How to Get Roofing Leads: What Each Channel Costs

Every real source of roofing leads for an Ontario contractor, what each one costs, and where each one breaks. Published benchmarks, no shortcuts.

By Gavin Sevastian · Updated August 4, 2026 · 18 min read

The phone was busier last month. You are booked through next week and after that it thins out, and you would rather sort that now than in three weeks when it is a problem.

This guide is about how to get roofing leads without guessing: every real source, what each one costs, and where each one breaks. It does not promise a shortcut, because there is not one. If you never hire anybody to help with this, you should still be able to act on what is here.

Here is the honest conclusion up front. You can build a lead channel you own instead of renting one, and that difference compounds, because a marketplace keeps the customer relationship and you keep nothing when you stop paying. Doing it properly is ongoing work rather than a weekend setup.

Where roofing leads come from

A roofing job finds you in one of five ways, and every one of them trades money, time, or control. Nobody gets all three.

The five channels, and what each one trades

Referral and repeat work. The cheapest leads you will ever get and the slowest to scale. They cost you nothing in ad spend and everything in years. The problem is not quality, it is volume: you cannot turn referrals up in February because the schedule looks thin.

Door knocking and canvassing. Fast to start, free to try, and entirely your time.

Buying leads from a marketplace. You pay per enquiry and somebody else owns the relationship, the reviews, and the phone number. It is the fastest way to put work in front of a crew this month and the worst way to build something you own.

Local Services Ads. Google's pay-per-lead product. It sits above the paid search results, you pay per lead rather than per click, and getting in requires verification, insurance documents, and calendar time.

Search advertising and SEO. You buy the moment someone decides they need a roofer. Ads start and stop on command and cost money per click. SEO compounds and costs months. Both leave you owning the lead outright.

You are competing with 1,777 Ontario roofing companies, and most of them are small

Statistics Canada counts 1,777 roofing contractors with employees in Ontario in its December 2025 business counts release, out of 4,721 nationally. Ontario holds 37.6% of the country's roofing businesses. Of those Ontario firms, 1,041 have between one and four employees, 393 have five to nine, and only 52 have fifty or more. That works out to 80.7% under ten employees.

That count only includes businesses with payroll. Owner-operators working alone are not in it, so the real field you are bidding against is larger than the table shows.

Read that as good news rather than bad. The field is overwhelmingly small operators, which probably means much of your competition is not running a properly maintained advertising account either.

Why doing all five is bad advice for a small roofer

You will run out of attention before you run out of budget. Two channels run properly beat five run badly, every time, because the channels that produce reward maintenance and the ones that do not just quietly drain money.

A reasonable pair for most Ontario roofers is one channel you own and one that fills gaps. What that pair should be depends on your capacity, your close rate, and how much of the year you need to cover.

What a roofing lead costs

LocaliQ's 2025 home services benchmarks are the closest thing this trade has to a public price list, drawn from more than 3,200 US search campaigns run between April 2024 and March 2025. WordStream's 2026 benchmarks cover a broader set of industries over a later window.

The published benchmarks, and the two things wrong with them

Two problems, and they matter before you read a single figure.

First, these are US campaigns reported in US dollars from US panels. An Ontario contractor's real numbers will differ, and reading a USD median as your expected cost is the most likely way this page misleads you.

Second, LocaliQ and WordStream publish under the same parent company. Treat them as one source rather than two that happen to agree.

Use the numbers as a sanity check on order of magnitude and nothing more. If you want the broader picture on what contractors pay to run search advertising, that is a separate question worth reading properly: what Google Ads cost for contractors.

Roofing carries the highest cost per lead in home services

Of all the home service categories LocaliQ measured, roofing showed the highest cost per lead and one of the lowest conversion rates. The consequence matters more than the statistic.

At a conversion rate under 4%, a modest monthly budget buys very few conversions. Few in the sense that the monthly count is a small enough number that random variation swamps it. Compare that to plumbing at 7.63% or air conditioning at 6.56% and you can see why roofing advice borrowed from other trades tends to mislead.

A short quiet stretch in a low conversion category is not by itself evidence that anything is broken. Check the obvious mechanical things before changing the account, because reacting by pausing keywords or switching bid strategies destroys the only data that would have answered the question.

