The six-input severity model behind the tool, the six-rung escalation ladder it routes between, the gentlest-first principle that prevents over-treatment, the substrate-cap rules that protect coated and irreplaceable glass, and the difference between a deposit that will release under the right chemistry and an etched surface that no chemistry can recover.
What the Hard-Water Stain Removal Triage does, in five points:
The tool's job is to put the right chemistry on the panel the first time. The trade default — reach for the strongest cleaner that's lying around and scrub until the deposit moves — produces both unnecessary substrate damage and unnecessary chemistry cost. The escalation ladder substitutes a calibrated decision for the reach-for-the-strongest reflex.
There is a conversation that happens every time a homeowner walks past their shower door, or the south-facing kitchen window above the sink, or the storm door on the back patio, and notices that the glass no longer looks clean even when it is clean. The conversation runs through a sequence of escalating attempts: a wipe with regular cleaner, then a wipe with vinegar, then a wipe with CLR, then a wipe with something stronger from the hardware store, then a scrub with a Magic Eraser, then the slow realization that whatever is on the glass is not going to come off without a different kind of work. By the time the homeowner calls a professional, half of the wrong-rung treatments have already happened, and the operator who arrives needs to figure out where on the escalation ladder to start without redoing what has already been done and without skipping past what hasn't.
The Hard-Water Stain Removal Triage is the tool we built for that conversation. It takes six inputs — deposit type, age, glass substrate, water source, prior treatments attempted, and time available — and selects the right rung of a six-rung removal-escalation ladder. The output is not a single chemistry recommendation; it is a complete protocol for the selected rung (chemistry, dwell, rinse, post-treatment) with adjustments for what to do if the rung fails to move the deposit. This piece is the methodology behind the tool: what the rungs are, how the severity score routes between them, why the substrate cap exists, why the gentlest-first principle is strict, and where the framework ends and the replacement conversation begins.
The escalation ladder runs from gentlest to most aggressive. Each rung is defined by a specific chemistry family and a specific application protocol.
Rung 1 — RINSE. The gentlest first step. Plain water or pure water, applied with a brief contact-and-wipe, then flood-rinsed. This rung exists not because rinse alone removes mineral deposit (it doesn't, in most cases) but because the rinse-first move surfaces what is loose-versus-bonded on the panel. Light film and surface-soiling often release under rinse alone. If the deposit moves at this rung, no further escalation is needed and no aggressive chemistry has been applied to a panel that didn't need it.
Rung 2 — SURFACTANT-SOAK. Warm water plus surfactant — dish soap, the house cleaning mix, any standard window-cleaning solution — applied to the panel and allowed to dwell for five to fifteen minutes before being wiped and rinsed. The surfactant lifts surface bonding and the dwell time gives the calcium-carbonate forms in light spotting and recent deposit a chance to release. Most film-grade deposits move at this rung; spotted-ring patterns sometimes do, depending on age.
Rung 3 — MILD-ACID. White vinegar at concentration (not diluted), or CLR mixed per label, or commercial lime-and-scale remover. The acid neutralizes calcium-carbonate by protonating the carbonate to carbonic acid which then releases as carbon dioxide. This is the trade-default first-line treatment for visible hard-water spotting; the bubbling at the deposit surface during application is the calcium-carbonate releasing, and the operator who sees the bubbling knows the chemistry is working. Three to ten minutes of contact, then extensive rinse to clear the acid residue.
Rung 4 — STRONG-ACID. Oxalic acid (the active ingredient in Bar Keeper's Friend) for iron-bearing deposits; phosphoric acid blends for calcium-silica composites; specialty stronger formulations for cooling-tower-source deposits. This rung is the catastrophic-failure rung for coated and tinted-film substrates — the chemistry attacks the coating layer before it touches the deposit. The dwell time at this rung is short and watched continuously; two to five minutes maximum, with rinse triggered the moment the deposit shows visible movement. Over-dwell etches the underlying glass.
Rung 5 — MECHANICAL. Cerium-oxide polishing slurry on a felt pad or fine polishing wheel. This rung is for deposits that have resisted all chemistry but where the underlying glass surface is still intact. The mechanical action removes a thin layer of glass along with the deposit; on annealed glass the result is usually an acceptable surface, on tempered glass the result is usually a fabricating-debris scratch event (the polishing wheel is the steel scraper in this scratch mode). The tool routes to this rung sparingly and with a hand-off to the Scratch Risk Calculator for tempered substrates.
Rung 6 — REPLACE. The terminal rung. The deposit has resisted every removal attempt, or the substrate has been etched, or the substrate is irreplaceable and aggressive treatments are off the table. The recommendation is panel replacement rather than continued removal attempts; the math usually favors replacement after the second or third unsuccessful escalation.
