What an IICRC Card Is Worth, Measured

Data collected October 24, 2026. Every dollar figure below is in the dataset linked at the end of this article, or is arithmetic on that file. The rejects log — every lead that did not survive a page read, and why — is linked there too.

The design this collection was judged against was written first, before any page was opened. Every dollar below is in the dataset or is arithmetic on that file. Annual bands are turned into hourly at 2,080 hours (40 × 52) only in the model. The dataset keeps the unit the page printed.

The question nobody had answered

An earlier research pass looked for an authoritative dollar value on an IICRC certification and did not find one. No study. No table. One proposed pay model, Ken Larsen’s, written up in Restoration & Remediation in April 2022. The measurement here is the other way: read real postings, price the credential the way labor economics prices a credential, and publish the method with the number.

The number, from this pass, is a wide interval around zero. That is a result. It is not a slogan.

The method

Two borrowed tools, named on purpose.

Jacob Mincer’s earnings equation puts the wage in logs and holds schooling and experience. Sherwin Rosen’s hedonic price reads a wage as the sum of the attributes bundled into the job. The certificate is one attribute. The posting also buys a role, a city, and a tenure screen. The certificate’s price is whatever gap is left after those are held still.

The pay figure is the midpoint of the band on the page. A single figure is a band with the same low and high. Preferred is its own status. It is not folded into required, and it is not treated as silence.

Two comparisons were specified before any page was opened.

1. Inside a role × metro cell, the median midpoint of postings that require the named cert, against the median of postings that do not mention it. A cell with fewer than 4 postings on either side does not produce a result. It produces an inventory line.
2. Across the file: log of the hourly midpoint on required-cert indicators, role, metro, a coarse experience band, and franchise status. A cert with fewer than 8 required postings does not get a price. The coefficient can still be shown, with the N, so a reader can see that the interval is too wide to use.

“Associated with” is the verb wherever the card cannot be separated from the rest of the posting. One clean dollar per cert was the failure mode this design was written to avoid.

The dataset

48 rows. 47 URLs. One Spyder Construction page in Littleton stated two bands, technician and lead, and is two rows. Five of these rows were added after a browser pass over Indeed, and only after this run fetched a page that itself stated the pay. The other names from that pass had no posting URL, or the page did not return a current band. They are in the rejects log.

Read on 2026-10-24, except four pages that printed their own posted date: Right-Now Restoration of Seattle (2026-08-25), Kustom US in San Diego (2026-03-11), A1 Restoration of San Diego (2026-07-15), A1 Restoration of Dallas (2026-03-20). Those four were still up when read. A March posting read in October can be stale. It is in the file because the page still stated the pay.

Roles: 22 water/mitigation technicians, 18 lead technicians or crew chiefs, 8 project managers, 0 mold remediation technicians. Mold work is inside the water and lead ads. It never arrived as its own paid posting.

Employers: 34 franchise, 14 independent, 0 carrier-program. Franchise pages (CareerPlug for SERVPRO, Paul Davis, ServiceMaster, PuroClean, Restoration 1, Rytech, AdvantaClean, Voda) were the pages that actually opened. Indeed returned HTTP 401. ZipRecruiter returned HTTP 403.

The target was 120. The file stops at 48 because the rest of the leads did not survive a page read. The rejects are in the linked rejects log. The sample was not padded.

Posted hourly midpoints, before any cert split: technicians median $22.50 (N=22, range $16.50–$29.50), leads median $27.00 (N=18, range $22.50–$33.65), project managers median $34.50 an hour (N=8, range $25.00–$42.50). The project-manager median includes two hourly postings: SERVPRO of Morton Grove / East Niles at $22–$28, and Greenwood Heating in Everett at $40–$45. The other six PM bands are annual. Their median midpoint is $71,750.

Per-cert findings

Status counts across all 48 rows:

Four more rows require an IICRC card without making one named cert mandatory. Restoration 1 of North Fort Worth, MVP Restoration, and SERVPRO of Morton Grove / East Niles name no cert. All Pro in Denver requires at least one of WRT, OCT, CCT, or FSRT. Eleven rows prefer an IICRC card and do not name it. One, a SERVPRO of Aurora Central technician posting, tells the hire to go get the card after starting. Those unnamed cards are not assigned to WRT.

