SEO for lawyers: 15 law firms, measured

The numbers, before the explanations

Thirty-two points between the how and the what

The SEO median is 67, the GEO one 41. These are sites built well from a technical standpoint and empty from a machine's standpoint: the firm's address is never in the structured data (100% of cases), and the pages talk about «practice areas» instead of the problems they solve.

The flaws, in order of how many sites have them

• 100%

  No date showing how up to date the page isMedium

  An assistant choosing which source to cite also looks at how fresh it is. «Updated October 2026» written on the page, and declared in the structured data, shows that someone looks after it. Without it, the page could be five years old, and with equal content the one with the date wins.

• 100%

  The address is not in the structured dataHigh

  Out of fifteen firms, fifteen. It is written in the contacts, where a person reads it; it is not in JSON-LD, where Google reads it and an assistant reads it. For a profession chosen by city, it is the most important piece of data that is missing.

• 100%

  No coordinatesLow

  Here too, fifteen out of fifteen. Latitude and longitude cost two lines and remove any doubt about where you are, which for a firm on a central street with a common name is no small thing.

• 100%

  Numbers, timeframes and prices are missingMedium

  An assistant gladly reports what can be verified: durations, hours, prices or price ranges, years in practice. Phrases like «serving clients for years» give nothing to cite.

• 93%

  The business profile in the structured data is incompleteMedium

  Name, address, phone, hours and official profiles (Google, LinkedIn, Wikipedia if they exist) read by a machine in a single block. When a piece is missing, the assistant doesn't know whether you are the same business it finds elsewhere and when in doubt it doesn't name you.

• 93%

  The site doesn't answer the real questionsLow

  «How much does an uncontested divorce cost?» «How long do I have to challenge a dismissal?» These are the questions people ask an assistant. If there is no sentence on the site that resembles them, there is nothing to cite.

• 93%

  No business schemaHigh

  Schema.org has LegalService and Attorney: types made for exactly this, holding the area of expertise, the address and the hours. Declaring them is the way to tell a machine what your profession is without making it infer it.

• 93%

  Headings don't resemble clients' questionsMedium

  Whoever asks an assistant writes a full question. A heading written as a question, with the answer right below it, is the easiest piece to cite: «How long does the first consultation last?» works better than «Information».

The three things to start with

1. 1
   Put address and practice area in the structured data. Don't remove them from the contacts: add them in JSON-LD too, with the LegalService schema. It is the fix that earns the most points for the time it takes, and none of the fifteen firms measured has it.
2. 2
   Rewrite the practice areas as problems. A page for «dismissal» that starts with the question the client asks, not with the statute. The expertise stays the same, what changes is who manages to find it.
3. 3
   Sign what you write. Name, photo, bar association, years in practice. For a model it is the difference between an attributable text and an anonymous one.

How we measured

• How many sites15 on the list, 15 reached.
• How many pagesUp to 5 pages per site, starting from the home page.
• With whatThe same function that analyzes your site when you press «Analyze for free». No hand-assigned scores.
• How they were chosenBy searching Google for «law firm» and «lawyer» together with the names of five Italian cities and four practice areas, and taking the firms' sites, not the directories. They are therefore among those Google already shows.
• What we do not publishThe names. Measuring a public site is what the product does every day; publishing the report card with the name on it is something else.

Fifteen sites are a small sample, and it is worth saying so: the numbers above describe these fifteen, not the forty thousand Italian lawyers. They serve to show that the flaws repeat, not to produce a national statistic.

Questions