The severity score is the additive sum of point contributions from six inputs. The score is the primary input to rung selection, but it is not the only input — the gentlest-first principle and the substrate cap both modify the score-based recommendation.
Deposit type carries the largest single-input contribution (0 to 50 points). Light film contributes 8 points; spotted-ring contributes 22; heavy crust contributes 38; etched contributes 50 and forces the replace rung regardless of other inputs. The spread reflects what the trade observes: film releases easily, spotted-ring is the standard mild-acid case, heavy crust requires strong-acid or mechanical, and etched is past the framework.
Deposit age contributes 0 to 18 points. The age premium reflects the chemical-bond strength of aged deposits — calcium-carbonate crystallizes over months, silica cross-links, iron-oxide sets — and the operator's expectations should adjust accordingly. A six-month-old deposit at the same visible thickness as a one-month-old deposit will resist treatment more strongly because the underlying chemistry has matured.
Glass substrate contributes 0 to 12 points, but more importantly it sets the substrate cap that limits maximum recommendable rung. Annealed uncoated has no cap (full ladder available); tempered uncoated caps at mechanical (with Scratch Risk Calculator hand-off); low-E coated, tinted film, and heritage all cap at mild-acid. The points contribution reflects the additional treatment complexity even within the allowed range — tinted film at mild-acid requires more careful technique than annealed at mild-acid.
Water source contributes 5 to 20 points and routes to source-specific advice. Municipal hard at 5 points is the trade-default baseline (calcium-carbonate dominant, predictable response); well-iron at 15 points triggers the two-stage treatment protocol (calcium first, then iron); sprinkler at 12 triggers the source-control advice (adjust the head before the next cycle); cooling-tower at 20 triggers the source-assessment advice (the cleaning alone will not resolve the recurrence).
Prior treatments attempted contributes 0 to 22 points and is the key input to the gentlest-first principle. None contributes 0 (start at rinse); soap-and-rinse attempted contributes 4 (start at surfactant-soak); vinegar attempted contributes 8 (start at commercial mild-acid); CLR attempted contributes 14 (start at strong-acid if substrate allows); mechanical tried contributes 22 (the terminal-rung signal — only replacement remains).
Time available contributes -2 to 6 points. The negative value on unrestricted-time reflects an honest empirical observation: long dwell times let the gentlest effective chemistry do its work, and the optimal-time case actually scores lower than the time-pressured case because the operator can use a gentler treatment. The time-pressure input flags when the recommendation is being forced upward by the available window rather than by the deposit itself.
The score-based rung selection is the first half of the recommendation pipeline. The second half is the gentlest-first principle: even when the score indicates a higher rung, the recommendation steps down to the lowest rung the operator has not yet attempted.
The principle is strict by design. When the prior-treatments input reads none, the recommendation always steps down to the rinse rung, regardless of how high the severity score has gone. The reasoning: an operator who has not yet rinsed the panel may be looking at a deposit that releases under rinse alone, and the tool cannot tell the difference between "this is a heavy crust" and "this looks like a heavy crust but is actually a layer of dust over a light film" without the first treatment attempt to clarify.
The principle steps up the ladder as prior treatments are recorded. If the operator updates the input from "none" to "soap-rinse" after rinsing and observing no change, the recommendation moves to surfactant-soak (the next untried rung). If that also fails and the input moves to "soap-rinse" remains, the operator is in surfactant-soak mode; if they document a surfactant-soak attempt by moving to "vinegar" (the wrong label for what they tried but the closest in the tool's vocabulary — this is a known limitation), the recommendation advances to mild-acid.
The principle prevents the over-treatment failure mode the trade default produces. The default reach — apply the strongest cleaner that's available and scrub until the deposit moves — burns chemistry and damages substrate on panels that would have released under a fifteen-minute soak. The gentlest-first principle is the calibrated alternative.
Substrate compatibility cuts across the score-based recommendation by capping the maximum allowable rung. The cap is not about gentleness; it is about preventing permanent damage to substrates that cannot tolerate higher rungs.
Tinted film caps at mild-acid because anything above that — strong acid, mechanical work, even some commercial mild-acid formulations — attacks the film polymer, the bond layer, or both. The damage is visible immediately (cloudy patches, lifting at the edges, color shift) and is not recoverable; the film must be removed and replaced. The cap forces a different conversation: if the deposit will not release under mild-acid on a tinted-film panel, the answer is professional film replacement, not aggressive removal.
Heritage glass caps at mild-acid for the irreplaceability reason. Pre-1940 cylinder glass cannot be replaced with modern float without losing the optical character that gives the panel its value; any aggressive chemistry or mechanical work that damages the surface is permanent damage to an irreplaceable object. The honest recommendation for a heritage panel with a deposit that mild-acid won't release is to leave the deposit in place rather than risk damaging the panel further.