Ask which cost per lead you are being quoted

Branded searches are people typing your company name. They already knew who you were. Non-branded searches are people who typed roofing contractor and had no idea you existed. Those are different products and they cost different amounts.

SearchLight Digital's Q1 2026 roofing data, covering 15 roofing contractors and 3,304 leads, reported $124 USD per lead on non-branded search against $44 USD on branded. That is one agency's own book of business rather than a platform-wide panel, so the useful part is the roughly 3x gap and not the dollar amounts.

The practical use is a question. When anyone quotes you a cost per lead, ask whether it is branded, non-branded, or blended. A blended number averages in demand that already existed and would have found you anyway.

Buying leads from marketplaces

A shared lead means the same homeowner enquiry goes to several contractors at once. You are not buying a customer. You are buying a place in a race, and the homeowner is collecting other quotes while you drive over.

Why the price is not the cost

The real cost is the lead price divided by the share of those leads you close. Close one in ten and a lead costs you ten times its invoice price to turn into a job. Close one in three and it costs you a little over three times. Same invoice, completely different business. Do that division before you sign anything, using your own close rate rather than the one in the sales deck.

The operators active in Canada include HomeStars, which is Angi's Canadian marketplace, plus TrustedPros and Bark.

Canadian marketplaces do not publish per-lead pricing, and any figure you find in a blog post has no methodology behind it. Get your number in writing before you sign.

The FTC action, and exactly what it covers

The US Federal Trade Commission went after HomeAdvisor, a company affiliated with Angi, in January 2023 over how it sold leads. The proposed order required payment of up to $7.2 million USD.

One allegation was about matching. The complaint alleged HomeAdvisor represented that providers would only receive leads matching the services they provide and their preferred geographic area, when many did not.

The other was about conversion. The complaint alleged the company told providers its leads turned into jobs at rates much higher than it could substantiate. Providers at the time paid an annual membership fee of $287.99 USD on top of per-lead charges.

Be precise about scope. That is a US regulator acting against a US entity. It is not a finding against HomeStars or against any Canadian operation. What it is useful for is knowing exactly which promises to get in writing before you sign.

When buying leads is the right call anyway

There are situations where it genuinely is. You have crew capacity you cannot fill this month and empty trucks cost more than expensive leads. You have no website and no time to build one. Or you want a floor of volume while a slower channel matures underneath it.

The failure mode is treating purchased leads as a permanent channel rather than a bridge to something you own. Five years in, you are still renting, the price has gone up, and you have nothing to show for it.

Google Local Services Ads for roofers

Local Services Ads sit above the regular paid results and charge per lead rather than per click. Most roofing content written about them is out of date, so here is what is currently true.

Roofers are an eligible category in Canada

Google's Canadian eligibility list includes Roofers, alongside plumbers, electricians, HVAC, and a dozen others. If you want the longer treatment of the product itself, read Local Services Ads for contractors.

Google describes the product in terms worth quoting, because it is the cleanest statement of how it differs from a marketplace. Customers choose you, and you only hear from customers who have specifically selected your profile out of all the rest. You pay only for leads related to your business and the services you offer.

Google's Canadian screening page sets out what roofing specifically requires: a business background check and an owner check, general liability and professional liability insurance, and business and owner licences where applicable by local law. The same page states that verification takes three to four weeks on average after documents are submitted, and that advertisers must have a public and verified Google Business Profile.

That licence requirement resolves at the city level in Ontario. The province's own guide for home renovation and roofing businesses does not point roofers to a provincial licence. It tells them to check with their local city or municipality for requirements including licences.

1. 1

   Get your Google Business Profile verified first

   Google requires a public, verified profile before a business can advertise with Local Services Ads. If yours is unclaimed or unverified, that is step zero and it gates everything else.

2. 2

   Gather insurance documents

   General liability and professional liability insurance certificates are hard requirements for roofing. Not optional, not waived for small operators.

3. 3

   Clear the background checks

   Google runs a business check and an owner check. Pre-badge ads can run before verification is complete, but they rank below verified providers, so treat the checks as the thing that unlocks real performance rather than the thing that unlocks the switch.

4. 4

   Sort your municipal licence

   Ontario points roofing businesses to their municipality for licensing. Toronto, for example, requires a Building Renovator licence for any business providing repair or renovation services, including any business that advertises them.

