Exact match on a borough plus a matter is the only affordable entry point
Broad phrases in this market are priced by national advertisers with far more budget than a single firm, so the winnable inventory is narrower and more specific.
The broad divorce and custody terms in the five boroughs sit at the top end of what legal search costs anywhere. Bidding into them with a modest budget buys a handful of clicks a day, most of them unqualified, and starves everything else.
Start with phrases that carry both a matter type and a place. Child custody attorney plus a borough. Contested divorce plus a neighborhood. Order of protection plus a county name. Tighter match types cost more per click on paper and less per consultation in practice.
Add the circumstance terms that broad campaigns never surface. High asset, business owner, apartment, relocation. People type these when they already know their situation is not simple, and they are the clicks worth paying for.
Google also screens legal advertisers, and verification can hold up a new account. Start that process before you plan a launch date rather than after.
Your license stops at the Hudson, and your geo targeting should say so
The metro spills into New Jersey and Connecticut, and paying for clicks you cannot serve is the most common leak in a New York legal account.
Radius targeting drawn from a Manhattan office reaches across the river within a few miles. Those clicks look local in the interface and are worthless to you unless the firm is admitted there.
Set targeting to people in or regularly in your locations rather than the default that includes people showing interest in them. Otherwise you fund searches from anywhere in the world that mention New York, and there are a lot of them.
Decide separately about Nassau, Westchester and Long Island City style commuter geography. Plenty of households live in one county and work in another, so the person searching from Midtown at lunch may be filing in Nassau. If you take that work, target it deliberately and write ads that say so. If you do not, exclude it.
Then check the location report monthly. It is the single fastest place to find money going somewhere you did not intend.
Half of this auction is people who will never hire an attorney
Negative keywords do more for a family law account in this city than bid adjustments ever will.
A large share of divorce searching is procedural. People looking for forms, for the clerk, for filing fees, for how to do it themselves, for free help. Every one of those clicks is billed at attorney prices.
Build the negative list before the campaign goes live and keep adding to it weekly for the first two months. Free, forms, DIY, online, cheap, template, pro bono, jobs, salary, paralegal, and the name of every online divorce brand advertising here.
Watch for the ambiguous city names too. Queens, Chelsea and Harlem all exist elsewhere, and broad or phrase match will happily bring you traffic from other states or other countries.
Read the search terms report as a document, not a dashboard. Twenty minutes with it each week will do more for cost per consultation than any bidding change you could make in the same period.
In a one bedroom walk-up there is nowhere private to make the call
Spend should be timed to the hours when a person can talk without being overheard, and in this city that is rarely during a normal business day at home.
A parent in a small apartment with a spouse in the next room cannot call a divorce attorney at seven in the evening. The call happens at a desk at lunch, in a stairwell, on a platform, or after everyone else is asleep.
Look at your own hourly data before assuming anything. Most firms find clusters they were not funding, and daytime hours they were overfunding out of habit.
Only fund the hours somebody actually responds. An ad running at eleven at night is fine if the inquiry gets a reply first thing, and wasted if the form sits until Thursday.
Call extensions and call ads earn their place here, with one caveat: they are graded on whether a human picks up. If reception is stretched, send the click to a page instead of to a ringing phone.
A Park Avenue asset case and a support petition cannot share one bid
Matters with wildly different values need separate campaigns with separate budgets, because a shared account will always spend toward the cheaper click.
A complex distribution matter involving a business, a co-op and a house upstate is worth a different amount of acquisition spend than a straightforward support petition. Put them in one campaign and the algorithm optimizes toward whichever produces more form fills, which is usually not the one you want.
Split the account by matter type first, then by geography inside it. Custody and visitation, contested divorce and asset matters, support and modification, and protective orders each behave differently and deserve their own budgets and their own ad copy.
Write the ads to sort people. Naming the borough and the matter in the headline costs you clicks from everyone else, which is the point. A narrower ad with a lower click volume and a higher consultation rate is the better outcome.
Be careful with remarketing on family law. Following someone around the internet with a divorce ad can appear on a shared tablet or a family laptop. Keep it off entirely for the sensitive matter types.
The retainer signs weeks after the click, so the account always reports late
The lag between an inquiry and a signed matter means your platform data is describing a decision that has not finished yet.
A form fill is not a result. The path runs inquiry, contact, consultation booked, consultation attended, retainer signed, and the gap between the first and the last can be weeks or longer.
Import the later stages back into the account. Feeding signed matters in as offline conversions, even with a delay, is what stops the bidding from chasing cheap and useless form fills.
Until that is running, judge the account on consultations attended, broken out by borough and by matter type. It is imperfect and it is far better than cost per lead.
Expect a month where spend looks bad because the matters it produced have not closed yet. Deciding in advance how long to wait before judging a change prevents the panic that kills accounts right before they work.
Questions we actually get
- What monthly budget does this actually require?
- It depends on how many matter types and boroughs you want to run, and any specific figure quoted without seeing your auction data is invented. The useful test is different: work out what a signed matter is worth to the firm and how many consultations it typically takes to sign one. If the arithmetic does not support the click prices you are seeing, run fewer campaigns properly rather than all of them thinly.
- Should we run Local Services Ads as well as search?
- They operate under separate rules and separate verification, and they are billed on a different basis. Some firms find them worthwhile alongside search, others do not. Treat them as a distinct channel with its own measurement rather than as an extension of the search account.
- Is paid social worth anything for family law here?
- It reaches people before they search, which suits modification, relocation and post-judgment work more than an active new divorce. The risk is delivery onto shared devices in a shared apartment. If you run it, keep the creative neutral, avoid remarketing lists built from sensitive pages, and never let the audience settings imply you know something about the viewer.
- How aggressively should we exclude the suburbs?
- That is a practice decision, not an advertising one. If you take Nassau or Westchester matters, target them on purpose with their own campaigns and ad copy. If you do not, exclude them, because they are close enough to consume a meaningful share of an untended budget.
- How long before we should judge the account?
- Long enough for matters to close, which is longer than most people are comfortable with. Set the review window in advance, keep the structure stable inside it, and change one significant thing at a time so you can tell what caused what.
