Three Times the Floor: The Quiet Convention Inside Almost Every Price Range

The number nobody negotiates

A price range is two numbers, and buyers argue about one of them. The floor is what you hope to pay; the ceiling is what you brace for. Almost nobody looks at the distance between them — which is a shame, because in the VIT MARKET catalogue, the services arm of the VITON13 platform, that distance turns out to be the most consistent thing on the page. All 100 listings carry a floor and a ceiling.

The intuitive reading of a wide range is that the seller does not yet know what the job involves: a narrow range means defined scope, a wide one means the gap is an honest measure of what remains unknown. That reading is testable, because a hundred bands are a hundred chances for the width to vary with something real.

A disclosure before the arithmetic. VIT MARKET is VITON13's own storefront and VJOURNAL is VITON13's newsroom, which is precisely why nothing here is asserted from memory. Every figure below was computed from the catalogue source file. Band width throughout means ceiling divided by floor — a listing quoted at $500 to $1,500 has a band of 3 — and where a sentence is our reading rather than the data's, it says so.

Almost every band is three times its floor

One listing has to leave the sample first. A template sold on ThemeForest stores $19 and $10,000 as its two numbers, but its own label explains that $19 to $59 is the price of one copy and $600 to $10,000 is the lifetime total. Those are two different quantities, not the ends of a band, and dividing one by the other gives a meaningless 526-fold ratio. It is excluded here, and named so the exclusion can be checked.

That leaves 99 bands. Their median width is 3.33 times the floor. The lower quartile is exactly 3.00 and the upper quartile is 3.75, so half the catalogue sits inside a band-width range of three-quarters of a multiple. Counted outright: 73 of the 99 bands fall between three and four times inclusive. Fifteen are narrower than three, eleven are wider than four.

The clustering is sharper still at the round figures. Twenty listings quote a band of exactly 3.00 — $300 to $900, $500 to $1,500, $600 to $1,800, $1,000 to $3,000. Thirteen quote exactly 4.00. Between them, a third of the catalogue lands on one of two whole numbers.

That is not what unsettled scope looks like. A hundred genuinely different jobs, each with its own uncertainty about what a client will ask for, would not produce a third of the list agreeing on a ratio to two decimal places. The width is mostly a convention.

Difficulty moves the level, not the width

If the width reported uncertainty, harder work should carry wider bands. Every listing already declares its own complexity tier, so the test is direct. Sorted by tier, the median band width reads 3.33 for the entry tier, 3.25 for the middle tier and 3.42 for the top tier. There is no trend, and the whole spread across the three is smaller than the gap between the catalogue's own quartiles.

This finding needs a methodological guard, because the same question asked differently gives the opposite answer. Compare median floors across the three tiers — $150, $300, $500, a 3.3-fold climb — against median ceilings across those tiers — $400, $1,050, $2,500, a 6.25-fold climb — and bands appear to widen sharply with difficulty. They do not. That comparison takes the floors of one set of listings and the ceilings of a different set, and the ratio between two separately computed medians is not the median of the ratios.

Measured the only way that answers the question — inside each listing, one floor against its own ceiling — the width does not move with difficulty at all. What difficulty moves is the level. Our reading: the catalogue prices harder work as more money at the same relative uncertainty, which is the behaviour of a rate card rather than of an estimator.

Where the band genuinely widens

The width does vary, but not by complexity. It varies by discipline, and one group separates cleanly from the rest.

The four service groups, by median band width: AI for business at 3.00, website development at 3.06, automation at 3.33, and SEO and traffic work at 4.00. The AI group's quartiles are 2.92 and 3.33, the tightest in the catalogue. SEO's quartiles are 3.50 and 4.67 — its narrow quarter is wider than the AI group's wide quarter.

The widest bands in the catalogue are almost all SEO. Placement in ChatGPT, Perplexity and AI Overviews is quoted at $300 to $2,500, a band of 8.33. A technical SEO audit runs $150 to $900, a band of 6. Programmatic landing pages, a traffic-drop investigation and e-commerce SEO give three bands of exactly 5. Only one non-SEO listing reaches 5: an urgent repair of broken layout at $50 to $250 — a different phenomenon, since an emergency job cannot know what it will find.

Our reading of this, offered as reading: SEO is the one group whose result is delivered by a third party. A developer controls whether the site is built; nobody controls whether a search engine ranks it. The catalogue's widest bands are exactly where the outcome depends on a system the seller cannot instruct, and the extra width looks like the honest part of the price list.

Two shelves that never pass four times

The other end of that observation is sharper than the medians make it sound. In the automation group, no listing anywhere has a band wider than 4.00. In the AI group, the same: 4.00 is the ceiling, reached but never exceeded. Fifty listings between them, and not one quotes a range wider than four times its floor.

