Evidence first
Pages prefer official careers pages, active postings, employer policy pages, government references, and public filings.
HireTea Independent hiring research
Methodology
How HireTea checks a hiring claim
HireTea compares dated employer pages, active posting language, government references, and structured research records. This page shows what evidence carries weight, what stays out, and what you should still verify for your exact role and location.
13 company hubs in this review scope
158 source-backed topic facts
253 linked source records reviewed
242 claims checked in source audits
Quick answer
What has to be true before a claim is published?
A page is strongest when it combines a current public source, company-specific hiring context, a clear limitation note, and applicant-focused next steps. Official employer pages and current postings carry the most weight. Secondary sources can help identify what applicants ask about, but they do not override employer language or current public instructions.
Limits stay visible
When details vary by role, location, department, property, franchise, or posting, the page says so directly.
Weak pages wait
Pages with thin evidence can stay out of the index until the source trail is strong enough for applicants.
Company dossier evidence key
How to read every company guide
Company hubs keep the employer-specific decision, cited evidence, unresolved questions, and next actions close to the applicant. The shared interpretation rules live here once so they do not interrupt every dossier. When a current posting, candidate account, recruiter message, or written offer is more specific than a retained public record, use the current role-specific instruction.
| Label in a company guide | What it means | What should control your decision |
|---|---|---|
| Source-backed or verified | A cited public record was retained and reviewed. It does not promise that a dated opening is still available. | The active posting, final Apply destination, candidate account, and written offer. |
| Dated posting sample | A page or requisition was observable on the row's date. It can change, close, refresh, or be reposted later. | The saved job ID, URL, observation date, current posting state, and application receipt. |
| Current public dossier | The facts, source links, and latest source-review date visible on the company page now. | The company page's current citations and its company-specific known limitations. |
| Historical audit batch | A past research batch used to check source rows and claims. Its totals are not an opening count or the current public source count. | The current dossier for applicant decisions; historical totals only describe the review process. |
| Editorial inference | HireTea's synthesis of cited duties, qualifications, role language, or source patterns—not an employer scorecard or published rule. | The exact role language and only examples that are true for the applicant. |
| Practice prompt | A preparation aid derived from the role evidence, not a question attributed to the employer. | The interview invitation, current duties, and the applicant's truthful experience. |
Application route
Confirm that the final Apply destination belongs to the employer, operator, or named recruiting vendor for that requisition.
Application status
Record the exact account or message state and its date; a portal label alone does not establish a hiring outcome.
Pay and schedule
Verify the amount, pay basis, shift, location, and employing entity in the active posting or written offer.
Review scope
What this methodology currently covers
The numbers below describe the same company set currently used by the public sitemap. The purpose is not to claim perfect coverage of every employer. The purpose is to keep the published guide set focused on pages with enough public evidence, clear caveats, and applicant value to support real application decisions.
| Methodology signal | Current count | How it is used |
|---|---|---|
| Company hubs in scope | 13 | Defines the employer set used by the public directory, source inventory, and consolidated company facts. |
| Source-backed topic facts | 158 | Counts company-specific facts consolidated into the published company hubs. |
| Linked source records reviewed | 253 | Shows whether pages are grounded in public references rather than only broad archetype assumptions. |
| Preferred-source signals | 252 | Counts sources that are not classified as low-confidence signals by the source-quality filter. |
| Lower-confidence signals | 1 | These can identify questions to verify, but they should not be used as the final rule by themselves. |
| Source-backed topic facts | 158 | Counts fact-sheet entries that connect a hiring topic to a public source and access date. |
| Audited source rows | 353 | Historical Source Register rows across recorded audit batches from 2026-05-24 through 2026-07-12; this is not the current public-link count. |
| Curated audit sources added | 163 | Historical audit-batch count of sources promoted into fact sheets. Later refreshes can change the current public source list without rewriting that batch total. |
| Checked hiring claims | 242 | Counts claims reviewed for support, limitations, exclusion, or fresh verification in the source-verification files. |
| Claims excluded pending stronger evidence | 21 | Counts items kept out of public guidance until a fresher or stronger source resolves them. |
| Latest source access date | 2026-07-22 | Helps readers judge whether a page should be rechecked before a high-stakes decision. |
Verification process
How historical source-verification batches are used
Each company fact sheet preserves totals from a historical source-register and claim-verification batch. The recorded batch dates currently run from 2026-05-24 through 2026-07-12, while individual public sources and field reports can have later access dates. Audit rows do not automatically become page copy. HireTea promotes a curated subset, keeps limitations visible, and excludes claims that need fresher or stronger evidence. Historical batch totals and the current public source list are reported separately.
