One campaign for five boroughs hides which borough is losing you money
Averaging Manhattan, Brooklyn, Queens, the Bronx and Staten Island into a single line of reporting guarantees you cannot see where the budget is being wasted.
Split the account by borough first and by keyword second. Each one has different competition, different housing stock and a different job mix, and each deserves its own budget you can turn up or down without touching the others.
Manhattan buys differently. A large share of the work is buildings rather than houses, and the person searching often reports to a board or a managing agent. Staten Island and parts of eastern Queens look far more like a conventional single family market with pitched roofs.
Brooklyn splits internally. Brownstone belt work in Park Slope and small flat roofs in Bushwick are the same borough and not the same sale.
Nassau County and Westchester are separate decisions entirely, not an extension of the radius. If your crews go there, fund it deliberately and watch it on its own line. If they do not, exclude it and stop paying for the overlap.
A ceiling stain on the top floor is this week's job; a full replacement is next year's
Emergency and planned work convert on completely different clocks, so putting them in one campaign means the fast one eats the budget and the slow one never gets measured.
Leak and repair searches convert within days. Somebody has a bucket on the floor and is calling three companies before lunch. Those jobs are small, but they are how you meet a building before the replacement is budgeted.
Replacement, recover and coating searches move slowly. In a co-op or condo, the decision generally waits on a board, a reserve study and a meeting. Judging that campaign on this month's bookings will get it shut off before it ever pays.
Run them separately, with different landing pages, different bids and different expectations. Repair is measured weekly. Replacement is measured across quarters, and the conversion you count first is a completed roof survey rather than a signature.
Commercial and building work should be third. A super calling about one leak may be attached to a portfolio, which is not something a keyword level report will ever show you.
In a city where the roof is also a terrace, half the clicks are not customers
Roofing keywords in New York catch an enormous volume of traffic that has nothing to do with fixing a roof, and the negative list is worth more than any bid adjustment.
Start with the terrace problem. Roof deck construction, roof garden, rooftop bar, rooftop venue, roof access, roof top party. Apartment listings advertising a roof deck as an amenity will pull your ads in constantly if you let them.
Then the trade adjacent traffic: solar installers, green roof grants, chimney and fireplace work, gutter cleaning if you do not sell it, skylight retailers, and every variation of roofing jobs, roofer salary and roofing apprenticeship.
Then the researchers. Cost calculators, DIY patching, material suppliers and anybody typing the word wholesale. None of them are buying a roof from you this month.
Build the list before launch, then mine the search terms report weekly for the first two months and monthly after that. In a market this expensive, negatives protect more budget than any change to a headline will.
Bridges, tolls and the BQE decide your targeting, not a circle on a map
Radius targeting is drawn on a map with no water, no tunnels and no traffic, which is not the city your crews actually drive.
Target by borough and by zip code, not by miles. A circle drawn around a yard in Long Island City takes in half of Brooklyn, a slice of Manhattan and a piece of the Bronx, and pays the same price for all of them regardless of whether you would take the job.
Set location targeting to people in or regularly in your area rather than people showing interest. Interest based targeting in a city this famous means paying for someone in another state reading about a rooftop in Chelsea.
Bid by how expensive the work is to reach, not just by how expensive the click is. A morning in Manhattan with no place to park a truck is not the same cost of service as a job eight blocks from the yard.
If you do not cross into New Jersey, exclude it. If you take Westchester only for larger buildings, run it as a separate campaign with different keywords rather than letting it bleed into the residential budget.
Budget pacing should track rain, ice and the board's meeting calendar
Roof demand in this city arrives in bursts that a flat monthly budget will always miss the front of.
Leak calls spike during and just after heavy rain. If the budget is capped evenly across thirty days, the ads stop showing at exactly the hour the phone would have rung. Build headroom for those days and take it back later in the month.
Winter narrows what can actually be installed, and ice on a flat roof stops crews. Rather than going dark, shift the money toward survey and planning searches, because that is when buildings decide what they will fund for spring.
Board driven work runs on its own calendar. Buildings that approve in late winter start in spring, so the campaign that produces those jobs has to be funded well before the season you want to fill.
Pacing also has to respect your own capacity. Advertising for emergency leak work on a week when every crew is committed to a tear-off buys you angry callers and nothing else.
A managing agent with nine buildings breaks every cost per lead average
One inquiry from somebody who controls a portfolio is worth more than a month of homeowner form fills, and a cost per lead column cannot tell them apart.
Decide what you are actually buying before you report on it. For repair, the number is cost per booked visit. For replacement, it is cost per completed roof survey, and eventually cost per signed contract.
Feed signed jobs back into the platform. Importing offline conversions with real values teaches the bidding which clicks turned into work, which matters far more here than in a market where every job is roughly the same size.
Keep buildings and houses in separate reports permanently. Blending a four hundred dollar repair on a Bay Ridge two family with a full replacement on a prewar elevator building produces an average that describes neither.
When a report shows cost per lead going down and signed work going flat, the campaign is buying cheaper clicks from worse buyers. Catching that early is most of the value of measuring properly at all.
Questions we actually get
- What should we budget to start?
- We will not quote you a number before looking at your service area and the mix of work you want. What we can say is that funding a single borough properly beats spreading the same money across five, because a campaign that runs out of budget by mid morning never learns anything. Start where your crews are strongest, prove the cost per booked job there, then expand with numbers instead of hope.
- Are Local Services Ads worth it here?
- They sit above search ads and take the calls first, so for repair and leak work they are usually worth testing alongside search rather than instead of it. They also demand real responsiveness; missed calls hurt you inside that product. For larger building work and board driven replacement they matter much less, because that buyer is not calling a badge, they are checking whether you have done their building type before.
- How do we stop paying for people searching about roof decks?
- A negative keyword list built before launch, then maintained. Roof deck, roof garden, rooftop bar, roof access and every hiring variant go in on day one. After that it is a weekly read of the actual search terms for the first two months. In this city that single habit protects more money than bid tuning ever will.
- Should the ads run at night?
- Only if somebody answers. Leak traffic does arrive in the evening and during storms, and it converts well, but an ad that produces a voicemail at nine at night is worse than no ad, because the caller has already moved to the next company by the time you listen to it. Decide the coverage first, then set the schedule to match it honestly.
- Can you promise a number of leads per month?
- No, and you should be wary of anyone who will. Auctions move, weather moves demand, and your close rate is part of the outcome. What we do commit to is structure you can inspect, reporting that traces spend to signed roofs by borough, and a monthly conversation about what to cut and what to fund next.