5. 5

   Allow three to four weeks

   That is the average Google states on its screening page once documents are in. Plan the timeline backwards from the season you want to be live for, not forwards from today.

Google Guaranteed no longer exists

If an article about roofing leads mentions the Google Guaranteed badge or the money-back guarantee, it is out of date and you can discount the rest of it.

Google consolidated its advertiser badges into a single Google Verified badge and discontinued the money-back guarantee that came with the old one. Eligible consumers could submit reimbursement requests only for services booked before December 7, 2025. Existing verified advertisers moved across automatically, and Google states that ad ranking was not affected by the change.

Nothing to do here except know it, and use it as a quick freshness test on any roofing marketing advice you are handed.

You cannot bid your way to the top of Local Services Ads

This is an easy way to waste a first month. Bid is one input among several, and it is not the dominant one.

Google's documentation on LSA ranking says rank is set by your bid and by how likely your ad is to result in a lead, which is based on your responsiveness to customer enquiries, the context and relevance of what the customer searched for, and profile quality. It specifically names your rating, your number of reviews, average response time, use of high-quality images, and completed verification checks. Proximity to the searcher and the hours you keep are the other well-understood inputs.

So an LSA that is not ringing is usually a reviews problem or a responsiveness problem. If you are letting calls go to voicemail at 4pm on a Thursday, no bid fixes that.

Two details worth knowing. The reviews that count for Local Services Ads are your Google Business Profile reviews, so every review you have ever asked a customer for is already working here. And lead credits, booking tracking, and Google's recording of inbound LSA calls are all available in Canada, which means a lead that arrives outside your services or your area can be disputed rather than absorbed. SMS lead notifications are US only, so build your response process around the app and email.

Google Ads for roofers: what you are buying

Roofing is not one demand type, it is at least three

Emergency repair. Active leak, water coming through a ceiling, shingles in the driveway after a windstorm. These people are urgent, they call rather than fill in a form, and they decide in minutes. They are not comparing four quotes. They want somebody who can come today.

Planned replacement. The roof is twenty years old, or the house is going on the market, or the insurance company sent a letter. These people gather several quotes over two or three weeks, they convert on a form more often than a phone call, and the job is worth many times an emergency patch.

Insurance claim work. Different again. Longer cycle, an adjuster in the middle, different objections, and a homeowner who is often more worried about the claim than about you.

Those three want different ads, different pages, and different response speeds. An emergency searcher who lands on a page about financing options for a full replacement leaves. A replacement searcher who gets a page shouting 24/7 EMERGENCY CALLOUT assumes you are a patch outfit and leaves too. Any campaign that models roofing as one funnel is wrong before it launches.

What search advertising does that the other channels do not

You buy the moment of intent instead of somebody's attention. A billboard interrupts a person who was thinking about something else. A search ad appears at the second someone typed roof repair into their phone.

You own the lead outright. Nobody else got sent the same enquiry. There is no race, no membership fee, and no vendor sitting between you and the homeowner.

You keep what the account builds. Stop advertising for a month and it is all still there when you switch it back on.

Against that: you pay per click whether the click converts or not, and roofing clicks are expensive. LocaliQ put roofing at $10.70 USD per click. WordStream, on a different panel and a later period, put home improvement broadly at $8.33 USD and called it among the highest-CPC categories it measures, behind only legal services. Different panels, so do not read the gap between those two figures as precise. Read it as roofing sitting at the expensive end of an already expensive sector. Every mistake in the account is charged at that rate.

Where the money goes when nobody is watching

Four settings are worth checking before anything else, and every one is visible on a settings screen right now.

Display Network left on inside a search campaign. A search campaign is supposed to show ads to people searching. Leave the Display option ticked and your budget also buys banner impressions across apps and websites where nobody was looking for a roofer. The clicks look cheap, and that is the problem, because cheap clicks pull budget away from the expensive ones that convert.

Auto-applied recommendations. Google can change your match types, add keywords, and switch your bidding strategy automatically. It is opt-out, not opt-in. On an account nobody logs into, this quietly rewrites what you built while you are on a roof.