Website development goes as high as 5, SEO reaches 8.33. So two of the four shelves impose a hard cap on their own uncertainty, and the two that break it are the emergency-repair end of development and the search-dependent end of SEO.

For a buyer this beats the median. A range narrower than about four times its floor is unremarkable here — it says nothing about how well defined your project is, because nearly everything is quoted that way. A range wider than four times is genuinely unusual and worth a question: in every case here it marks work whose result is not fully in the seller's hands.

Round numbers at both ends

There is a second sign that these endpoints are chosen rather than derived. Of the 100 listings, 61 have a floor that is a whole multiple of $100 and 88 have a ceiling that is. Fifty-nine have both.

The exceptions are informative. Thirty-nine floors are not whole multiples of $100, and of those only ten distinct values escape multiples of $50 altogether: $15, $19, $30, $60, $80, $120, $156, $180, $341 and $399. Three of them — $156, $341 and $399 — are not prices the catalogue set at all. They leaked in from market-benchmark notes, which is the subject of the next section.

An estimator working from measured variance would not land on whole hundreds nine times out of ten at the top of the band. Prices ending in two zeros are prices a person wrote down. That is not a criticism — almost every price list is built this way — but it settles what a band is. It is a quote shaped by convention, not a computed interval.

Five rows where the stored band contradicts the printed one

Every listing carries its numeric floor and ceiling in machine-readable fields, and separately carries a human-readable price label. Those should agree. In five of the 100 listings they do not, and reading all five together shows a single mechanism at work.

The GEO/AEO listing prints "$300 – $2,500" and adds a market note: basic packages on Fiverr run $156 to $291. Its stored floor is $156 — a competitor's price, not its own. The link-profile audit prints "$400 – $1,400" with a market range of $399 to $1,399 in brackets, and stores $399. The first-line AI autoresponder prints "$350 – $1,200" with a Fiverr median of $341 to $520 in brackets, and stores $341. In all three the stored floor was captured from a benchmark clause rather than from the seller's own quote.

The fourth is the ThemeForest row already set aside. The fifth fails differently: the weekly position-monitoring service prints "Setup $150–400" and stores $150 as both its floor and its ceiling, losing the top of its band entirely and becoming, in the raw fields, the only single-price listing in the catalogue.

None of these is a large error, and none changes what a customer is charged, since the label is what a customer reads. They matter here for one reason: they sit at the extremes.

The correction that changes nothing

The honest way to handle five bad rows is to recompute everything on the published labels and publish both results. Reading the band printed at the head of each label instead of the stored fields, across the same 99 listings: median width 3.33, lower quartile 3.00, upper quartile 3.75, and 73 of 99 between three and four times. Twenty are exactly 3.00 and thirteen exactly 4.00.

Every one of those figures is identical to the uncorrected version. The five defective rows sit entirely in the tails, and correcting them moves only the tails: the widest band drops from 16.03 to 8.33 and the narrowest rises from 1.00 to 2.67, which becomes a four-way tie.

That is the strongest thing in this article, and it is worth being plain about why. A pattern that survives the correction of its own worst data is a pattern that does not depend on that data. The three-to-four-times convention is not an artefact of five bad fields. It is what the catalogue does.

What a range cannot tell you, and what to ask instead

It cannot tell you where inside the band you will land. The catalogue records asking prices, never invoices — no record of what was billed, what was discounted, or whether buyers cluster near the floor, the ceiling or the middle. A midpoint is arithmetic, not a forecast.

It cannot tell you whether the convention is grounded. A band of 3.33 might encode real measured variance from past projects, or it might be a number somebody settled on years ago and applied ever since. Those two possibilities are indistinguishable from the outside, and the catalogue stores no history that would separate them.

And the group comparison rests on 25 listings per group, one seller, one moment. It describes the shape of one price list. It is not a market benchmark, and no figure here should be quoted as one.

What survives is a practical instruction. Do not read a three-times band as a statement about your project; nearly every listing here has one. Read a band wider than four times as a signal worth a question, because in this catalogue it consistently marks work whose outcome depends on somebody other than the seller. And when a quote arrives with an unexpected range, ask what would have to be true for it to land at the floor. If the answer names something you control, the range is negotiable. If it names a search engine, it is not.

Decision framework: what a price range in a quote means

A wide price range is a statement about unresolved information, not automatically a sign of overpricing. Its value depends on whether the supplier identifies the assumptions that move the estimate from the lower to the upper bound.

Official estimating guidance expects ranges to reflect scope maturity, schedule and risk and to narrow as evidence improves. An unexplained three-times band is not yet a decision-grade estimate.

Ask for the three largest range drivers, the evidence needed to resolve each, the owner of that evidence and the date when the supplier can issue a narrower baseline.