The regression below is log(hourly midpoint) on a required-cert indicator for each of the four named certs, plus lead, project manager, metro, experience (none / some / two-years-or-more), and franchise. Reference row: technician, Seattle–Tacoma, experience not required, independent. N=48. R²=0.80. The fit is high because role and metro explain the wage. The cert intervals are the part that does not.

WRT

WRT is the only named cert that reaches the pre-registered floor of 8 required postings. The adjusted association is about −3.9%, and the 95% interval runs from −18.4% to +13.1%. That interval is the headline range. It covers a raise, a cut, and zero. It is not a price. Adding the five verified pages from the browser pass moved the point estimate down from the first-pass figure, because the new high bands do not require WRT. The interval still covers zero.

The eight required rows are not eight clean WRT screens.

• Paul Davis of Tacoma, lead, $28–$32, 2 years, WRT only (row 6).
• Paul Davis of Greater Phoenix, lead, $26–$32, 2 years, WRT required, ASD and FSRT preferred (row 10).
• SERVPRO of Buckeye, crew supervisor, $24–$26, 2 years, WRT only (row 15).
• Bill Howe, San Diego, lead, $25–$30, 5 years, WRT only (row 17).
• ServiceMaster by Weibye, Boise, lead, $20–$25, 3 years, WRT and FSRT. The sentence says required, and it sits under the heading “Preferred Qualifications” (row 42).
• A1 Restoration of Seattle, technician, $27–$31, prior experience, WRT + FSRT + AMRT together (row 2).
• Spyder Construction, Littleton, lead, $24–$30, 2 years supervising, WRT + FSRT + AMRT together (row 26).
• ServiceMaster of Seattle, project manager, $72,000, 2 years, page spells them WTR, ASD, and FRST (row 7).

Every ASD, AMRT, or FSRT requirement in this file sits inside a WRT requirement. No posting asked for the advanced card by itself. The separate coefficients later in this draft are splits of that stack. They are not four prices.

Inside a single role and metro, the WRT contrasts that exist are one posting against one posting. All of them fail the size rule.

• Denver leads: Spyder required, midpoint $27.00, against three leads that do not mention WRT (All Dry $27.00, SERVPRO of Aurora Central $24.00, BMS CAT $24.00). The not-mentioned median is $24.00. Delta $3.00 an hour, about 12%. Two of those three prefer an unnamed IICRC card. N=1 and 3. Still under 4 on the required side.
• San Diego leads: Bill Howe required, midpoint $27.50, against Paul Davis of North County, WRT not named, midpoint $24.00. Delta $3.50 an hour, about 15%. Bill Howe also asks for 5 years. N=1 and 1.
• Seattle–Tacoma leads: Paul Davis of Tacoma required, midpoint $30.00, against Paul Davis of Greater Seattle, which pays for the class instead of requiring the card, midpoint $27.50. Delta $2.50 an hour, about 9%. The Tacoma ad also asks for 2 years in restoration. The Lynnwood ad asks for 1 year in a trade. N=1 and 1.

Held inside the lead title and pooled across metros, the six WRT-required leads have a median midpoint of $27.25. The seven leads that do not mention WRT have a median midpoint of $25.75. Delta $1.50 an hour, about 5.8%. Several of those seven prefer an unnamed IICRC card. Set those aside and the three fully silent leads — Paul Davis Lynnwood at $27.50, All Dry Denver at $27.00, PuroClean Woodstock at $26.00 — have a median of $27.00. Delta $0.25 an hour. Same job family, card required or not, the posted midpoint barely moves.

The tempting technician gap is one row. The only technician posting that requires WRT is A1 of Seattle at a $29.00 midpoint, and it also requires FSRT and AMRT. The median technician posting that does not mention WRT is $21.00 (N=12). Delta $8.00 an hour, about 38%. That is the ad, not the cert. The adjusted model, which can see that this row is a stack in a high-wage metro, does not turn it into a WRT price.

Fourteen other postings prefer WRT and will hire without it, or will pay for the class. Preference is common. A hard WRT screen is not.