Location targeting on the default. Google's recommended setting includes people who have merely shown interest in your area, not only people in it. Roofing is unusually exposed to this because people research roofs for a parent's house, for a rental two towns over, or for a place they are about to buy.

Clicks bought at hours nobody answers. If the phone goes to voicemail after 5pm and your ads run to midnight, you are paying full price for calls that ring out. No bidding strategy repairs that.

None of this is a conspiracy. It is incentive alignment. Every one of those defaults increases the number of billable events Google can charge you for, and Google is not hiding them. They are switched the way that suits Google unless you switch them the other way.

Hire intent and browse intent are not the same market

Two searches can show similar volume in a keyword tool and represent completely different people.

Some searches come from homeowners learning: comparing materials, checking whether the quote they already have is fair, working out how long a roof should last. Others come from homeowners who have finished all that and are now looking for the person to do the work. A keyword tool shows both as demand, because a keyword tool counts searches and cannot see intent.

That matters for two reasons. The first is sizing. Counting research volume as your market overstates the opportunity, sometimes by a wide margin, and it is how contractors talk themselves into budgets that were never supported by real buyers.

The second is scope. The intuitive move for a contractor who wants full replacements rather than small patch jobs is to bid on words like premium, luxury, or high-end, on the theory that expensive words attract expensive jobs. It does not work that way. What a searcher types tells you very little about the size of the job waiting at the other end of it, and any article selling keyword choice as a way to filter for big work is selling something that does not exist. Job size gets sorted out in the conversation, on the page, and in how you qualify an enquiry.

Working out which side of that line a keyword sits on is most of the work in a roofing account, and it is not something a keyword tool will tell you.

Setting up a roofing campaign that does not leak

Negatives and positives match by opposite logic

Getting this wrong produces a false sense of safety that lasts for months.

Positive keywords expand. Google's own documentation on how keywords match to meaning explains that keywords match the meaning of a search, and notes that 15% of queries every day are ones Google has never seen before. Your keyword reaches synonyms, paraphrases, plurals, and phrasings nobody typed into the account. Exact match has not meant literally exact for years, so any advice premised on match type giving you precise control is out of date.

Negative keywords do the opposite. Google's documentation on negative keywords states they do not match close variants or expansions. A negative for the singular does not block the plural. A negative for one synonym does not touch the others. Negative phrase match respects word order, and one extra word in the search lets it through. The one thing negatives handle for you is misspellings.

The practical consequence: every variant has to be enumerated by hand, and the list decays as search behaviour shifts. Adding negatives once does not cover you. It covered you on the day it was done.

Which is why reading the search terms is not an optional refinement for big accounts. It is the only way to find out what you are buying, regardless of match type. The same logic governs broad match: it is workable alongside a cost guardrail and a maintained negative list, and it is a fast way to burn a roofing budget without them.

Tracking, and the phone specifically

Some roofing enquiries arrive as calls and some as forms, and an account that only counts one of them is being judged on part of its results.

Worse is call tracking with no duration floor. A conversion that counts any connected call counts voicemails, wrong numbers, and eight-second misdials as wins. Automated bidding then learns from those wins and buys more of whatever produced them. You end up with an account that looks like it is improving while the real enquiries decline.

If your ads are running and the phone is not ringing the way the reports suggest it should, the gap is usually in this layer rather than in the keywords. That is a diagnosis of its own: why your Google Ads are not getting calls.

What has to be true before you spend, in plain terms: tracking exists before the traffic does, the geography matches where your crews will drive rather than where people are merely interested, and the page answers the question that was typed. Getting those out of order is what costs money, because traffic bought before tracking exists is data you cannot read afterwards.

Where the ads land

Sending paid roofing clicks to a general homepage is the most common expensive mistake on this list. A homepage answers a different question than the search did, and it asks the visitor to navigate. Paid traffic does not navigate. It leaves. A page matched to the search, with the offer, the proof, and the contact method visible without scrolling, is the highest-leverage change available to most contractors and it does not touch the ad account at all.

Two related traps. Bidding on a competitor's company name looks like stealing their customers and usually is not: the searcher already has somebody in mind, the ad carries a relevance disadvantage that inflates the click cost, and the conversion rate is poor. For a contractor on a limited budget that money buys more on generic hire-intent searches where nobody has an incumbent.

