Skip to content

Tenure Comp Intelligence

How the Pay Index works

We collect compensation from primary sources, verify each row against a minimum sample size, and refresh the dataset every quarter. This page documents the categories, statistical approach, and what we deliberately do not publish.

Bands published
804
Sectors covered
12
Verified data points
3,899
Last refreshed
21 days ago

What is the Tenure Pay Index?

The Tenure Pay Index is verified compensation data across 12 white-collar sectors in UAE and Saudi. Each row pairs a role with a seniority level and reports a percentile distribution for total monthly cash. The unit of analysis is the role, not the individual, so no figure traces back to a named person. Coverage follows source depth rather than editorial preference: a band appears once enough independent sources support it, and stays unpublished until then. The dashboard shows the source count on every band, so a buyer can see the evidence behind a number before using it.

How does Tenure source compensation data?

Every figure traces to one or more primary sources. Live job postings supply compensation disclosed in ads on employer career sites, regional boards, and aggregator feeds. JD signal mining extracts pay language buried inside job descriptions, scored for reliability before it enters a band. LinkedIn open-salary data captures figures professionals have chosen to make public. Candidate submissions add compensation shared through the consumer side of Tenure, tagged verified when payslip-backed. Each band stores its source breakdown, so the dashboard can show how many of each type stand behind a figure. Tenure works from these primary sources, rather than crowd-sourced guesses.

What is the minimum number of sources before a band gets published?

Three. A role-and-level band appears in the Pay Index only once at least three independent sources support it. Below that line, Tenure holds the data back rather than ship a row too thin to trust. Bands with three or four sources carry a limited-data marker and read as directional. Bands with five to nine sources are suitable for setting pay ranges and benchmarking. Bands with ten or more are suitable for offer decisions and leadership reporting. The source count next to each band is the signal. There is no separate rating to interpret.

How does Tenure handle outliers?

Every record passes a normalization and outlier check before publication. Currency, pay period, allowance separation, and bonus inclusion are resolved first, so figures compare on the same basis. Any band more than 2.5 standard deviations from the role-level median is flagged for human review before it can publish. Where sources disagree sharply on the same role, the conflict is surfaced internally and either resolved or held back, rather than averaged into a misleading midpoint. The result is bands that tighten when sources agree and widen when they do not, which is the honest signal HR teams need.

How often is the data refreshed?

The Pay Index refreshes every quarter. Underneath, source ingestion runs continuously for job postings and weekly for LinkedIn and candidate paths, so new evidence flows in throughout the quarter rather than in one batch, then lands in the published bands at each quarterly refresh. Every band stores its own last-updated timestamp, and the dashboard surfaces it on each row, so a buyer can see how fresh a specific number is before relying on it. This is the practical difference from an annual survey: the bands move when the market moves, not 12 to 18 months later.

Why only UAE and Saudi?

These are the two markets where Tenure has enough verified sources per sector to publish credible percentile distributions. Narrowing the scope is how the accuracy holds: two markets covered well beats nine covered unevenly. Qatar bands were unpublished in June 2026 when their source depth fell below the publish bar, and the market returns when re-verified sources clear it. Tenure would rather hold a market out of scope than ship bands it cannot stand behind.

What does Tenure not claim?

Tenure does not claim to cover every employer or every role in the Gulf. Coverage depth varies by sector, and the source count per band shows exactly where it is thin. Tenure does not back-fill missing data with synthetic figures: where the evidence is thin, the row is held, not interpolated. It does not adjust bands for inflation, currency drift, or cost-of-living differences, since those calls belong to the customer. Equity, long-term incentives, and deferred compensation are out of scope for now, because almost no source captures them cleanly in the Gulf.

How is candidate-submitted data handled?

Candidate submissions enter through the consumer side of Tenure, tagged verified when payslip-backed and unverified otherwise. No row in the Pay Index identifies a person. Aggregation requires at least three sources per band, and the submitter's identity stays in the source table, never joined into the published index. The band-level source breakdown holds no identifying metadata. Verified submissions weigh above unverified ones in the percentile calculation. The Privacy Policy at business.tenurecareers.io/privacy documents PII handling end to end, including every data processor Tenure uses.

What does the source count mean on a sector ladder table?

The Sources column on a sector seniority ladder shows the total verified observations at that level, summed across every role mapped to that level in the sector. If a VP row reads 23, that is 23 verified observations across all VP-level roles in the sector added together, not the count behind a single title. This is deliberate: the ladder answers what evidence stands behind the band a buyer is about to use, and the level total is the honest figure for that. Per-band counts are more granular: the sample table higher on each sector page shows the source count for each individual band, and the dashboard breaks every band out by title, firm tier, and city.

How do we normalize Gulf job titles?

Job titles in the Gulf vary widely between employers. The same person might be a Vice President at one bank, a Director at a second, and a Senior Manager at a third. Tenure maps each observed title to a canonical role within its sector before computing bands, so percentiles are not distorted by title inflation at one employer. Normalization accounts for jurisdiction qualifications, language requirements, and regional experience where they shift the level. The published ladder reflects the canonical role, while the dashboard preserves the underlying titles, so a buyer can trace how a figure was placed.

How does Tenure compute the percentile bands?

Each band reports five percentiles: P10, P25, median, P75, and P90. Verified sources weigh more than unverified ones, recent observations weigh more than older ones, and statistical outliers are excluded before the band is computed. The approach is deliberately more conservative than a flat average: it produces tighter bands where sources agree and wider ones where they diverge. Tenure lists the categories of sources behind a band without publishing the exact weights or thresholds, because disclosing those would invite gaming of the source pool and degrade the data for every customer.

Where does Tenure's data not come from?

Tenure does not build bands from anonymous internet forums, single-source assertions, recruiter rumor, or self-reported figures that cannot be corroborated. A lone unverified number never becomes a band on its own; it waits for independent corroboration or stays out. Tenure also does not publish its vendor stack, recruiter-partner names, the prompt structure behind JD signal mining, or the exact weighting parameters, because those details would let the source pool be gamed. The full list of data processors, including the AI provider used for signal mining, is disclosed on the Privacy Policy as PDPL and GDPR require.

Sector-specific methodology

The methodology above applies across all 12 sectors. Sector-specific notes on source mix, role normalization, and sample-size caveats sit at:

Change log

Date Version Change
17 May 2026 v1.0 Initial methodology, two-market launch (UAE, Saudi).
25 May 2026 v1.1 Qatar added. B2B Pay Index now covers UAE, Saudi, and Qatar.
25 May 2026 v1.2 Q&A restructure for AI extractability; ladder source-count aggregation method disclosed; sector cross-links added.
6 June 2026 v1.3 Qatar unpublished pending source re-verification. Sector taxonomy consolidated to 10 canonical sectors; per-sector links re-pointed.
9 June 2026 v1.4 Real Estate & Construction and Energy added; coverage now spans 12 canonical sectors. Refresh cadence stated as quarterly throughout.