If you build or maintain sites for criminal defense firms, someone will eventually ask you to "run an A/B test" on a charge page. Before you pick a tool, check whether the site can measure the thing the firm cares about. On most defense sites that is a phone call from a mobile visitor, and a surprising number of sites record it poorly or not at all.
Testing without reliable conversion data produces confident charts about nothing. This post covers the measurement groundwork that should be in place first, with notes on the constraints that make legal sites different from ecommerce.
Define the conversion with the firm, in writing
Developers often default to whatever the analytics tool tracks automatically. For a defense firm that is the wrong starting point. Sit down with whoever runs intake and agree on what counts:
- A tap on a tel: link from a mobile device
- A completed contact or case review form
- A chat conversation that reaches a human, if the firm uses chat
- A call to a tracking number shown on the site
Then agree on which of those count as primary. Many firms care far more about calls than forms. Write the definitions down, because six months later nobody will remember whether chat openings were included.
Track tel: link taps and form submissions as events
Google Optimize was shut down on September 30, 2023, so many teams now pair GA4 with a separate testing tool or run server-side experiments. Whatever you choose, the conversion events have to exist in analytics first.
GA4 has a recommended event called generate_lead for form submissions. Phone taps need a custom event. A minimal version looks like this:
document.addEventListener("click", (e) => {
const a = e.target.closest('a[href^="tel:"]');
if (a) gtag("event", "phone_tap", { link_location: a.dataset.loc || "body" });
});
// after the form endpoint confirms success, not on button click
gtag("event", "generate_lead", { form_id: "case_review" });
Fire the form event only after the server confirms the submission. Firing on button click counts failed validations and double taps. Add a data attribute to each phone link so you can tell header taps from in-content taps, which matters when the test changes where the number appears.
Never send what the visitor typed
Google Analytics policies prohibit sending personally identifiable information, and a defense intake form collects some of the most sensitive data a website can hold: names, phone numbers, charges, arrest dates. None of it should reach analytics, a heat map tool or a session recorder.
Check a few things before any test goes live. Form fields should be excluded or masked in any recording tool. URLs should not carry form values as query strings after submission. Event parameters should describe the form, never its contents. If the firm uses a third-party form service, confirm where submissions are stored and who can see them.
Connect site events to calls the firm actually received
A phone_tap event shows intent. It does not prove a call connected. Call tracking with dynamic number insertion can close that gap by showing a different number per source and logging calls on the provider's side. If you use it, keep the firm's main number consistent in the footer, schema markup and Google Business Profile so local search signals stay clean.
The richest data lives in intake. A short field asking how the caller found the firm, plus a weekly export, lets you compare test variants on signed cases rather than raw contacts. For the form side of that work, designing an intake form a frightened caller will finish covers how many fields to ask for and how to word labels for a scared person typing on a phone.
Size the test honestly
Most defense charge pages see modest traffic and few conversions. Run the numbers before you promise a result. If a page produces twenty calls a month, a split test chasing a ten percent lift will take far longer than anyone will wait.
With small numbers, test large changes to the first mobile screen, run variants for fixed periods and keep a change log in the repo. Before and after comparisons are weaker evidence than a true split, but they are often the only realistic option, and they are fine if you say so in the report.
Keep every variant compliant
Each variant is attorney advertising. If a test changes headlines, claims or disclaimers, the firm's lawyers need to approve it, and state bar advertising rules apply to every version a visitor might see. Build the approval step into your deployment process so a variant cannot ship by accident.
More on testing ideas for legal sites is collected in the conversion testing library for law firm websites.
Questions about conversion tracking on law firm sites
Should phone taps count as conversions in GA4?
Mark them as key events if the firm treats calls as its main goal. Report them separately from forms so changes in one do not hide changes in the other.
Is a session recording tool safe on a defense site?
Only with strict masking of every input and a review of where recordings are stored. Many firms decide the risk is not worth it.
Which testing tool should I use?
Pick one that works with your stack and supports server-side or flicker-free variants. The tool matters less than clean conversion data and an honest sample size.
SBD Marketing works only with criminal defense firms, one firm per city. It builds websites for defense practices, handles SEO for the charges a firm defends, works on visibility in AI search tools and runs SEO audits.
Top comments (0)