Background and Context

San Bernardino County is the largest county in the continental United States by land area. It stretches from the Inland Empire’s urban core to mountain communities, high desert posts and rural places such as Joshua Tree. Roughly 2.2 million residents rely on county services. County government employs about 25,000 people in more than 40 departments. The service footprint looks more like a small state than a typical county: public health, behavioral health, sheriff and probation, land use, airports, regional parks, social services, information technology and more.

Representation is complex. Employees belong to numerous collective bargaining units, each with its own agreement structure, premium rules and bargaining cadence. Recruitment and retention pressures vary dramatically between urban and remote worksites, so labor data must be segmented and defendable at a granular level.

Classification and compensation work sits in Human Resources under Jenna York, who leads the Classification & Compensation function. Jenna reports to Gina, the County Human Resources Director. Gina reports to Leo in the County Administrative Office and ultimately to the Chief Executive Officer. That chain matters. The Class & Comp team must reconcile department requests, labor proposals and countywide policy before changes move forward. Timely, credible data is the fuel that keeps that process moving.

Challenges

Three intertwined issues dominated the workload: scale, data quality and speed.

Out-of-class claims. Analysts received a steady stream of complaints that employees were working “out of class.” Each claim required fact finding, a review of duties, and a decision on whether to reclassify, adjust pay or take no action. The breadth of the county’s catalog made expert review difficult. No one analyst could know every job.

Market data burden. Each review also required salary comparisons to benchmark jurisdictions. Staff pulled ranges, trimmed outliers and tried to calculate a defendable median or market position. This clerical work absorbed large amounts of analyst time. California vacancy reporting rules added pressure: persistent vacancy levels in a bargaining unit can trigger additional scrutiny, so HR needed fast, well sourced numbers.

Aging IT classifications. Jenna knew a long list—well over one hundred IT-related classifications covering many hundreds of incumbents—had drifted away from current labor-market practice. Only two HR staff members were comfortable drafting changes to highly technical IT roles, creating a bottleneck.

PDQ backlog. Departments submitted Position Description Questionnaires (PDQs), often long narratives. Analysts had to convert them into structured class specifications with minimum qualifications, duty statements and KSAs. Historically the county cleared only the low dozens of PDQs in a full year. That pace could not keep up with a 25,000-person workforce.

Ramifications

Lingering out-of-class work drove equity and morale concerns and raised potential back-pay exposure. Outdated minimum qualifications and duty statements weakened recruitment in competitive areas such as information technology and health care. Vacancy hot spots triggered repeated data calls and reactive spreadsheet work. The time spent assembling and cleaning market data left less analyst capacity for proactive consultation with departments and labor. Version control suffered as PDQs, draft specs and salary files moved by email.

Summary of Solution

San Bernardino County partnered with Holly to create a common data foundation and accelerate analysis. Using Holly Poly, the county loaded its current salary schedules and comparator data from peer jurisdictions into one environment. Holly’s AI mapped relationships among internal classifications and external benchmarks. It produced similarity scores that showed which titles truly compared even when naming conventions differed.

On that data layer, Holly generated structured draft class specifications from PDQ inputs in county format. The platform flagged patterns consistent with out-of-class work, highlighted where minimum qualifications or duty clusters diverged from market practice, and surfaced salary range anomalies that could help explain vacancy problems.

Brendan Hellweg, Holly’s founder, underscored the scale that drove the engagement: “San Bernardino County is one of the largest counties in the United States. It has about 25,000 employees across more than 40 departments.” Reflecting on throughput, he described the shift the platform enabled: “In the low dozens per year versus 600 PDQs processed in about a month.”

Traditional alternatives were weak fits. A comprehensive consulting study would likely run a year, cost well into six figures and result in a static report that would begin aging immediately. Staying fully in-house would have kept the team on a low-dozens-per-year treadmill while vacancy and bargaining pressures grew.

Outcomes Achieved

With Holly, the county processed roughly 600 PDQs in about a month, a dramatic increase over historic pace. Analysts received first drafts already structured and source-linked, so time shifted from clerical assembly to professional judgment.

Early analytic passes surfaced a sizable set of employees whose duties suggested they were working above or below their assigned classifications. Roughly one quarter of the reviewed positions showed some degree of misalignment. That signal helped the county target deeper follow up where exposure was greatest.

The long-stalled review of IT classifications finally moved. Current labor-market data and templated draft language made it easier for the county’s two technical SMEs to scale their input. Because salary and comparator data now lived in one system, HR could respond to vacancy and bargaining data calls more quickly and with clearer sourcing notes. Labor discussions shifted from disputes over spreadsheet lineage to a focus on options. Cleaner duty statements fed directly into recruitment materials for hard-to-fill roles.

Future of the Partnership

San Bernardino County is expanding use of Holly analytics to monitor vacancy clusters and refresh benchmark data between bargaining cycles. The team is exploring automated conversion of approved class specs into candidate-facing recruitment bulletins and adding compliance checks keyed to California regulatory updates. Having seen Holly operate at county scale, HR is planning a rolling review cadence so classifications do not fall years out of date. Lessons from San Bernardino are informing features Holly is building for other large, unionized jurisdictions.