A date_accessed value is the documented source-check date. URLs checked in one research pass can
honestly share a batch date; a newer date is recorded only after a new check. HireTea does not disperse dates
with random, hash-based, or cosmetic rules, and an access date never guarantees that a posting is still open.
| Review outcome | How it is used | Public-page effect |
|---|---|---|
| Retained with support | The claim or source can support a fact-sheet source trail when it matches the current page context. | Eligible for source badges, citation metadata, and public source-coverage reporting. |
| Retained with a limitation | The source is useful, but role, location, time, or policy limits must remain visible. | Shown with caveat language rather than treated as a universal rule. |
| Excluded | The item should not support public copy without replacement evidence. | Separated into cleanup notes instead of silently changing applicant-facing facts. |
| Needs fresh verification | The item may be useful but needs a fresh source, current page, or stronger confirmation. | Held out of stronger claims until a later review resolves it. |
Source hierarchy
Which sources get the most weight
Official employer sources
Careers pages, active postings, benefits pages, policy pages, press releases, recruiting pages, and public company materials are preferred because they are closest to the employer's current process.
Public authority sources
Government labor, civil-rights, consumer, and workplace references are used when a page needs legal-rights or state-specific context. They explain applicant rights, but they do not replace the current posting.
Current posting evidence
Active postings can clarify pay range, schedule, location, certification, role title, department, remote status, travel, and first-week expectations. Saved posting evidence is often the best local detail.
Secondary signals
Review sites, scraped listings, forums, and short social posts can reveal questions applicants ask, but they are treated as prompts for verification rather than final proof of a company-wide rule.
Publication gates
Publication gates that protect the guide set
HireTea does not publish every possible route as if it were equally useful. Material outside the focused guide set stays unpublished until it has enough distinct evidence and applicant value. This keeps public pages focused on guidance that is specific, sourced, and useful to an applicant making a real decision.
| Gate | What must be true | Why it matters for applicants |
|---|---|---|
| Company fit | The employer is part of the current published guide set and has a structured fact sheet. | Readers get company-specific language instead of a generic career article. |
| Source-backed topic | The topic has a public source, access date, and clear caveat when local practice can vary. | Applicants can check the source trail before changing their plan. |
| Useful page body | The page includes decision guidance, evidence to save, comparison notes, and next-step context. | The page helps with an application decision rather than only repeating a short answer. |
| Safe published guide set | Higher-risk or thin pages are kept out of the published guide set until they are stronger. | The public site avoids foregrounding pages that could be misunderstood or lack enough context. |
| Correction path | Every public page links to contact and editorial pages so source updates can be sent for review. | Readers have a clear way to report outdated public information. |
Derived datasets
How HireTea calculates and labels comparison fields
HireTea publishes both source facts and derived fields. A source fact paraphrases a cited public page. A derived field is a reproducible editorial calculation or coding decision made from the structured fact sheets. Derived values help compare the current set, but they are not employer ratings, selection probabilities, or promises about an individual application. In an additive score, missing evidence contributes zero; the limitations below explain why zero must not be silently interpreted as a favorable employer fact.
The 2026 Application Transparency Report uses a separate, non-scored codebook. Its five indicators remain documented, partial, or not observed; missing public evidence is never converted into a zero or folded into a company ranking.
The version 2026.1 published-cohort appendix is a separate raw-evidence release:
it preserves 58 dated live_posting records for the
current 16-company public cohort, observed from
2026-07-12 through 2026-07-22. It does not
claim that a retained posting remains open, and it is not added to the report's fixed 25-employer or 37-record
analysis denominators.
| Derived field | Exact rule | Known limitation |
|---|---|---|
| 2026 report evidence states | One manually coded row per fixed employer dossier records application route, status return path, posting identity, pay specificity, and schedule detail as documented, partial, or not observed. The report and both CSV files use the same versioned 25-company module. Posting-level pay uses a separate denominator of 37 structured records with an ID, URL, and observation date. | Not observed means the reviewed public sources did not establish the indicator; it is not evidence of absence. The 25 employers and 37 postings are purposive research samples, not market prevalence estimates. |
| Published-cohort posting evidence appendix |
Version 2026.1 exports one normalized row for each retained
live_posting record in the current public cohort: employer boundary, location, pay and
schedule evidence, Apply destination, source URL, observation date, and an opening-state caveat.