ASD

ASD is required on one posting: the ServiceMaster of Seattle project manager, stacked with WRT and FSRT (the page’s FRST). N=1. The fitted coefficient is about +8.5%, and the interval runs from −28% to +64%. That interval is what a single stacked row does to a regression. It is not a range anyone should quote as the value of ASD. Twelve postings prefer ASD. Thirty-five do not mention it. No role × metro cell has an ASD contrast with both groups large enough to use.

AMRT

AMRT is required on two postings, both stacks: A1 of Seattle (with WRT and FSRT) and the Spyder lead (with WRT and FSRT). N=2. The fitted coefficient is about +40%, interval −4% to +105%. Too thin to conclude. The Denver lead contrast is Spyder at $27.00 against four leads that do not mention AMRT, median $25.50, delta $1.50 an hour, about 6%. N=1 on the required side. Six postings prefer AMRT. Forty do not mention it.

FSRT

FSRT is required on four postings: the two stacks above, the ServiceMaster of Seattle project manager, and the Boise ServiceMaster lead whose “required” sentence sits under a preferred heading. N=4, under the floor of 8. The fitted coefficient is about −23%, interval −46% to +10%. The sign flips relative to AMRT because the model is pulling one stack apart. That is collinearity, not a finding that FSRT pays less. Inside Seattle–Tacoma project managers the only contrast is ServiceMaster at a $34.62 midpoint, which requires FSRT, against Greenwood Heating at $42.50, which does not mention FSRT. Delta about −$7.88 an hour. N=1 and 1. Greenwood is a department supervisor at a home-services company, not the same job as the ServiceMaster posting, and the cell is too small to use. Nine postings prefer FSRT. One, All Pro in Denver, will accept FSRT as one of several cards. Thirty-four do not mention it.

The same All Pro posting is the only OCT appearance that treats the card as a way in the door: at least one of WRT, OCT, CCT, or FSRT, preferably two, pay $21–$26. PuroClean of North Seattle prefers OCT and does not require it. HST is absent from all 48 pages.

A coarser indicator — any of the four named certs required — has the same eight rows, because the advanced cards never appear without WRT. Its adjusted association is about −5.1%, interval −16.2% to +7.4%. Drop the six template-cluster rows (four A1-boilerplate ads, two near-duplicate “you are IICRC certified” ads in Fort Worth) and the association on the remaining 42 is about −10.3%, interval −22.6% to +3.9%. Both intervals cover zero.

What the regression can see is the job, not the card. Project manager, relative to technician, is about +51% (interval about +34% to +71%) in the full model. That coefficient has eight postings behind it, including an hourly band at $22–$28 and a department-supervisor band at $40–$45, so it is a sample fact, not a national PM premium. It is large enough to show why a cert study that ignores title will call a promotion a certificate.

The employer side

This is eligibility, not a percent.

On Accuserve’s own contractor page (https://www.accuserve.com/contractor, read 2026-10-24), water-mitigation membership requires an IICRC WRT technician on site for every visit. ASD is preferred and should be held by someone in the company. Mold remediation requires AMRT, plus whatever the state or locality requires. The same page also requires insurance limits, background checks, a weighted extractor, and two kinds of moisture meter. The card is one gate in a list. The page does not attach a wage to it.

A second Accuserve page, the re-inspection protocol posted May 9, 2024, asks the inspector on a disputed water loss for WRT and ASD, and for AMRT when the issue is mold. That is a credential for a specific assignment, stricter on ASD than the join-page preference. Both are on Accuserve’s site. They are not the same rule, and neither is a pay scale.

Contractor Connection’s requirements URLs, fetched the same day, returned a homeowner matching page. The certification matrix was not on the page. Alacrity’s join URL returned HTTP 404. Sedgwick, Lionsbridge, and BrightServ were not opened on a program-owner page that named a cert. Trade blogs that say every program requires WRT were not used. One verified program gate is the finding. A percentage of programs is not.

The two anchors

Larsen’s model, as printed in Lorne McIntyre’s April 20, 2022 Restoration & Remediation article “Building a Sustainable Restoration Workforce” (https://www.randrmagonline.com/articles/90069-building-a-sustainable-restoration-workforce), adds a percent of the project manager’s base, at a tenure stage, for each card earned.

• Technician assistant, months 3–9, is at 37% of PM base. WRT adds 4 points of PM base. FSRT adds 4.
• Technician apprentice, months 9–18, is at 45% of PM base. ASD adds 5.
• Restoration technician, months 18–36, is at 55% of PM base. AMRT adds 5. OCT adds 3. HST adds 3.

