Three buys, not one metro: the City, the county, and across the Mississippi
Split the account by geography first, because the offer, the ad copy and the landing page all change when the reader lives in a different state.
A campaign for the City of St. Louis, a campaign for St. Louis County and St. Charles County, and a campaign for the Metro East. Three separate budgets, three sets of ads, three landing pages. Anything less and you cannot see which of the three is paying for itself.
The split is not decorative. A reader in Edwardsville needs to know whether you take Illinois clients and what happens if a related state matter comes up. A reader in Kirkwood does not care about any of that. Same firm, different conversation, so the ad has to be different.
Check your location settings while you are in there. Most accounts run on the default that reaches people merely showing interest in an area, which is how a firm ends up paying for clicks from another country. Presence based targeting is usually the setting you want, and it is worth verifying in every campaign rather than assuming.
Budget the three separately, and be willing to switch one off. A campaign that cannot be evaluated alone is a campaign that will quietly subsidize failure for a year.
Both Springfields will spend your budget before anyone in Kirkwood calls
Missouri and Illinois each have a Springfield, the region is full of place names that repeat, and loose targeting finds all of them.
Geographic negatives matter more here than in a single state metro. Exclude the places you do not serve by name, then watch the location report weekly for the first month. It fills up with towns you have never driven to.
Keyword negatives do the heavier work. Free legal help, government agency names, form downloads, status checks, job seekers looking for a sponsor, students writing papers. All of it looks like immigration intent to a match algorithm and none of it retains a lawyer.
Build the list before launch, then treat it as a standing weekly job. Pull the search terms report, mark what did not belong, and add it. A neglected account bleeds gradually, which is why it goes unnoticed until a quarter is gone.
Watch for one local trap in particular. Searches about county offices, city services and utility accounts pull in traffic that has nothing to do with legal work, and generic near me phrasing picks up a lot of it.
The keyword list should read like the matters you want more of
Bid on the work you actually staff and want to repeat, not on the whole category.
Start from your own file list. Which case types were profitable, which ones you have capacity for, which ones you would take ten more of tomorrow. Those become campaigns. Everything else stays out until you decide otherwise.
Intent tiers matter more than volume. Someone typing a lawyer phrase with a case type is closer to retaining than someone typing a category name alone, and much closer than someone typing a form number. Bid accordingly and keep the tiers separated so you can see the difference in cost.
Employer side work is a different animal and probably a different account. The person searching is doing it for the company, on a work computer, at a different hour, and the sale takes weeks. Mixing it with consumer campaigns hides both.
Resist the urge to add volume when results are slow. Wider keywords in a category this expensive is usually how a modest problem becomes a large one.
Write the disqualifier into the ad, because a wasted consultation costs more than a wasted click
Ad copy that filters costs you clicks and saves you attorney hours, which is the trade you want.
Say what you do and what you do not. Name the case types, name the states you are admitted in, say whether the consultation carries a fee. Every one of those lines reduces click volume and improves what arrives.
The arithmetic is one sided. A click costs what it costs. A consultation that should never have been booked costs a slot on a lawyer's calendar plus the staff time to arrange it plus the time to unwind it. In a practice where consultation capacity is the real constraint, filtering early is the whole game.
Match the landing page to the promise word for word. If the ad names a case type, the page headline should name the same case type. Mismatched pages are the most common reason a well built account underperforms.
Give the phone a fair chance. Call extensions, call only ads during staffed hours, and a number that rings somewhere a human picks up. Buying clicks at legal category prices and sending them to voicemail is an expensive way to fund your competitors.
Automated campaigns will find the cheapest clicks in the metro, and you do not want them
Broad automated campaign types optimize toward whatever converts fastest, which in this category is rarely a client.
Give an algorithm a form fill as its goal and it will find you form fills. Some will be from outside your service area, some will be from people looking for free help, and a few will be from people who are not going to retain anyone. The dashboard will look excellent.
Guardrails are the answer, not abstinence. Feed the account real outcomes: consultation attended, matter opened, fee agreement signed. Offline conversion import from your case management system takes an afternoon to set up and changes what the machine learns from.
Keep exact and phrase coverage on your core terms in a campaign you control. Let automation work at the edges once you have outcome data flowing, not before.
Listen to the calls. An account whose recordings nobody reviews is an account nobody can actually fix, and in this category the recording tells you more in ten minutes than the reports tell you in a month.
A thin budget spread across two states clears nothing
Legal search has a price floor, and splitting a small budget across three geographies leaves each one too quiet to learn from.
If the total will not support three campaigns at a spend that produces regular clicks, run one. Pick the geography where your office sits and your admissions are cleanest, win there, then expand. A budget stretched over the City, the county and the Metro East at once often produces three campaigns with no signal.
Pace against the calendar of the person doing consultations. Spending into a week where nobody can seat a new client buys inquiries that go cold before anyone calls back.
Report on cost per matter opened, with the state and the case type attached. Cost per lead flatters everything, and in this practice it flatters the wrong things. The number that matters arrives late, so expect to read paid performance on a lag and hold the line when a monthly report looks thin.
Nobody can promise a cost per case or a volume. What can be committed to is structure, a negative list that gets maintained, outcome tracking that reaches back to signed files, and the discipline to turn something off.
Questions we actually get
- What should we budget to make search ads work here?
- There is no published figure to quote you, and legal keyword prices move. The practical test is whether the budget supports enough clicks per week in one geography to learn something within a couple of months. If it does not, run a single tighter campaign rather than three underfunded ones across the metro.
- Should we advertise on Meta or LinkedIn as well as search?
- Possibly, but check current platform policy first. Ad platforms apply extra restrictions to categories touching immigration status, and the rules change. Search captures people already looking. Social buys attention you have to earn, which suits employer side work better than urgent consumer matters.
- Do we need a separate campaign for the Illinois side?
- If you take Metro East clients, yes. The ad, the landing page and the offer all change once the reader is in Madison or St. Clair county, and mixing them into a Missouri campaign hides both performance and cost. Keeping them separate also lets you switch one off without disturbing the other.
- Should the consultation be free or paid in the ads?
- It depends on your capacity and your case mix, and both choices work. What matters is deciding before launch and saying it in the ad. A fee filters, which protects the calendar, and it also reduces volume. Changing it midstream makes every number before and after incomparable.
- What should the monthly report actually show?
- Spend, consultations attended and matters opened, broken out by geography and case type, with search terms and call recordings available on request. Cost per lead alone tells you almost nothing in this category, and any report that leads with impressions is describing activity rather than results.
