A Dallas contractor running a fifty-mile radius is bidding on searches in Denton, Waxahachie, Terrell, and Weatherford. Same radius setting, wildly different drive times, competition, and job economics.

That single configuration choice wastes more Texas ad budget than any other. Here is how to run an account properly across a large territory.

Stop using radius targeting

Radius is a circle drawn on a map with no knowledge of highways, traffic, or where your crews actually are.

Target named locations instead. Cities, and where the platform supports it, specific areas within them. A Houston contractor should be targeting Katy, Cypress, Sugar Land, Pearland, and the Heights individually — not a thirty-mile ring that includes places they will never drive.

Use presence-only targeting. By default, location targeting can include people merely showing interest in an area rather than physically located there. For a service business that is usually wasted money.

Apply bid adjustments by zone economics. A job forty-five minutes out is worth less to you than one fifteen minutes out at the same ticket, because the transit cost eats margin. Your bids should reflect that. Most accounts bid identically across a territory where profitability varies by half.

Structure campaigns by zone, not just by service

The standard structure (one campaign per service) breaks in a large Texas territory because it averages performance across areas that behave completely differently.

Split by zone as well. A DFW plumbing account might run North, Mid-Cities, and East as separate campaigns, each with its own budget and its own bid strategy.

Two things become possible. You can see that your north-side cost per booked job is half your east-side cost, which is invisible in a combined campaign. And you can shift budget toward the geography that is actually profitable, which in most Texas accounts is a meaningful reallocation.

Feed it booked jobs, not form fills

The biggest determinant of automated bidding performance is what you count as a conversion.

Most Texas accounts count form fills and calls over thirty seconds. That set includes wrong-zone inquiries, price shoppers, existing customers calling about scheduling, and vendors. The algorithm cannot distinguish them and optimizes toward the whole mixture.

Send back booked jobs with revenue values from your CRM. In Texas this matters more than almost anywhere, because the same keyword set produces both a $200 drain clear and a $30,000 roof replacement, and an algorithm optimizing on undifferentiated lead count treats them identically.

Build the three storm presets

Texas gets three annual surges, and an ad account that reacts to them a week late captures the leftovers.

Freeze preset. Elevated budget, activated when the forecast turns. Copy written for burst pipes and no heat. Statewide rather than zone-constrained, because a freeze hits everywhere.

Hail preset. Elevated budget with geography tightened hard to the affected corridor — the opposite of your normal configuration. Copy written for storm damage and insurance claims. This is the highest-value preset of the three for exterior trades.

Hurricane preset. Two phases. Pre-storm preparation copy in the days before landfall, then post-storm damage copy for the following weeks, with sustained rather than spiked budget.

Each takes an hour to build in the off-season and activates in minutes when needed. Accounts without them consistently spend their surge budget after the surge.

The post-surge window is where money burns

Automated bidding learns from recent history, and Texas demand does not decay gradually, it falls off a cliff.

Two weeks after a hail event, your algorithm has learned that these keywords convert beautifully and will keep bidding aggressively into a market that has already been served. This is the single most predictable source of wasted Texas ad spend.

Set a calendar reminder to review budgets ten days after any storm activation. It is the least sophisticated advice in this article and it saves more money than most of the rest.

Spanish-language campaigns are an actual opportunity

Roughly two in five Texans are Hispanic or Latino, and in large parts of the state Spanish-language search is a substantial and less competitively bid segment.

Running Spanish-language ad groups with genuinely Spanish landing pages (not machine-translated ones) is frequently cheaper per booked job than the English equivalent, because fewer competitors bother.

The prerequisite is that you can actually handle a Spanish-language inquiry when it arrives. Generating leads you discard is worse than not generating them.

Copy that is not interchangeable

Every Texas service ad says fast, reliable, licensed, insured, free estimates. It differentiates nothing.

What works: a real response window, a named service area at the zone level, a concrete guarantee, and honest availability. During a surge, honest lead time in the ad copy actually improves account performance because it filters out people who will not wait, so you stop paying for clicks that were never going to close.

Performance Max with guardrails

Performance Max will spend a local service budget across placements that generate volume and no jobs.

Run it with brand exclusions, account-level negative keyword lists, asset groups split by service line, and a monthly placement report review. Unconstrained, a meaningful share of Texas Performance Max spend routinely lands somewhere useless.

What to measure

Cost per lead is the metric that makes bad accounts look good. Track cost per booked job and return on ad spend using real revenue, then break both out by zone and by month.

By zone, because your territory is not uniformly profitable. By month, because an annual average blends a February freeze week where you could not respond fast enough with a mild October where clicks were cheap and intent was low.

The story in a Texas account is always in the segmentation. The aggregate number tells you almost nothing.