Four points of the PM’s base is not a 4% raise on the technician’s check. At the assistant stage, 4 / 37 is about 11% of that stage’s pay. ASD’s 5 points on a 45% stage is about 11% of that stage’s pay. AMRT’s 5 points on a 55% stage is about 9% of that stage’s pay. The article is explicit that this is one restorer’s recommendation, tied to a career path, not a measured market wage.

Applied to the one single-figure PM base in this file, ServiceMaster of Seattle at $72,000, Larsen’s WRT step is 4% × $72,000 = $2,880 a year, or $1.38 an hour at 2,080 hours. The ASD or AMRT step is 5% × $72,000 = $3,600 a year, or $1.73 an hour. Applied to the median of the six annual PM midpoints in this file ($71,750), the same steps are $2,870 and $3,587.50 a year. Those are illustrations of Larsen’s rule on observed PM bases. They are not posting premia.

Agreement: several franchise ads in this file say the band moves with certifications (Paul Davis of Tacoma, SERVPRO of Buckeye, PuroClean of Boise, ServiceMaster of Seattle’s spelling variants). They do not print the step. Larsen’s direction — the card is a reason to move someone up a ladder — matches that language.

Disagreement, stated as far as the postings can carry it: no page read here publishes a 4% or 5% step, and the adjusted association of a WRT requirement with the posted midpoint is an interval from −18% to +13% that includes zero. The lead-title comparison, the closest thing to holding the job still, is $0.25 to $1.50 an hour depending on how strictly “not mentioned” is drawn. That is smaller than an 11% step on a $27 lead midpoint (11% of $27 is about $3 an hour), and the sample is too small to rule Larsen’s step out. The postings do not confirm it. They also describe a different object: a hiring screen on a band, not a raise after someone already on the crew earns the card.

Franchise ladder: HOLD. No corporate pay ladder was read on SERVPRO, Paul Davis, ServiceMaster, PuroClean, or 911 Restoration’s own site. A WayUp copy of a 911 Restoration schedule in Appleton, Wisconsin, was outside the sample and was not the employer’s page. It is not an anchor.

What this study cannot say

The card is bundled with the rest of the ask. The eight WRT-required ads are leads, a project manager, or a technician who must also hold FSRT and AMRT. Experience rides along: 2 years, 3 years, 5 years. A regression with 48 rows and a stack of dummies does not unbundle that. The interval around zero is the honest summary of the unbundling this file can support.

Silence is not zero. “Not mentioned” means the posting did not use the named cert as a screen. The hired person may already hold it. Preferred, which is where WRT actually lives in this file (11 ads), is a third state and was kept out of the contrast on purpose.

A posting is not a paycheck. Bands are recruiting statements. Rytech of Seattle prints $20–$37. SERVPRO of Oldsmar / Westchase prints $50,000–$85,000. The midpoint treats those as one wage. All Pro’s $55,000–$65,000 and DRYCOR’s $62,000 are bonus illustrations on the page, and they were not used as the band. Overtime, on-call bonuses, and production pay are outside the midpoint.

Metro price levels differ. The model includes metro indicators for that reason. A national dollar gap would have mixed Seattle with Tampa. The technician median in this file is $22.50, and one Atlanta technician ad prints $16–$17 with a blank in the body and $16.00–$17.00 in the footer. That row is kept because it is what the page printed. It pulls Atlanta down. It is one row.

Pay-transparency laws are not the whole selection story. Washington, California, Colorado, and Illinois account for 27 of 48 rows. Phoenix, with no statewide mandate in view here, still contributed 7, because franchise career pages stated pay anyway. Spokane, inside a disclosure state, contributed 1. The binding limit was which pages opened, not only which states require a number.

Four A1-family ads share a boilerplate paragraph and are flagged a1-boilerplate in the CSV. Two Fort Worth ads share the sentence “You are IICRC certified” and are flagged you-are-iicrc-certified. They are real pages with different pay. They are not six independent wage policies. The sensitivity that drops them still covers zero.

