A burst riser at midnight and a planned bathroom gut are two accounts
Emergency demand and planned work behave so differently that letting them share a budget hides the performance of both.
Emergency intent converts on speed. The searcher wants someone answering now, tolerates a higher price, and decides within minutes. That traffic can justify a click cost that would look absurd on a spreadsheet, provided the phone is actually answered.
Planned work behaves nothing like that. A repipe, a bathroom gut, a riser replacement or gas line work involves comparison, sometimes a board, and a decision that lands weeks later. The same click cost applied to that traffic looks like a disaster until the work closes.
Put them in one campaign and the emergency keywords consume the daily budget before lunch, while the planned work never accumulates enough impressions to tell you anything.
Separate budgets, separate landing pages, separate definitions of what counts as a conversion. Then the reporting stops lying to you.
Long Island City and Forest Hills belong on separate lines
Geographic targeting here is not a radius, it is a list of places you can get a truck into and back out of.
A radius drawn around a yard crosses water, tolls and neighborhoods you have no intention of serving. It also buys clicks from people who will never be dispatched to.
Target by borough and then by the specific neighborhoods you actually work. Nassau County and Westchester are separate business decisions, not overflow from a circle on a map. If a Staten Island job means a long crossing for a two hour visit, decide that on purpose rather than by default.
Time of day belongs in the same conversation. An outbound run in the late afternoon can cost an hour that a morning run does not, and bid adjustments are the right place to express that.
Report geography separately. Averaging Manhattan and the Bronx into a single cost per job hides which one is carrying the account and which one is being subsidized.
Parking, freight and stair carry belong in the bid, not just the invoice
The real cost of a New York job includes hours nobody bills for, and your bidding should reflect where those hours pile up.
No driveways. Restricted street parking. Walk-ups with narrow stairwells. Freight elevators booked in advance and released on somebody else's schedule. Equipment that has to be carried when it will not fit in the lift.
A water heater into a fifth floor walk-up is not the same job as the identical heater in a suburban garage, and the difference is not small. The gross margin on a job varies by building type before anyone touches a wrench.
So the ceiling you can afford to pay for a click varies by neighborhood and building type too. Use that. Bid up where the work is reachable and profitable, bid down where it is neither.
Say it on the landing page as well. Pricing that looks high against a national average is not high here, it reflects a different job. Explaining that plainly costs you a few price shoppers and saves your dispatcher an hour a day.
Steam heat searches wake up in October and go quiet in April
A steam and radiator season exists in this city that national ad playbooks have no concept of.
Prewar buildings running on steam produce their own demand curve. No heat calls, banging radiators, failed traps, boiler service, and owners who have finally had enough of the top floor being cold.
Budget should move ahead of the season rather than after the calls start. Shifting spend in November because October looked slow is the wrong order of operations.
Hold reserve for the first genuinely cold week. Demand does not arrive on a smooth line, and an account that hits its daily cap by nine in the morning on the day it matters most has failed at the only moment it was being tested.
Watch your own numbers rather than a generic seasonality chart. Every book of business skews differently depending on which boroughs and which building types you serve.
Renters, supers hunting parts and license lookups are clicks to refuse
Most of the ongoing work in a plumbing account is deciding what not to buy.
The traffic that arrives uninvited is predictable. Renters trying to find out who is responsible for a repair. Owners looking up how to bleed a radiator themselves. Supers searching for a part number. People looking for apprenticeships, licensing information or what a plumber earns.
Aggregator and franchise brand names belong on the list too, along with anything that reads like research rather than a problem.
Build the list from your actual search terms report, weekly at first and monthly once it settles. Nobody predicts the whole list in advance, and the odd ones are usually the expensive ones.
Exclusions are the fastest available lever on cost per booked job. They cost nothing, take effect immediately, and compound.
A call you cannot get freight access for was never a lead
Count conversions at the invoice and every other decision in the account gets easier.
Lead counting flatters an account. A call that ends with no elevator window, no board approval and no scheduled date is a cost, not a conversion.
Push outcomes from your scheduling system back into the ad platform, or reconcile by hand each month if the integration is not there yet. Either is better than optimizing toward form fills.
Give every inquiry an outcome code. Booked. Refused as out of area. No access. Waiting on board approval. Lost on price. Wrong trade. The codes themselves change how the office thinks about the phone.
Then bid toward what closes. Some of the cheapest clicks in the account produce the worst work, and you will only see that once the numbers are joined up.
Questions we actually get
- How should we split budget across the boroughs?
- Start by splitting at all, then let the numbers move the money. Manhattan, Brooklyn, Queens, the Bronx and Staten Island behave differently on click cost, job type and drive time, and a single blended report hides which one is paying. If you also work Nassau County or Westchester, treat those as separate decisions with their own budgets rather than as spillover from a radius.
- Are Local Services Ads worth running alongside search?
- They do a different job. Local Services traffic skews toward immediate, smaller work and is priced per lead, which suits an emergency book. Search campaigns are where you build the planned work: repipes, riser replacements, board approved projects. Running both and reporting them separately is normally more useful than choosing one, provided you can tell which channel produced which booked job.
- What is a reasonable cost per booked job for a New York plumber?
- We will not invent a figure, and any agency quoting one before seeing your numbers is guessing. The right approach is to establish your own baseline from real outcomes, then track it by borough and job type. A blended average across a metro this varied is close to meaningless. What matters is the direction of travel on the segments you can actually service.
- Should the ads mention price at all?
- Usually yes, in the form of a range or a diagnostic fee, and for the same reason the landing page should. Access costs are real here and pricing that looks high next to a national average is not high, it reflects a harder job. Saying so filters the callers who were never going to book and reduces the number of conversations your dispatcher has to repeat all day.
- How quickly should we expect the account to stabilize?
- Longer than most people want, and we will not promise a date. The first weeks are mostly exclusion work, because the search terms report tells you things no keyword plan predicted. Planned work campaigns take longer still, because jobs that go through a board close on the building's schedule and not on yours.