Low-E coated glass caps at mild-acid because the coating layer is the most chemistry-sensitive surface in modern residential construction. Most low-E coating manufacturers spec mild-acid only (vinegar, dilute CLR) and explicitly contraindicate the strong-acid family. The damage to a low-E coating from oxalic acid is visible within seconds — the coating softens, then flakes, then leaves a bare-glass spot in the application area. The replacement cost of a low-E IGU runs $400 to $800; the cap exists to prevent that.
Tempered uncoated caps at mechanical (full ladder is available) but with a scratch-risk hand-off — any time the recommendation lands on mechanical for a tempered panel, the tool surfaces the Scratch Risk Calculator as the next required check before proceeding. The fabricating-debris failure mode is the catastrophic outcome for mechanical work on tempered glass, and the dedicated tool exists for that exact decision.
Annealed uncoated has no cap. Full ladder access; the only constraint is the score-based recommendation and the gentlest-first principle.
When the deposit-type input reads etched, the entire framework collapses to the replace rung regardless of every other input. The reasoning is that what appears as "stuck deposit" on an etched panel is not actually deposit — it is glass-surface roughening produced by the alkaline mineral chemistry attacking the silica matrix. The deposit was the cause; the etching is the consequence; the etching is not removable by any chemistry because there is no longer a clean glass surface underneath the apparent deposit.
The two visual tests for etching, both surfaced in the tool's deposit-type input descriptions: the deposit will not lift even partially under aggressive mild-acid, and the underlying glass feels rough to the touch when the surface is wiped clean. Both signals confirm that the silica matrix has been modified rather than a discrete deposit having built on top of an intact surface.
The cerium-oxide mechanical rung can sometimes recover the visual appearance of etched annealed glass by removing the etched layer entirely — but this is specialty restoration work, requires the equipment and skill to do without distorting the optical character of the panel, and is rarely cost-competitive with replacement. On coated or heritage substrates, mechanical recovery is not available, and replacement is the only path.
Two of the water-source inputs trigger source-control advice that goes beyond the immediate cleaning recommendation.
Sprinkler overspray is the most common avoidable source of recurring hard-water staining on residential glass. The repair is mechanical — adjust the sprinkler head to direct away from the panel, or install a deflector — and is part of the work the operator should do or recommend. Without the source repair, the cleaning is recurring overhead with no resolution path; the homeowner pays for the same work on the same panel every six months.
Cooling-tower drift is the commercial parallel and is harder to address. The drift-eliminator condition on the tower, the distance from the tower to the affected glass, the prevailing-wind direction during the heating season — all of these are environmental factors the operator may not control but should document. The customer conversation needs to acknowledge that successful removal does not prevent recurrence; the tower assessment is a separate work-package the property management should commission.
Well-iron water triggers the two-stage treatment advice: calcium first (mild-acid), then iron (oxalic) as a follow-on. Running oxalic on a deposit that still has its calcium layer in place locks the iron stain under the calcium-carbonate matrix, which makes subsequent removal harder. The correct sequence — mild-acid until the calcium clears, rinse extensively, then oxalic to address the underlying iron — is non-obvious and the tool surfaces it explicitly.
The tool will not diagnose the deposit chemistry from a photograph. The deposit-type and water-source inputs are the operator's best read of what is on the panel and where it came from; the tool trusts those reads and routes accordingly. A misclassified deposit (heavy crust read as spotted-ring, well-iron source read as municipal-hard) will produce a misrouted recommendation, and the misrouting will surface at the treatment attempt rather than in the tool. The first treatment is the test of the diagnosis; if the deposit does not behave as the recommended rung expects, the diagnosis was wrong and the input should be updated before re-running.
The tool will not save the operator from skipping the gentlest-first step. The prior-treatments input is honest — it records what has actually been tried — and the recommendation depends on its accuracy. An operator who claims none-attempted to get a low-rung recommendation, then proceeds to apply mild-acid anyway because they "know" the deposit needs it, is working outside the framework. The framework only protects the substrate when the operator works inside it.
The tool will not replace the operator's eye on the panel during treatment. The dwell times in the per-rung protocols are guidelines, not commitments. A mild-acid treatment that has visibly cleared the deposit at three minutes should be rinsed at three minutes, not held for the maximum ten. A strong-acid treatment that has not visibly moved the deposit at four minutes should not be held longer in the hope that more time helps; the rinse is the safer move and a second cycle at the same dwell is the right next step. The tool's protocols are starting points; the panel in front of the operator is the actual guide.
Mara Whitfield is Senior Editor at Window Washing Guide and covers the chemistry-and-damage editorial beat, with particular focus on substrate-specific cleaning chemistry. Editorial content is researched and reviewed in collaboration with the Giordano Inc. editorial team, with field input from practicing operators in well-water and hard-water markets, plus published trade and materials-science references on mineral-deposit removal.