There is no such thing as "thinset coverage"
Search for how far a bag of thinset goes and you will get a single number. It is not a real number. Hold the trowel and the bag weight constant and the published figures still diverge: CUSTOM VersaBond gives 90–100 sq ft per 50 lb bag at a 1/4″ × 1/4″ square-notch, while MAPEI Ultraflex 2 gives 75–90. That is a 20% gap at the conservative end, on the same job, for the same trowel.
Mortars differ in density, in how much water they take, and in how their ridges stand and then collapse under the tile. So the honest tool is one that carries each product's own chart, which is what this page does, and why the mortar selector is the first thing that changes the answer.
The notch is chosen for contact, not for coverage
It is tempting to read the coverage chart backwards — pick the notch that makes the bag go furthest. That inverts the actual requirement. ANSI A108.5, which both data sheets cite, requires at least 80% mortar contact behind the tile in dry interior areas and at least 95% in wet areas and exteriors, with all tile edges properly supported. MAPEI's instruction is to choose a trowel "with sufficient depth to achieve" those figures. Coverage is what falls out of that decision.
Which means larger tile, tile with deep ribbed backs, and ungauged natural stone all push you toward a bigger notch and use more mortar — and CUSTOM notes those cases may need flat-back or notched-back troweling on top. The way to check is the way the sheets describe: pull a tile up periodically and look at the back.
One trowel comes with a manufacturer warning attached. CUSTOM publishes coverage for the 1/2″ × 1/2″ square-notch and then recommends against using it, because the 1/2″ spacing between notches makes it harder to bed the tile and reach proper coverage. This calculator surfaces that note on screen when you select it. A coverage number handed over without the caveat that the manufacturer attached to it is worse than no number.
Back-buttering doubles it
The coverage chart on the bag assumes mortar goes on the substrate only. Back-buttering — a skim on the back of each tile as well — is routine on large format and stone, and it roughly doubles consumption. This is the usual reason a job runs dry halfway across a floor: the estimate was right for the method on the bag and wrong for the method being used.
The bag counts here come from the low end of each published range, then overage on top. Both manufacturers describe their charts as estimating figures that vary with jobsite conditions, and running short mid-floor with mortar already skinning over costs far more than an extra bag. The optimistic count is still shown in the working so you can see the spread.
What this does not cover
Only the two product families whose charts are cited below, at their published trowel sizes. It does not cover medium-bed or large-and-heavy-tile mortars at their thicker application, epoxy setting materials, self-levelling underlayment, or the mortar needed to fill an uneven substrate — an out-of-flat floor can absorb a surprising amount before any tile goes down. It also assumes a flat rectangular area with no waste from cuts, which is a tile question rather than a mortar one.
For the tile itself use the tile calculator, for a shower the shower tile calculator, and for the joints afterwards the grout calculator — or the per-brand MAPEI and LATICRETE versions, which work the same way this page does.