The Word After the Money: Sixteen Prices in a Hundred Say What They Buy

Eighty prices that are only a number

Every listing in the VIT MARKET catalogue publishes its price twice as text, once in rubles and once in dollars, on top of the two numeric fields a machine reads. Take the dollar text and apply a single test: is the entire string two dollar amounts joined by a dash, and nothing else? Eighty of the hundred pass. They are a number, and only a number. The other twenty carry words as well as digits, and those words are the only place in the whole catalogue where a price states what it is the price of.

That ratio is the finding. Four fifths of this shelf publish a sum without a unit. The title says what the service is and the description says what the client ends up with, but nothing in those eighty rows says whether the money buys one delivery, a month of them, a machine that produces them, or a single item out of a quantity nobody has agreed yet. The twenty annotated rows are not the anomaly needing explanation. They are the ones doing what a price is supposed to do.

The noun after the money, and the three things it can be

Sixteen of the twenty name what this seller's own money buys. Six cite an outside market instead — an Upwork range for the same kind of audit, a Fiverr median for chatbots, a market band for link audits — and two of those six do both at once. One listing explains why the count is sixteen and not seventeen: a first-line answering bot whose dollar label does contain the phrase per project, attached not to its own range but to the Fiverr median printed beside it. Read at speed, that label looks like a scope statement. It is a scope statement about somebody else's price.

Sort the sixteen by what the noun actually refers to and there turn out to be exactly three kinds. Seven name a phase of work: a setup, a pipeline, a diagnosis. Six name a unit: an email, a page, a video, a marketplace card, a copy of a template. Four name the job as a whole, as a project or a package. One listing appears in two of those groups because its label does two things at once, which is why seven plus six plus four comes to sixteen. Three kinds of noun, three completely different shapes of invoice, and no field anywhere records which kind a listing is.

A rate wearing a price tag

The six unit-priced rows are the easiest to misread, because a range like $60 to $250 looks exactly like a project price. Email templates that survive Outlook are $60 to $250 per email. Rewriting a page for search intent is $50 to $200 per page. YouTube optimisation is $15 to $50 per video. Marketplace card optimisation is $30 to $80 per card. A theme sold on a marketplace is $19 to $59 per copy. Batch generation of product cards is $400 to $1,500 for the setup and then ten to fifty cents per card.

For all six, the published figure is a coefficient and the bill is that coefficient multiplied by a quantity the catalogue never sees. Fifty pages at the floor of the page rate is $2,500, which is more than seventy-six of the hundred listings ask at their ceiling. The buyer supplies the multiplier, usually late, often after the price has already anchored the conversation. What a unit price gives you, and a project total never can, is a clean basis for comparison — provided both sides agree what a page is.

Seven prices that buy the machine, not the output

Phase-priced rows deserve a slower read, because their noun quietly moves the finish line. Local search and map listings are $200 to $800 for a one-off setup. Rank monitoring does not lead with money at all: its dollar label reads Setup $150–400, the only price in the catalogue that puts the scope word first. Article production is $1,000 to $3,000 per pipeline. Ad-creative generation is $400 to $1,500 per pipeline. Lead-email personalisation is $600 to $1,800 per pipeline. Translation and localisation is $500 to $1,800 for the setup. Product-card generation is $400 to $1,500 for the setup.

A pipeline is a machine. Buying one is not the same purchase as running one, and in six of these seven the running cost is either stated somewhere else or not stated in dollars at all. This is the cleanest test in the catalogue for telling a fixed scope from an open one: if the word after the money is a stage rather than a deliverable, the published band is a floor for the engagement and not its price, and everything after it is a second negotiation that has not happened yet.

The emptiest word on the shelf is package

Four labels name the job as a whole, and they are not equally informative. A traffic-drop investigation at $500 to $2,500 per diagnosis is precise, because its delivery estimate says in the same breath that the remedial work is a separate project. A technical SEO audit at $150 to $900 per project and a product-description rewrite at $300 to $1,200 per project are ordinary. Bulk generation of page titles and descriptions at $150 to $700 per package is the outlier: a package of what? The dollar label does not say. The ruble label does, and prices a hundred-page package separately from a thousand-page one.

Nor does the service page narrow it. Each listing renders a short list of what is included, and that list is generated rather than written: four lines, of which three are identical for all 100 services and the fourth is identical for all 25 services on the same shelf. Four distinct included-lists exist for a hundred distinct jobs. The first line of the shared three is the most honest sentence in the whole system — it promises a concise brief and an agreed scope before work begins, which is the catalogue saying plainly that the scope is settled somewhere other than on this page.

