Benchmarking data maturity against peer jurisdictions

A city CIO in a mid-size US city (population roughly 250,000) pulls up a self-assessment: "we have an open data portal, a chief data officer, and a few dashboards." Sounds solid. Then someone lines it up against five comparable cities using a public data-maturity index. That city ranks last on data governancedata governanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → and second-to-last on staff capacity. The dashboards were a distraction from a much weaker foundation. This is the value of benchmarking: it converts a vague sense of "we're doing okay" into a specific, comparable score.

Why benchmarking beats self-assessment

Self-assessment is biased by default. Every department believes its data practices are "pretty good" because there's no external reference point. Benchmarking fixes this by scoring jurisdictions against a shared rubric, then comparing.

Two well-known frameworks:

• GovEx Open Data Maturity Model (from the Center for Government Excellence at Johns Hopkins), used by cities in the What Works Cities network.
• Urban Institute's data capacity assessments, which look at cross-agency data sharing, analytics staffing, and use of data in budget decisions.

Both frameworks score a jurisdiction across dimensions rather than giving one overall grade. That's the key design choice: an aggregate score hides exactly the weak spot you need to find.

The data that matters: what gets scored

Maturity indices don't evaluate abstract "data culture." They score concrete artifacts and processes. Typical dimensions include:

1. Data inventory and cataloging

Does the city know what datasets it holds? A data catalogdata catalogA centralized inventory of an organization's data assets, enriched with metadata, that helps people find, understand, and trust the data they need.View full definition → (a searchable inventory of datasets with descriptions, owners, and update frequency) is the baseline. Many cities fail here first: nobody can list all datasets held across departments.

2. Open data publishing

Number and quality of datasets on an open data portal (e.g., Chicago's, San Francisco's, or NYC's Open Data portal). Quality matters more than count: a dataset last updated in 2019 counts against you even if it's technically "published."

3. Data governance structure

Is there a Chief Data Officer (CDO) role, a data governance committee, and documented data-sharing agreements between departments (e.g., between police and public health)? Governance without a named accountable owner rarely survives budget cuts.

4. Data qualityData qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → management

Documented standards for accuracy, completeness, and timeliness. Example: does 311 (non-emergency service request) data get validated for duplicate entries and address accuracy before it hits a public dashboard?

5. Analytics capacity and staffing

Number of FTEs (full-time equivalents) doing data analysis versus data entry. A city with 3 data analysts per 100,000 residents is structurally different from one with 0.3.

6. Use in decision-making

Is data actually used in budget hearings, performance reviews, or service redesign, or does it sit in a PDF nobody reads after publication?

Peer selection: the step most cities get wrong

Comparing a city of 200,000 to New York City is useless: different budget scale, different staff depth. Peer groups should match on:

• Population band (e.g., 150,000 to 400,000)
• Metro/regional context (urban core vs. suburban)
• Government structure (council-manager vs. strong mayor, which affects who controls IT budgets)
• Comparable fiscal capacity

The International City/County Management Association (ICMA) and the U.S. Census Bureau's Government Finance and Employment data are common sources for building a defensible peer set. Get this wrong and every subsequent comparison is noise.

A worked example: scoring three dimensions

Say a city self-scores against a simplified 0 to 4 maturity scale (0 = no practice, 4 = institutionalized best practice) across three dimensions, benchmarked against the median of four peer cities:

The composite average (1 + 2 + 0.8-equivalent-scaled) might look "middling" at first glance. But averaging hides the story: this city is roughly on par on governance, yet two full maturity levels behind on cataloging and running at about 38% of peer staffing capacity (0.8 / 2.1 ≈ 0.38). That staffing gap is the actionable finding, not the composite score.

This is the core lesson: never report a single blended maturity score to decision-makers. Always show the dimension breakdown. A composite score of "2.3 out of 4, slightly below peer average" invites no action. A staffing gap of 62% invites a budget conversation.

Governance metrics that signal real maturity

Beyond the index scores themselves, a few governance metrics are worth tracking independently because they predict future maturity:

• Dataset freshness rate: % of published datasets updated within their stated schedule. Below 70% is a red flag common in under-resourced open data programs.
• Data sharing agreement count: number of active, documented interagency agreements (e.g., housing department sharing eviction data with public health). Zero or one is typical for low-maturity cities; five-plus signals institutionalized collaboration.
• CDO tenure and reporting line: a CDO reporting directly to the city manager or mayor's office tends to have more durable authority than one buried three levels down in IT.

A simple query analysts often run internally to check freshness (using a dataset's last-updated field from a catalog export):

import pandas as pd
from datetime import datetime

catalog = pd.read_csv("open_data_catalog.csv")
catalog["last_updated"] = pd.to_datetime(catalog["last_updated"])
catalog["days_stale"] = (datetime.now() - catalog["last_updated"]).dt.days

stale_pct = (catalog["days_stale"] > 365).mean() * 100
print(f"Datasets stale over 1 year: {stale_pct:.1f}%")

This is the kind of check a data governance team should run quarterly, not something discovered during an external audit.

Analytics and measurement benchmarks

Once governance basics are in place, benchmarking shifts to outcomes: is data actually improving services?

Common benchmarks used in the What Works Cities certification process (a recognized US standard for city data practice):

• % of budget decisions citing performance data: mature cities target above 50% for major program budget lines.
• Time to fulfill public records requests: under state open-records laws (varies by state; e.g., California's Public Records Act, Texas Public Information Act), turnaround time is a measurable, comparable metric across jurisdictions.
• Dashboard usage analytics: page views and repeat visits on public dashboards, a proxy for whether residents and council members actually use published data rather than it existing for compliance only.

In Europe, the equivalent reference point is the European Data Portal's open data maturity reports, which score EU member states annually on data policy, portal quality, impact, and quality dimensions using a similar 0 to 4-ish tiering (methodology summarized at data.europa.eu). Countries like Ireland and France have ranked consistently in the top tier in recent editions (check the current year's report for exact rankings, as they shift annually).

🎬 [VIDEO: "What Works Cities: Data-DrivenData-DrivenAn approach where decisions are systematically informed by data analysis rather than intuition alone.View full definition → Government" - youtube.com/@BloombergCities - case examples of mid-size US cities using data maturity assessments to redesign services]

Key Takeaways

• Benchmark against a matched peer group (population, government structure, fiscal capacity), not against the biggest or most-cited city.
• Use established frameworks (GovEx, Urban Institute, What Works Cities, European Data Portal) rather than inventing an internal rubric from scratch.
• Never present a single composite maturity score to leadership; break it into dimensions (cataloging, governance, staffing, usage) so the actionable gap is visible.
• Pair index scores with hard governance metrics: dataset freshness rate, number of active data-sharing agreements, CDO reporting line.
• Track whether data actually changes decisions (budget citations, service redesigns), not just whether it's published, publication without use is a common false signal of maturity.