And more budget does not fix a broken account. Increasing spend on an account with weak tracking, leaky targeting, and a mismatched page scales the waste proportionally. Fix the foundation, then spend.

Ontario roofing seasonality, and why a flat budget loses money

Roofing demand in Ontario swings hard across the year, and most contractors budget as though it does not.

The season does not ramp, it switches on

The shape below comes from a keyword tool pull of public Ontario search volumes on hire-intent roofing terms across a single twelve month window. Keyword tools group close variants together, so read it as the shape of demand rather than a market size, and a heavy storm year would move it.

Demand sits at the floor in December and January and stays low through most of February. Then it steps up hard from one month to the next as the weather breaks and the frost comes out of the ground. It holds high through the spring, peaks in early summer, and falls away as the cold arrives. Peak to trough on the terms measured is roughly ten to one.

Practically, your calendar has one long window where demand is abundant and expensive and one where it is thin and cheap. Those are two budgets, not one.

Two other seasons that are not the roofing season

Ice dam work is its own market with its own narrow window, concentrated in the coldest weeks and effectively absent through the summer and autumn. If you sell it, it needs its own campaign that wakes up and goes to sleep on a completely different schedule from your replacement work.

Eavestrough runs on a third calendar. It is an autumn service, with demand building as the leaves come down, months after roof replacement has already peaked and started falling away.

A roofer selling all three is running three calendars. Bundling them into a single always-on campaign means two of the three are always being funded at the wrong time.

What to do about it

A flat monthly budget underspends the window that pays and overspends the trough. It is spending the same money in a month where buyers are scarce as in the months when most of the year's replacement decisions get made.

There is a second trap that catches people who understand the first one. Structural changes to a campaign trigger a re-learning period, which means the worst possible time to restructure is the peak of the season you are restructuring for. Do the rebuild in the quiet months and let it settle before demand arrives. If you are working the Toronto market specifically, the pacing question is worth reading on its own: Google Ads in Toronto.

How to tell whether the account is working

A low cost per lead is not automatically good

Cost per lead is the number every agency leads with, and it is the easiest number in the business to make look good without doing anything useful.

A cheap cost per lead can be built out of duplicate form fills from one person, misdials, eight-second hang-ups, enquiries from two hours outside your service area, and people who wanted a quote on a shed. Each of those counts as a conversion. None of them counts as a job.

So before you compare anyone's number to anyone else's, ask what a lead is defined as in that report. If the answer is vague, the number is meaningless.

Five questions worth asking out loud

Ask these of yourself, or of whoever you are paying. Every one has a specific answer if the work is being done and a vague one if it is not.

Are conversions defined as things that could become a job? Not clicks on a phone number. Not page views. Not every connected call regardless of length. A conversion should be an event that a reasonable person would call an enquiry. If a report counts anything looser than that, everything built on top of it is wrong.

Has somebody read the search terms recently, and can they tell you what they blocked and why? This is the sharpest question on the list, because it cannot be answered from a dashboard. Either a person has sat down with the search terms report in the past few weeks and made judgment calls, or they have not.

Does the reporting separate company-name searches from everything else? Mixing them makes an account look better than it is and hides whether the growth channel is working. If the report is one blended number, ask for it split.

Is there a page built for the search, or is traffic going to the homepage? A homepage answers a different question than the one that was typed.

Can somebody explain what changed in the last month and what it was expected to do? A short account of what was tried and why, not a task list. If nothing changed and nothing was meant to change, that might be fine. If nobody can say, then nobody is watching, and you are paying for an account running on the settings it launched with.

The search terms report does not show you everything

Low-volume queries are withheld from the search terms report. Running the account longer does not reveal them, and spending more does not either.

In a small local account, which produces an unusual amount of one-off phrasing, that matters. Absence from the report is not absence from the auction. Somebody typed something, your ad showed, and you were charged, and you will not see the words.

Anyone who tells you they will see everything once there is more data does not understand how the report works. What you can do is read what is disclosed, block what is clearly wrong, and accept that the list will never be complete. What you cannot do is treat a clean-looking search terms report as proof that the traffic is clean.