ServiceMaster of Seattle spells the cards WTR and FRST. ServiceMaster Phoenix spells the water card WTR. Those were coded as WRT and FSRT, and the page spelling is in the notes. Paul Davis of NW Chicago lists ORT among preferred extras. ORT was not coded as OCT. Restoration 1 of Chicago Southland lists MRST. MRST was not coded as AMRT. Paul Davis of Greater Seattle offers paid training in “IICRC FRST” and was not coded as requiring FSRT.

No cell in this file can support the sentence “the cert is worth this much.” The sentence the file supports: a named IICRC requirement, as these employers wrote it in October 2026, is associated with a posted-wage gap that this sample cannot distinguish from zero, once role, metro, and a coarse experience control are in the equation. The place the card shows up cleanly is on the employer’s side of the market, as a gate Accuserve publishes for who may take the work.

The full table

R = required on the page. P = preferred, including “or willingness to obtain.” 1of = accepted as one of a list, not required by itself. A dash means the named cert is not mentioned. Unnamed IICRC language is in the CSV column generic_iicrc and is not in this grid. Experience bands: none, y1, y2, y3, y5, or stated without a year count. URLs and notes are in the CSV. Row 25 and row 26 are one Spyder URL.

Rows 9, 14, 16, 19, 21, 22, 25, 28, 30, 31, 33, 36, 38, 45, and 47 carry unnamed IICRC language (preferred, required, or “seek it after hire”) that this grid does not show. Read generic_iicrc and notes in the CSV before treating a dash as “this employer does not care about a card.”

Friction log and debrief

Blocked sources. Indeed (HTTP 401), ZipRecruiter (HTTP 403), LinkedIn (not opened). Those three were supposed to carry the sample. Franchise career pages and a handful of Paylocity, Rippling, and JobTarget pages were what actually returned a job body with a dollar figure. Career.com timed out on the one CRC Seattle URL that looked usable. Glassdoor was outside the source list and was not used, even where an index showed an employer-provided band.

Shell pages. A recurring CareerPlug failure: the URL resolved, and the body was the franchise welcome page, with no job and no pay. Search indexes of several of those URLs contained a pay figure. Those figures stayed out. The rejects log names each URL. This is the same failure mode as an earlier snippet pass, caught at the fetch.

Thin cells. Every within-cell cert contrast is 1 against 1. Mold is empty in all 10 metros. Spokane has a project manager only. Boise has two leads and nothing else. Dallas has no lead. Denver, Phoenix, and San Diego have no project manager with a readable band in the opened set. Atlanta’s technician cell is one ad at $16–$17. No cell clears 4 on each side of a cert contrast. The pre-registered rule was to say so and not borrow a gap from another city. That rule bound.

What a second pass fixes. A browser session that can pass Indeed and ZipRecruiter, or a data pull the boards license, is the difference between 48 and 120. The Seattle-through-Denver browser pass still could not supply Indeed URLs for 19 of the 20 ads it named. Those ads stayed out. The Chicago-through-Boise browser pass stopped before it returned a single posting, so those five metros were not added to. Pay-transparency states will still over-contribute if the pull is public postings, and that should stay as a reported bias rather than a weight invented after the fact. The second pass should also open the current Contractor Connection, Alacrity, and Sedgwick contractor manuals on the program owner’s domain. Those three were holds, not findings. A franchise corporate compensation guide, if one exists on the franchisor’s own site, is the missing ladder next to Larsen. Calling ten employers and asking what they added to the check when a tech passed WRT would measure the object Larsen wrote down. This file measures the object the postings contain, which is a hiring screen on a band.

What this pass should not be asked to do. It should not be rounded into one dollar per card. The publishable unit from these 48 pages is the WRT interval, −18% to +13%, N=8, with ASD, AMRT, and FSRT left unpriced because they never appear except on top of WRT and their counts are 1, 2, and 4.

The homework, in files

The full dataset — all 48 postings, with employer, metro, role, the band as printed, each cert’s status on the page, experience requirement, and the source URL — is published with this article: cert-value-dataset.csv. The rejects log — the blocked sources, the shell pages, and every lead that did not survive a page read — is here: cert-value-rejects.txt. The method is the second section of this article; the limitations are stated in “What this study cannot say,” and they are part of the finding.

Related, on this site: the IICRC certification and restoration training guide this measurement was built to inform.