Before you spend another dollar on ads, answer one question: what percentage of the leads you already paid for did a human being actually talk to?
Most firm owners do not know. The ones who go and find out are usually unhappy with the answer. And that number, not your cost per lead, is where the next batch of signed cases is sitting.
Why this is the highest-return thing to fix
A person filling out a legal intake form at 9pm on their phone is not filling out one form. They are working through a page of search results, and they will talk to whoever reaches them first with a real conversation.
That means your competition is not the firm with better ads. It is the firm that called back in four minutes.
Everything else in marketing is expensive and slow to change. Creative testing takes weeks. New channels take months. Response time can change tomorrow afternoon and costs nothing, which is why it should be first every single time.
Measure it honestly
Two numbers, and the difference between them is the whole story.
Attempted contact rate: the percentage of leads where somebody tried. This is the number your intake team will quote you, and it is usually fine.
Actual conversation rate: the percentage where a human had a real conversation with the person. This is the number that predicts signed cases, and it is usually much lower.
Also measure median time from lead received to first attempt, not average. One lead contacted three days later drags an average into meaninglessness while the median tells you what normally happens.
Pull last month’s leads and compute all three. If your systems cannot tell you, that is finding number one, and it is worth fixing before anything else. This belongs on the firm dashboard permanently, not in a one-time audit.
Where the leads actually get lost
After hours, which for most consumer practices is where a large share of inquiries arrive. Evening and weekend leads go into a queue and get called Monday, by which time the person has retained someone else.
Second and third attempts. Most firms attempt once, mark it unreachable, and move on. The person who did not answer an unknown number at 2pm is not uninterested, they are at work.
Channel mismatch. Calling a lead who contacted you by web form, and only calling, misses everyone who does not answer unknown numbers, which is most people under 40.
Handoffs. The lead arrives in one system, someone has to see it in another, and the gap between those is measured in hours.
What to automate
Immediate acknowledgment, within a minute, by text. It confirms you received the inquiry, sets expectations for when a human will call, and asks one or two qualifying questions. This alone changes outcomes, because the person stops shopping.
Be direct that it is automated. “This is an automated message from our intake team, someone will call you shortly” is honest and works. Pretending a bot is a person is a bad trade for a law firm and it reads as fake anyway.
Routing and alerting. New lead fires a notification to whoever is on intake, in the tool they already have open. Zapier handles this in an afternoon.
Persistent follow-up. A defined sequence across a few days with a mix of call and text, so that the second and third attempts happen because the system schedules them, not because someone remembered.
After hours, voice AI can carry more of the conversation, and the guidance in AI intake for law firms covers how far to let it go before a human takes over.
What to keep human
Qualification decisions on anything borderline. Automated screening should sort the obvious cases and route everything else to a person, because a wrongly rejected lead is invisible. You never learn about the case you turned away.
The actual retainer conversation. People hire lawyers they have spoken with.
Anything emotional. Someone describing a serious injury or a family matter should reach a human quickly, and an automated sequence that keeps texting them is worse than nothing.
Do this today
Pull every lead from last week. For each one, find the timestamp it arrived and the timestamp of the first genuine conversation. Compute the median gap and the percentage with no conversation at all.
That single number tells you whether your problem is lead volume or lead handling. Most firms discover they are buying leads to replace the ones they already lost, which is an expensive way to solve a scheduling problem.