| The 58 rows are dated source records from 2026-07-12 through 2026-07-22, not a current-opening census. They document the 16-hub publication cohort without changing the fixed report's 25-company matrix, 37-record posting sample, or 22-of-37 pay denominator. |
| Pay transparency score (0–100) | +25 when the structured starting-pay field contains an explicit currency amount; +25 when the source-backed pay fact contains at least two currency amounts; another +20 at four currency amounts; +20 when that multi-anchor fact was accessed 0–365 days before the dossier's update date; +10 when it names role or location context; and +10 for union or collective-bargaining context. The total is capped at 100. | This measures evidence present in HireTea's fact sheet, not employer generosity. A high score can describe a wide range, and a low score can reflect missing public data rather than a low wage. |
| Assessment clarity score (0–100) | +30 for a named, non-generic first structured assessment; +20 for assessment-specific timing; +20 for an assessment-specific item count; +20 when structured criteria, dimensions, measures, or content contains at least three entries; and +10 for restart, resume, return, or completion behavior. The score is capped at 100. | The score measures public explanation, not difficulty. A role can have no assessment, and a privacy notice listing possible assessment data does not prove every applicant receives one. |
| Source depth score (0–100) | Legacy rows without a verdict plus sources marked KEEP or KEEP_WITH_CAVEAT are scoreable; dead or unverified rows are excluded. Each observed tier contributes 25 points and up to five distinct scoreable URLs contribute 5 points each. Tier 1 is official-company evidence; Tier 2 is regulator, official-vendor, or franchise/operator evidence; Tier 3 is mirror, third-party, user, or archive evidence. | A caveated source can add depth without removing its visible limitation. The value measures source variety and distinct URL count, not whether every claim is complete. |
| Documented process load / application friction field (0–100) | +20 when at least one structured assessment has a meaningful name; +20 for a meaningful async-video tool; +10 per meaningful recorded interview round up to three; and +15 for a predefined franchise-sensitive employer set. The score is capped at 100; other screening fields do not add points. | This is a documented-evidence index, not an ease ranking or a statement that every local applicant completes every recorded step. A low value can reflect missing public detail, while franchise sensitivity reflects process variation as well as effort. |
| Tracker group | One dominant attention group and a company-specific reason are explicitly stored for every published slug: local-site, fast-response, interview handoff, pay/schedule, portal status, or evidence-first. The current CSV exports both fields, so the editorial code can be inspected without reconstructing a keyword heuristic. | The group organizes tracker attention; it is not an employer-defined application status. A company can display several patterns even though the snapshot records one dominant code. Changing the code requires an explicit editorial data change rather than an accidental keyword match. |
| Interview practice themes | HireTea's structured editorial layer codes likely signals, practice prompts, and evidence angles from duties, qualifications, values language, and cited role evidence. Public pages label these fields as editorial inference unless an individual source explicitly attributes the wording to the employer. | They are preparation aids, not leaked questions, employer scorecards, or predictions. The active posting and interview invitation take priority. |
Snapshot date
The downloadable CSV is regenerated at build time from the published company set and carries each dossier's own update and field-review dates.
Missing values
Blank fields remain blank or receive zero only where the published formula is additive. They are not imputed from another employer or category average.
Reproduce the row
Open the linked company hub, inspect its source records and dates, then compare the displayed structured fields with the rules above.
Open the 2026 transparency report and codebook Download the 25-company evidence-state CSV Download the 37-record posting CSV Download the 58-record published-cohort appendix Download the current research snapshot CSV See the pay-transparency formula applied See the combined application evidence model
Source mix
Most common source domains in the current indexed set
Domain counts are not quality scores by themselves. A domain can appear often because it hosts many current job postings, benefits pages, or recruiting pages. The useful question is whether the source is close enough to the employer's own process to support the specific claim on the page.
| Source domain | Links | Lower-confidence signals | Example company coverage |
|---|---|---|---|
| careers.fedex.com | 24 | 0 | FedEx |
| chick-fil-a.com | 21 | 0 | Chick-fil-A |
| careers.walmart.com | 17 | 0 | Walmart |
| hiring.amazon.com | 17 | 0 | Amazon |
| careers.dunkindonuts.com | 12 | 0 | Dunkin' |
| jobs.tacobell.com | 11 | 0 | Taco Bell |
| corporate.target.com | 10 | 0 | Target |
| jobs.tjx.com | 10 | 0 | TJX Companies |
Role-family coverage
How source coverage differs by role family
Some categories have centralized employer careers pages. Others depend more on store, property, franchise, facility, or location-level postings. That is why HireTea pages separate broad company preparation from details that must be verified in the current posting.
| Role family | Company hubs | Linked source records | Source-backed facts | Examples |
|---|---|---|---|---|
| Retail | 6 | 90 | 83 | Walmart, Target, Kroger, and TJX Companies |
| Restaurant | 5 | 106 | 51 | McDonald's, Starbucks, Chick-fil-A, and Taco Bell |
| Warehouse | 2 | 57 | 24 | Amazon and FedEx |
Reader use
What applicants should still verify
A sourced page is a starting point, not a substitute for the active posting. Before applying or accepting an offer, verify the exact role title, job ID, pay language, schedule, location, department, required documents, candidate-portal messages, and written next steps. If a posting or recruiter message conflicts with a HireTea summary, follow the current employer instruction and send the source so the page can be reviewed.
Use company pages for preparation
Company hubs explain the likely hiring context, terminology, source-backed topics, and clearly labeled editorial practice inferences.
Use comparison tools for planning
The comparison worksheet and pay research help organize evidence, but they are not promises about a specific location or manager.
Use resources for evidence
Save postings, portal messages, interview notes, and offer details before links expire or instructions change.
Use updates for corrections
Send the page URL, current public source URL, and date viewed to jeongyuseck@gmail.com.
Reader check
Did this page help you decide your next step?
Your response stays in this browser. Analytics receives only the selected category.
Found a source problem? Email the editor.
What was still missing?