How Holly Sources Data
Last updated: October 13, 2026
Learn about the types of data Holly collects, where it comes from, how Holly's team checks it for quality, and how salaries are calculated for benchmarking.
What Data Holly Collects
Holly collects three types of data for your agency and your peer agencies within the Holly platform:
Class Specifications (Job Descriptions): The full text of each position's class spec, including definition, typical tasks, minimum qualifications, knowledge and abilities, and license requirements. This powers Holly's similarity analysis, AI-assisted editing, and benchmarking features.
Salary Schedules: Pay ranges and step tables for each position, including minimum, midpoint, and maximum salaries. Holly uses this data to populate salary benchmarking in both the Internals and Market tabs.
MOUs (Memoranda of Understanding): Collective bargaining agreements linked to each classification by bargaining unit. These are stored as reference documents in ChatMOU and surface alongside class specs in the Market tab's detailed view.
Where to Find This Data in Holly
The data Holly collects surfaces in three places within a class spec:
Source documents (spec, salary schedule, and MOU) — To see the underlying files Holly has on file for a classification, see The History Tab.
Salary comparisons — To compare a classification against your agency's own classifications and salary ranges or against peer agency data, see How Benchmarking Works in Holly.
Where the Data Comes From
Holly's data team researches and collects compensation data from official, publicly published sources. For most jurisdictions, this includes:
Job description platforms like GovernmentJobs (NeoGov), which host class specs
Salary schedule PDFs published on agency websites or within MOUs / CBAs
Direct agency websites for jurisdictions not on NeoGov
For jurisdictions where compensation data is not publicly posted, Holly may pursue public records requests (FOIA) or work with third-party labor market data providers with client permission. Because Holly collects data from the agency's own official documents, the published source is treated as the source of truth. Every salary figure and class spec in Holly is linked back to the document it came from, so you can always verify it yourself in the History tab.
How Data Gets Into the Platform
Holly's data pipeline follows these steps for every jurisdiction:
Research — Identify official public sources for class specs, salary schedules, and MOUs, and confirms each source belongs to the correct jurisdiction.
Collect — Extract data from PDFs, web pages, and structured files into a standardized format.
Clean and Match — Standardize formatting, convert pay to annual figures, and match job titles and codes between salary schedules and class specs.
Validate — Quality check the data (process described below) before it is approved for import.
Import — Load approved data into Holly with links to each source document and its effective date.
How Holly Checks Data Quality
Holly's data quality process is human-led and supported by automated checks. Every dataset is reviewed by a member of Holly's data team before it is first imported into the platform. As our tooling grows, more of these checks will be automated, but a person will always stay involved in the review.
Checks on every new dataset
Before any jurisdiction's data is first imported, it goes through both automated and manual review:
Source checks — Confirming that each source link points to the correct jurisdiction (for example, the right "Milpitas" in the right state) and that the number of classifications looks reasonable for the agency's size.
Completeness checks — Flagging records missing required fields such as job title, job code, base salary, effective date, or bargaining unit.
Plausibility checks — Flagging salaries that fall outside a reasonable annual range, salary steps that decrease instead of increase, and non-numeric or malformed values.
Duplicate checks — Flagging duplicate job titles and placeholder or administrative entries that aren't real classifications.
Manual spot checks — A reviewer compares a sample of positions directly against the agency's source documents to confirm the salary data matches.
If a check turns up an issue, the data is corrected and re-checked before it is approved for import.
Checks during quarterly refreshes
During each quarterly refresh, Holly compares incoming data against what's already in the platform to identify what has changed. For more on the refresh process, see How and When Holly Refreshes Your Data.
How issues are tracked
When Holly's team finds a data quality issue, it is logged internally against that agency's record, along with any notes on the latest import. This gives Holly's team a running history of known issues for each jurisdiction, which is reviewed again at the next refresh.
When Source Data Is Inaccurate or Incomplete
Holly's accuracy depends on the quality of each agency's published data. Here's how Holly handles common problems.
