The Client Research You Cannot Track: Referrals, Reviews, and Repeat Searches

Some of the most important events in a legal hiring decision never enter the firm’s analytics.

A friend names an attorney over dinner. A business owner forwards a biography to a partner. A prospect reads reviews, closes the browser, asks another lawyer for an opinion, and searches the firm’s name days later. The eventual report credits paid search because that was the observable click before the call.

Paid search may deserve credit for helping the person return. It still does not explain why the person searched for that firm.

The practical response is to maintain two forms of evidence at once: instrument the interactions the firm can observe, and run a lightweight research habit for the influences it cannot. This is not permission to assign imagined credit to “brand.” It is a way to learn what prospective clients check, whom they trust, which doubts delay contact, and how the firm’s public presence helps or hurts a recommendation.

See the hiring journey in three layers

Use three separate views:

Do not force staff to choose which layer is “correct.” A referral and a paid click can both be true. An untracked article and a direct visit can both matter. The error is turning the observed layer into a complete causal story.

JD-005’s attribution framework handles the data architecture. This article focuses on the research discipline around the part that remains unobserved.

Start with the moments the dashboard cannot answer

Write down the business questions that attribution leaves open:

• Why did a referred prospect trust or reject the recommendation?
• Which reviews, biographies, case explanations, news mentions, or videos did the person check?
• Did a repeat branded search represent growing confidence, unresolved doubt, or simple navigation?
• Which words did the referral source use to describe the firm?
• What made the prospect contact this firm now rather than continue researching?
• Which expected proof was missing?
• Did a partner, family member, accountant, doctor, former lawyer, or colleague influence the choice?
• What did a prospect see that created the wrong expectation for intake?

Those are research questions, not fields to infer from a last-click report.

Ask without turning intake into a survey

Connect those bounded questions to intake metrics that preserve inquiry and retention outcomes; do not turn remembered influence into a forced source field.

The person’s legal need comes first. Intake can usually capture useful context with two short prompts when the moment is appropriate:

1. “How did you first hear about the firm?”
2. “Was there anything you saw or heard that helped you decide to contact us?”

If the first answer is “Google,” one neutral clarification can help: “Was Google where you first found us, or were you looking up a name someone had already mentioned?”

Allow multiple answers. Include not asked, does not remember, and declined so staff do not guess. Preserve a short verbatim note where the person offers something specific, then classify it separately.

Useful categories might include former client, attorney referral, professional adviser, personal contact, review, search result, attorney biography, article/video/podcast, media coverage, social post, event, sponsorship, directory, prior relationship, and unknown. Do not make the list so long that it overwhelms the legal intake.

The field must not determine whether the inquiry receives attention. It exists to understand the journey, not to reward the channel someone happened to name.

Interview a small, purposeful sample every month

Analytics is continuous. Qualitative client research can be small and still useful if the sample and limits are visible.

Select a varied set from a recent mature cohort:

• retained clients from different matter types and sources;
• qualified prospects who chose another firm, where follow-up is appropriate;
• referred prospects who did and did not engage;
• people whose observed source conflicts with their self-reported influence; and
• longer-decision matters with repeated visits or searches.

A five-to-eight-person monthly sample will not estimate population percentages. It can reveal repeated questions, contradictions, and language worth investigating in larger data. Record why each person was selected so the team does not present a convenient sample as representative.

Use a 15-minute interview guide:

1. What was happening when you first decided you might need legal help?
2. Who or what did you consult first?
3. When did our firm first enter the process?
4. What did you look at before contacting us?
5. What made you hesitate or continue?
6. Did anyone else take part in the decision?
7. What did you expect to happen after you contacted the firm?
8. What nearly caused you to choose another path?

Avoid asking the participant to validate a marketing theory. “Did our reviews convince you?” invites agreement. “What did you look at?” gives the person room to contradict the team’s assumption.

Juris Digital’s earlier focus-group research on how people find a lawyer used 25 people in one market and found reported routes that included referrals, search, work connections, and advertising among participants who had previously hired an attorney. The bounded study is useful for hearing how people describe their choices. It is not a population estimate for all legal consumers or practices.

Turn stories into a pattern table

After the interviews, code each observation without stripping away context.

These rows are hypothetical examples. “Three of seven” describes only that small sample. It is a clue, not a conversion rate.

The last column matters most. Research should produce a next check, not a confident story from a memorable quote.

Understand what each hidden influence is doing

Use the dark-social guide for untracked sharing, the reviews, PR, and social-proof framework for public trust signals, and the attorney thought-leadership guide for expert content that may shape a later decision.

