The Cleanest Numbers in the Catalogue, and the Rate Still Triples

The only shelf where the arithmetic needs no assumptions

Almost every rate comparison in the VIT MARKET catalogue rests on a borrowed constant. Sixty-two of its 100 listings quote delivery in days or minutes, and turning those into hours needs a declared eight-hour day that nobody measured. One shelf is exempt. All 25 web-development listings quote hours directly, and 23 of the 25 state a bare quantity with nothing after it — "8-14 hours", "35-60 hours", "3-5 hours". Nothing is converted, nothing is truncated, nothing is assumed.

That makes web development the one place in this catalogue where implied hourly rates are directly comparable: 25 prices and 25 durations counted in the same unit, by the same seller, for the same kind of buyer. It is also the shelf where the catalogue's outlier test finds nothing at all — not one of the 25 sits at twice or half the shelf median. And yet the rate a buyer is implicitly paying runs from 24.21 dollars an hour to 68.75, a factor of 2.84 from one end to the other.

Half the shelf lives inside a factor of 1.44

Sort the 25 implied rates and the shape is a tight middle with visible ends. The median is 47.06 dollars an hour. The lower quartile is 37.14 and the upper quartile 53.33, so the middle half of the shelf occupies a band only 1.44 times wide. To a buyer with a calculator, most of these listings would look priced at essentially the same hourly figure.

The shelf also prices coherently in aggregate. Add all 25 price midpoints and the total is 33,864.50 dollars; add all 25 duration midpoints and the total is 725 hours; divide one by the other and the blended rate is 46.71 dollars an hour, within 35 cents of the shelf's own median. The house rate is real, and it is about 47 dollars.

The ends are what a buyer actually meets. The dearest implied hour on the shelf sits roughly 46 per cent above that median and the cheapest roughly 49 per cent below it. Two listings that both look like ordinary hour-quoted web work, published by the same seller on the same page, differ by almost a factor of three in what an hour of them costs — and nothing on either page says so.

The two ends, named

Both are reproduced here with the catalogue's own English title, its stored price band and its published estimate. Both quote hours, and neither carries any clause after its number.

Project size explains none of it

The obvious hypothesis is volume: long jobs get a lower rate, short jobs carry a premium for the interruption. The two ends above fit it neatly. The shelf as a whole does not. Rank the 25 listings by duration, rank them again by implied rate, and the correlation between the two orderings is 0.02 — indistinguishable from none.

The clearest illustration sits at the bottom of the price list. Sort the shelf by price midpoint and take the four cheapest projects: an urgent layout repair at 150 dollars, a set of Outlook-proof email templates at 155, a PWA wrapper at 250 and the 24-hour company site at 275. Rank those same four by implied hourly rate against the other 21 and they come out third, twenty-third, twenty-fourth and first. Four projects within 125 dollars of each other occupy both ends of the rate table.

Grouped by size, the same flatness appears. The nine listings quoted at 15 hours or less have a median implied rate of 47.06 dollars; the ten between 15 and 35 hours, 45.75; the six above 35 hours, 48.22. Across those three buckets the median project price climbs from 325 dollars to 1,000 to 3,625 — an elevenfold rise — while the rate moves by less than three dollars and does not move in one direction. On this shelf a bigger budget buys more hours at the same price, not cheaper hours.

The lowest rung charges the highest rate

Split the same 25 listings by the difficulty word each one carries. The seven Beginner listings have a median implied rate of 52.78 dollars an hour. The twelve marked as needing experience: 43.47. The six marked expensive and difficult: 48.22. The lowest median rate on the shelf belongs not to the bottom rung but to the middle one, and the top rung is not the dearest either.

The prices behind those rates go the other way, and they go a long way. Median project midpoint by rung: 325 dollars for Beginner, 1,000 for Experienced, 3,625 for Advanced. Median hours: 9, 22.5 and 70. The ladder is real in money and real in time. In rate per hour it is close to flat, with the lowest rung slightly on top.

A plain reading is available, and it is worth stating as an argument rather than dressing it as a finding. A short job carries fixed costs — a brief, a call, a handover — that do not shrink with the work, and spreading those over four hours produces a higher hourly figure than spreading them over seventy. The catalogue records no fixed cost anywhere and cannot confirm that reading; it can only show that the pattern is consistent with it.

The other three shelves, compared like for like

The fair comparison is bare estimate against bare estimate, because a converted day or a truncated clause widens a spread for reasons that have nothing to do with pricing. Restricted to listings whose estimate is a bare quantity, the four shelves spread like this: web development, 23 listings, 2.84 times from the cheapest implied hour to the dearest; automation, 9 listings, 2.52; SEO and content, 9 listings, 2.40; AI implementation, 20 listings, 1.87.

Two things follow. The shelf with the cleanest arithmetic carries the widest honest spread, which is the reverse of what an outlier hunt would suggest, since every flagged row in the catalogue lives somewhere else. And AI implementation, with 20 clean estimates and a spread under two, is the shelf whose rate is most nearly fixed — which is what pricing from a formula looks like when the units are clear enough to see it.

What a clean rate cannot tell you

It cannot tell you who does the work. The catalogue records no seniority, no team size and no split between design, build and testing, so two listings at the same rate may represent very different labour, and one at half the rate may simply be reusing something that already exists.

It cannot separate efficiency from padding. A low implied rate is arithmetically identical whether a seller quotes generous hours at a fair price or realistic hours at a low one, and nothing in the record distinguishes the two. The same ambiguity runs in the other direction at the top of the shelf, where a high rate may mean scarce skill or simply a short job.

And it cannot tell you what anything actually cost. These are published estimates on a menu, not logged hours on a finished job: no invoice, no timesheet and no revision history exists anywhere in this data. Every figure in this article describes what the price list claims, which is a different object from what a project came to.

What to do with a spread of 2.84

The practical move is to build the comparison the price list does not print. Take the two quotes you are weighing, divide each price midpoint by its own duration midpoint, and look at how far apart the two implied rates land. Inside this shelf that ratio tells a buyer more than any single price does: it is the difference between paying about 47 dollars for an hour of the same seller's attention and paying 69.

Then ask the question the number opens rather than the one it appears to close. A rate 46 per cent above the house median is not evidence of overcharging on a four-hour job; it is a signal that fixed costs are being carried by a small number of hours, and the sensible response is to ask what those costs are and whether a second job could share them. That is a conversation a price list cannot have, and it begins with arithmetic a price list makes possible.

Decision framework: effective hourly cost of web development

An implied hourly cost is a diagnostic ratio, not a market rate. It becomes meaningful only when the hours represent comparable work and the numerator includes the same discovery, management, review and support obligations.

A cleaner-looking listing can still imply a higher ratio because its duration omits waiting or because its price includes reusable expertise. The ratio identifies questions; it does not answer value.

Calculate separate ratios for active production, total supplier effort and calendar duration, then annotate the included outputs and risk ownership beside each number.