Dallas and Fort Worth are two auctions with an hour of freeway between them
Buying the Metroplex as one geography quietly funds clicks from people who will never make the drive to your office.
Set geography to where clients actually come from, not to where you would accept them from. Start with the last hundred files and map them. Most firms find a tight core, a ring that shows up occasionally, and a long tail of people who found them by accident.
Run the core as its own campaign with its own budget. Run the outer ring separately, at lower bids, with an ad that says plainly that consultations can happen by video. A person in Denton County responds to a different promise than someone ten minutes from your door.
Exclude nothing out of pride. If you have never once signed a file from Tarrant County and have no office there, running there is a donation to a firm that does.
Set location targeting to people in your targeted locations rather than people interested in them, unless you deliberately want out of state searchers. The default setting has funded a lot of useless clicks.
Structure the account around who writes the check, not around visa categories
A worried family member and a hiring manager are two buyers, and one account structure cannot serve both.
Family matters are bought at night, on a phone, often by a relative who is not the client. The ad has to answer a fear and offer a fast human. The landing page has to work in one thumb.
Employer work is bought during business hours by someone with a spreadsheet, usually in an office park in Plano, Irving or Richardson. That buyer wants scope, process and a sense of what the engagement costs the company. Same firm, different campaign, different budget, different hours.
Splitting by who pays also fixes reporting. Blended cost per lead across both is a number with no owner and no decision attached to it.
Inside each track, keep ad groups tight enough that the ad can echo the search. Broad match with a loose theme is how an immigration budget disappears in a fortnight.
Employment keywords bring job seekers long before they bring an HR director
The most expensive waste in an immigration account is traffic that is looking for work, a form, or a government office.
Anything shaped like sponsorship attracts candidates hoping a company will hire them. They click, they read, they occasionally fill in the form, and they are never the client. Build the negative list before the campaign goes live rather than after the first invoice.
The other three drains are the same everywhere: people hunting a form number, people checking a case status, and people looking for a government phone line. None of them wants to pay a lawyer today.
Review the search terms report weekly for the first two months and then monthly. Negatives are not a setup task, they are the maintenance.
Watch the free and cheap modifiers too. Someone typing free immigration help is not a bad person, they are just not a buyer, and your budget cannot subsidize the difference.
Local Services Ads and search ads compete for the same click, so decide which one wins
Running both without a plan means paying twice to appear on the same result page.
The screened placements sit above search results and charge per lead rather than per click, which changes the math and the intake burden. Eligibility and rules for legal categories change, so it is worth confirming what applies to your practice before restructuring the budget.
Where both are eligible, the usual approach is to let the screened placement take the broad high intent queries and point search spend at the more specific matter type queries where you can control the landing page.
Lead quality differs. Screened placements tend to send volume with less context, which puts pressure on whoever answers the phone. If intake is one person who also runs the front desk, adding volume is not a strategy.
Whichever you run, dispute bad leads promptly and log the reason. The pattern in those disputes tells you what to exclude next.
A click from McKinney should not land on a page about your downtown office
Landing page mismatch wastes more money than bid strategy ever will.
The page has to repeat the promise of the ad in the first line. Someone who clicked an ad about a work visa for an employee should not arrive on a general practice page listing eight services.
Build a small number of purpose built pages: one per major matter track, each with its own form, its own phone number if you can manage it, and a clear line about how a consultation works for someone who cannot drive in.
Say what the consultation costs and what it includes. A fee stated on the page filters out shoppers and raises the show rate for everyone else.
Keep the form short at the top and let it grow. First contact needs a name, a way to reach them, and a one line description of the situation. The paralegal questions belong on the confirmation step, not the front door.
Feed signed matters back into the account, or it optimizes toward junk
Automated bidding will happily buy more of whatever you told it counts, so tell it the truth.
A form fill is not a conversion. If every submission is fed back as a win, the system learns to find people who like filling in forms, and the account gets cheaper and worse at the same time.
Import the real outcomes. Consultation booked, consultation attended, file opened. Even a monthly manual upload from the CRM changes what the account chases.
Report cost per booked consultation and cost per signed file, split by matter track. Two campaigns with identical cost per lead routinely differ by several multiples once the files are counted.
Pace against your own capacity. An account that produces more consultations than the firm can staff is not a success, it is a queue, and people in a queue call the next firm on the page.
Questions we actually get
- What should we budget to start?
- Enough to gather signal in one clearly defined geography and one matter track rather than a thin spread across everything. We will not quote a cost per click before pulling data for your specific terms and market, because published averages are not your account. A budget too small to collect data is more expensive than no budget at all.
- Should we bid on our own firm name?
- Usually yes, because competitors can appear on it and the clicks are cheap relative to everything else. The counterargument is that you would have earned the click for free through your organic listing, which is sometimes true. Test it by pausing for a few weeks and watching total branded volume rather than arguing about it in theory.
- Can we advertise immigration services on Facebook and Instagram?
- Ads relating to immigration status sit in a restricted category on Meta with tighter targeting rules, and those rules change. Plan for broad audiences and a strong creative rather than fine grained targeting, and have someone confirm the current policy before building the campaign.
- How do we handle a spike after a policy announcement?
- Keep a reserve rather than raising every bid. Volume that arrives after news is a mix of urgent clients and anxious readers, and the ratio is not knowable in advance. Widen budgets on the campaigns already converting, leave the experimental ones alone, and make sure intake is staffed before you buy more calls.
- Should we run ads while our website is being rebuilt?
- Only onto purpose built landing pages that are ready now. Paid traffic pointed at a slow or unclear page is money spent proving the page is the problem. If a rebuild is underway, ship two or three campaign pages first and let the rest follow.