When the published data appears to contain an error
If Holly's team spots something that looks wrong in an agency's published data (for example, a salary that seems far too high), we note it internally but do not alter the underlying data. This keeps what you see in Holly consistent with the agency's official documentation, which is what your stakeholders and bargaining partners will reference.
If you believe a figure is wrong, check the linked source document first, then contact Holly Support at support@hollygov.com.
When data is incomplete
Not all jurisdictions publish complete data publicly. Here's how Holly handles common gaps:
No class specifications available — Holly will still include job titles and salary data. Classifications without descriptions are unlikely to appear as suggested benchmarks, since Holly's similarity analysis relies on the full text of a class spec. However, they can always be added to a benchmark set manually.
No salary data available — Holly will display the classification and description without salary information.
How to Verify a Comparator's Source
Because every figure in Holly is linked to its source, you can check any comparator's data yourself, which is useful when you need to document your methodology or answer questions from leadership or bargaining units.
Open a class spec and click the Market tab.
Scroll down to Configure market benchmarks.
Open the comparator classification you want to check.
Select Sources.
Click View under the Salary column to open the source document.
To see when a class spec was last imported and where it came from, open the spec's History tab.
Salary Calculation Methodology
What Source Holly Uses for Salaries
Holly pulls salary data from each jurisdiction's officially published salary schedule — the document an agency uses to define pay ranges for each classification. Holly only uses salary figures from a job description if no separate salary source exists, since salary schedule figures tend to be more current.
How Holly Calculates Annual Salary
Salary schedules publish pay in a variety of formats depending on the jurisdiction. If an annual figure is provided, Holly uses it directly. Otherwise, Holly calculates the annual value from the pay frequency shown. Holly standardizes all salaries to an annual figure using the following conversions:
Pay Frequency | Conversion |
Hourly (40 hrs/week) | × 2,080 |
Weekly | × 52 |
Biweekly | × 26 |
Monthly | × 12 |
Some jurisdictions use non-standard calculation methods that may produce slightly different annual figures than Holly's standardized approach. If you notice a discrepancy between what's shown in Holly and an agency's records, reach out to Holly Support at support@hollygov.com.
Salary Steps
Where a position has multiple salary steps, Holly retains each step individually. The minimum salary shown in benchmarking views reflects Step 1 (or the lowest listed rate), and the maximum reflects the highest step. Midpoint is calculated as the average of minimum and maximum.
Special Position Categories
Some position types do not fit a standard salary schedule format. Holly handles the following as best as possible given available data, though coverage may be limited:
Part-time and contract positions
Public safety roles with non-standard 40-hour workweeks (e.g., firefighters, police)
Healthcare workers
Seasonal and temporary workers
Elected officials and commissioners
Frequently Asked Questions
Are there minimum data quality standards for participating agencies?
Comparator agencies don't submit data to Holly, so there are no standards for them to meet. Instead, Holly applies its own standards to the data it collects: it uses only official published sources, requires key fields for each record, and runs the completeness, plausibility, and duplicate checks described above before import.
Does Holly run automated data quality checks?
Yes. Automated checks flag missing fields, implausible salary values, decreasing salary steps, malformed data, and duplicates. These checks support a human reviewer rather than replace one: every new dataset is reviewed by a member of Holly's data team before import.
How does Holly handle inaccurate or incomplete external data?
Issues are caught during review and logged against the agency's record. Holly corrects errors introduced during collection, but treats the agency's published documents as the source of truth and does not change the agency's official figures. Every data point links to its source, so you can verify it at any time.
If you believe data for your agency or a peer agency is missing or inaccurate, contact Holly Support at support@hollygov.com.
Related Articles
How and When Holly Refreshes Your Data: When quarterly refreshes happen and how your in-progress work is protected
How to Add and Manage Comparator Specs: How to add, remove, and adjust comparator classifications in your benchmarks
How Holly's Spec Similarity Score Works: How Holly calculates match scores between your specs and comparator specs
The History Tab: How to see when a class spec was imported and where its data came from