Referrals transfer a claim about fit

A referral is more than a source name. The referring person tells a story: this lawyer handles matters like yours, responds well, knows this industry, tries cases, explains things, or treats clients a certain way.

Research the claim being transferred. If accountants refer business owners because an attorney gives pragmatic advice, the website and intake script should help that person recognize the same strength. If referral partners repeatedly send work the firm does not accept, the problem may be the referral explanation rather than the advertising.

Track the specific referral source where appropriate, but also ask what they said. A name without the promise misses the useful part.

Reviews answer risk questions

Prospects may use reviews to assess responsiveness, respect, communication, process, or whether the firm has handled a recognizable situation. Star averages alone do not reveal which uncertainty a review resolved.

Group themes in public feedback and client interviews. Separate service feedback from legal outcomes. A pattern about missed updates calls for an operating response before a reputation campaign. A pattern of prospects seeking clarity about fees or next steps calls for clearer expectations in public content and intake.

Maintain an appropriate process for requesting honest feedback and responding in line with confidentiality and professional obligations. Do not write or suggest the desired sentiment.

Repeat searches show unresolved or returning demand

A branded search can mean a prospect remembered the firm from an ad, referral, article, billboard, event, or prior relationship. It can also be simple navigation by a current client, employee, job candidate, or vendor.

Separate branded and nonbranded search where the available data permits. Compare campaign timing, geography, page destinations, self-reported influence, and inquiry outcomes. Treat a trend as a signal with several explanations, not proof that one awareness channel created demand.

Paid branded search deserves the same nuance. It can protect a clear path back to the firm and provide useful message control. Its efficient conversions may also include people whose preference was created elsewhere. Evaluate that role explicitly when coordinating SEO and PPC.

Use research to repair the public decision path

The goal is not a prettier journey map. It is a more coherent experience for the next prospect.

For every repeated finding, identify the public or operational surface it affects:

• referral-partner explanation;
• attorney biography;
• practice or industry page;
• reviews and responses;
• search ad and branded result;
• article, video, podcast, or media mention;
• contact page and form;
• intake greeting and qualification language;
• consultation preparation; or
• follow-up after a missed contact.

Then write a testable change. “Build trust” is too vague. “Add the attorney’s role in the first consultation to the biography because four interviewees looked for who would actually speak with them” identifies the observation, surface, and change.

The law firm client journey in AI-assisted research adds another surface to this work. A prospect may use an answer engine to frame the legal problem or generate a shortlist without leaving a visit the firm can observe. The same rule applies: improve the public evidence, then ask real prospects what helped rather than inventing attribution.

Give untracked influence a bounded budget

“People saw us everywhere” is not a business case. For any awareness, sponsorship, thought-leadership, or referral-support investment, state:

• the audience and market;
• the belief or behavior the work is intended to influence;
• the assets and delivery purchased;
• the direct signals available;
• the qualitative questions the firm will ask;
• the observation window;
• other activity that could explain changes;
• total spend and renewal date; and
• the condition for scaling, changing, or stopping.

Where volume and design permit, a phased rollout, geographic comparison, or holdout may provide stronger evidence. Differences in markets, seasonality, competitor activity, and small samples can still limit the conclusion. Record those limits.

Untrackable does not mean unaccountable. It means the decision uses several imperfect forms of evidence.

Bring a contradiction to the review

The best monthly research review starts with something the current report cannot explain:

• Paid branded search receives the credit, but five retained clients name referrals.
• Review volume rises, while interviewees keep saying attorney biographies feel unclear.
• A sponsorship shows no tracked clients, but referral partners repeat its message.
• Repeat visitors grow, but the same unanswered fee question appears in intake notes.

Choose one contradiction. Compare observed interactions, reported influences, and business outcomes. Decide which public surface or research question to change next.

Juris Digital’s law firm PPC and paid media work connects campaign results with landing pages, intake feedback, retained clients, and broader channel decisions. If last-click reporting is obscuring the journey, bring an anonymized cohort that includes observed source, self-reported influence, qualification, and retention.

We can help separate what the campaign report shows from what the client research suggests, then identify the next measurable repair. That is a more useful objective than promising to track every influence a prospective client encounters.

Explore Partnership >

Casey Meraz Casey Meraz is an entrepreneur, SEO expert, investor, creator, husband, father, friend, and CEO of Juris Digital. Casey is a frequent speaker at industry events and the author of two books on digital marketing, including "Local Marketing for Personal Injury Lawyers" and “How to Perform the Ultimate Local SEO Audit”