State Legal Information Property

Owning the Channel: A State-Level Consumer Information Property

Families researching care options for a relative are among the most motivated researchers online, and among the worst served. The information they need is public — inspection findings, staffing levels, quality ratings, ownership — and it is published by a federal agency in a form nobody outside the industry can practically use. Meanwhile the commercial properties covering the subject are thin, advertising-heavy and largely uninformative.

That gap is a lead generation opportunity in the specific sense that matters: real demand, poor incumbents, and a data source that is free to use and expensive to use well.

The problem

Three constraints shaped the build. The data was public but difficult, published across several files with inconsistent keys and changing structures between releases. The vertical was regulated, because a property that connects consumers to attorneys is advertising and is subject to advertising rules. And the property had to be built from nothing, on a new domain, with no budget for paid traffic to bridge the indexing months.

Approach

The decision that shaped everything else was to lead with the data rather than with content. Most sites in this space write articles. This one publishes structured facility-level information derived from the federal dataset, which produces pages that are genuinely unique, genuinely useful, and impossible to replicate without doing the same engineering work.

The second decision was to treat the advertising structure as a design input rather than a launch task. Disclosure placement, advertiser identification and the absence of any attorney-client relationship were built into the templates from the first page, which meant the property could grow without the compliance layer drifting.

The third was to publish in stages. Coverage expanded by area rather than all at once, with index coverage watched at each stage and page types that failed to earn impressions revised or removed rather than left standing.

Build detail

• Federal open data ingested, cleaned and normalized with Python and pandas, then loaded into a PostgreSQL schema modelled on the entity relationships rather than on the page templates.
• Facility locations geocoded into PostGIS so proximity search runs in the database rather than in application code.
• Programmatic page generation at state, area and facility level, each page carrying real data rather than templated text with a place name substituted.
• WordPress with a custom theme and no page builder, with schema markup, disclosure blocks and internal linking implemented at template level.
• Enquiry routing built into the property, with tracking that attributes each enquiry to the page that produced it.
• A refresh path so a new dataset release can be ingested without hand-editing anything.

Results

[Pull each of these from Search Console and GA4 immediately before publication. Publish only figures you can evidence.]

• Indexed pages: [GSC coverage figure]
• Organic impressions, most recent full quarter: [GSC]
• Organic sessions, most recent full month: [GA4]
• Enquiry volume and routing: [describe in relative terms if absolute numbers are commercially sensitive]
• Data scale: facility records published and refreshed on the federal release cycle
• Performance: [Core Web Vitals figures for the facility template on mobile]

What I would do differently

Two things. The first is that the conversion path was built after the content rather than alongside it, which meant several months of traffic arriving at a property that was good at informing people and mediocre at helping them act. Building the enquiry path first would have cost nothing extra and would have started producing earlier.

The second is that I published broader coverage than the early data justified. Some page types were generated at scale before there was evidence that anyone searched for them, and pruning those later took longer than building them. The staged approach I now use on client work came directly from getting this wrong on my own.

The point of this project

It is the reason I can talk about owned lead generation without theorizing. I have carried the indexing months, the compliance constraints and the second-year operating cost on my own balance sheet. That is a different kind of knowledge from having built one for someone else.