The ruble label knows more than the dollar label

Run the same bare-or-annotated test down the ruble column and the count changes sharply. Thirty-six of the hundred ruble labels carry text beyond the range, against twenty in dollars. Nineteen listings are annotated in both currencies. Seventeen are annotated only in rubles. Exactly one goes the other way — a backlink audit whose dollar label carries a market comparison the ruble label omits. So on seventeen listings the two price lines of one service disagree about how much a reader is told, always in the same direction.

The seventeen are not decorative. Five of them are bare ranges in dollars and read for setup in rubles: a lead generator at $150 to $500, a backup routine at $80 to $250, an uptime monitor at $60 to $200, invoice digitisation at $350 to $1,000, and an industry-news digest at $50 to $200. The lead generator goes further in rubles and adds three to fifteen rubles per validated contact at volume — a per-unit meter with no dollar equivalent anywhere on the listing. A speed and Core Web Vitals job priced at a flat $200 to $800 in dollars splits in rubles into diagnosis with implementation against audit only without fixes, at roughly half the money.

The mechanism is visible in the repository. A per-service translation table carries five fields for each of the 100 listings in English and Spanish: title, category, description, delivery estimate and difficulty. Neither price field is in it. So the qualifier on a delivery estimate travels between languages — the email listing reads three to eight hours per letter in English — while the qualifier on a price never does. By shelf, the gap is starkest on automation: two of its 25 dollar labels carry scope words against thirteen of its 25 ruble labels. On web development the two columns agree, at two each.

Eight second meters, and only one of them speaks dollars

Count the labels that name a charge arriving after the first one and the asymmetry becomes concrete. Eight ruble labels do it: a subscription after the monitoring setup, three to fifteen rubles per validated contact, eight to forty rubles per generated product card, twelve to fifty rubles per rewritten description as an alternative to the project price, two thousand to eight thousand rubles per article after the pipeline is built, one hundred and fifty to eight hundred per ad creative, fifteen to sixty per personalised email, and eighty to three hundred per thousand characters of localised text.

One dollar label does it. Product-card generation publishes its ten-to-fifty-cents-per-card rate in both currencies; the other seven publish the second meter only in rubles. For a buyer this is the most actionable line in the analysis, and it needs no Russian: wherever a price mentions a setup or a pipeline, ask what the per-unit charge is afterwards — on seven of these eight listings the catalogue has already answered in a currency you may not be reading.

What a price label cannot tell you

None of this measures what anyone is actually billed. A label is a label: the eighty bare ranges may be fixed-scope quotes their sellers would confirm in the first reply, and the twenty annotated ones may be renegotiated on contact like any other. This analysis classifies published text and nothing else. It also cannot tell you whether an annotation is complete — a row that says per pipeline has told you the price is for a phase, but not what the next phase costs, unless the ruble line happens to add it.

One boundary is worth stating explicitly. Some of these labels also disagree with the numeric price fields stored beside them, a defect VJOURNAL has audited separately, and none of that arithmetic is repeated here. The test in this piece is about words, not figures: whether the price names its own object. A row can be numerically immaculate and still tell a buyer nothing about what the sum covers, which is the more common failure and the harder one to notice.

Five listings, one price, five different bills

The catalogue supplies its own closing example. Five listings publish exactly $400 to $1,500. One of them charges that for a setup and then ten to fifty cents per product card. One charges it per pipeline for ad creatives. One is a custom automation module whose label adds only that consultants elsewhere charge $1,000 to $3,000 for comparable API work, with the per-module scope stated in rubles. Two — multilingual SEO with hreflang, and payment acceptance with webhooks — say nothing beyond the range. Same five figures, five different agreements.

So read the noun before you read the number. If it is a unit, ask how many units and multiply before you react to the price. If it is a stage — setup, pipeline, diagnosis — ask what the next stage costs and what the per-item rate is afterwards. If it is a package, ask what is in the package, because in this catalogue that word is defined in only one of the two currencies. And if there is no noun at all, which is the case eighty times out of a hundred, the question is not whether the price is high. It is what the price is the price of.

Decision framework: fixed price versus time and materials

Fixed price transfers only the uncertainty that the scope actually defines. Time and materials exposes effort but still needs priorities, a budget ceiling and acceptance criteria; neither model repairs an ambiguous brief.

FAR performance-work guidance emphasises required results and measurable standards rather than hours alone. The decisive text after the price is what will exist, how acceptance works and what is excluded.

Underline every promised deliverable, acceptance test, revision limit, dependency and exclusion; if the commercial model changes but these lines do not, the risk allocation probably has not changed either.