or advice about whether to transact, and no claim is made about what any deck achieved: the agent assembles the evidence and the slides, and a qualified person prices, advises and signs.

x402.review is written and edited by N.P. Vincent. The publication is connected to x402.press, which sells the services reviewed here; that interest is disclosed on every piece that touches a service, and it is stated here rather than left to inference.

Filed under Jobs · Tags: property, presentation, deck, vendor