Door knocking, storm chasers, and the Ontario rules that govern both

Door knocking for roofing is legal in Ontario

Get this right, because a lot of published advice has it wrong. Ontario banned unsolicited door-to-door selling of a specific, enumerated list of products in March 2018. The list covers furnaces, air conditioners, air cleaners and purifiers, water heaters, water treatment and purification devices, water filters, water softeners, duct cleaning, and anything that performs those functions.

Roofing is not on that list. Anyone who tells you door-to-door roofing sales are banned in Ontario is wrong, and so is anyone who tells you a storm crew knocking doors is breaking the law by knocking.

What governs a contract signed at the kitchen table

The province publishes a guide written specifically for home renovation and roofing businesses, and it is the rulebook the crew down the street is failing. The rules it sets out:

• A customer who signs at home gets a ten day cooling off period. They can cancel for any reason, with no fee.
• You must refund everything including the deposit within fifteen days of the cancellation notice.
• Any contract worth more than $50 has to be in writing.
• If an estimate is written into the contract, you cannot charge more than 10% above it unless the customer asked for different work and signed a change.
• If the customer asked you to start work during the cooling off period and then cancels, you are entitled to reasonable payment for work and materials already provided that cannot be returned.
• The province recommends deposits of no more than 10% of the total project cost.
• A customer can cancel most contracts within a full year of signing if they never received a copy meeting the requirements set out in that guide.

That last one is the storm chaser's weak point. A contract that does not meet the requirements stays cancellable for a full year. If yours does meet them, that is worth saying on the doorstep: here is what the province requires this contract to contain, and here is mine.

The paperwork that separates a business from a crew

WSIB coverage is compulsory in construction whether or not you have employees. Independent operators, sole proprietors, some partners and some executive officers working in construction must have coverage and must register, with some exceptions. It is also part of what makes you verifiable when a homeowner checks, and it is the kind of document Google asks for during verification.

Working at heights training is a legal requirement for workers on construction projects who use fall protection devices, set out in O. Reg. 297/13, and it expires three years after the program is completed. Worth diarising, because an expired certificate is the sort of thing that surfaces at the worst possible time.

Municipal licensing adds another layer. Toronto's Building Renovator licence is required for all businesses providing services to repair or renovate buildings and structures, including any business that advertises renovation services, and it requires a criminal record and judicial matters check for owners, officers and partners.

One more thing worth knowing about how this is policed: information about a business stays on Ontario's Consumer Beware List for 21 to 27 months. That is a long time for a bad complaint to sit where homeowners can find it.

Why door knocking is still a poor primary channel

It is entirely your time, it does not compound, and it stops producing the day you stop walking. Nothing you built yesterday is working for you today.

It also puts you in the exact doorway homeowners have been warned about, which means you spend the first two minutes of every conversation proving you are not the thing they are worried about. That is a real cost even when it works.

Canvassing a defined area after a genuine storm is a reasonable supplement. It is not a plan.

What happens after the phone rings

Ads amplify whatever you already do with a lead

Two roofers can buy identical leads at identical prices and end up with completely different businesses. The difference is in how fast the phone gets answered, how quickly a quote goes out, and what happens on the third follow-up that most people never make.

Advertising turns the volume up on your existing process. If that process is strong, more leads means more jobs. If it is weak, more leads means more people telling their neighbours you never called back.

Which is where the arithmetic from the marketplace section applies to every channel, not just purchased leads: your real cost is cost per lead divided by close rate, so answering faster and following up more moves the number further than bid tuning does.

Why this is a job and not a setup

Search behaviour shifts. Competitors enter the auction and leave it. Storms move demand from one week to another. Google changes defaults and adds features that are switched on for you. Negative keyword lists decay because people invent new ways to phrase things. And seasonality swings the same budget from underspending to overspending inside a few weeks.

The setup is a project. It has an end. The management does not.

That is why contractors eventually hand this over, and it is also why plenty of them keep doing it themselves and do fine. Both are reasonable. What does not work is setting it up, walking away, and expecting it to still be right in six months.

What to do this week

You can build a lead channel you own rather than rent one. The channels that build something you keep are slower to start and cheaper to hold. The ones that produce this afternoon are the ones you rent, and whether that trade is worth it for you depends on your close rate and how long you intend to keep running.

Whoever ends up running your ads, here is the order to work through:

1. Claim and verify your Google Business Profile, and ask your last ten happy customers for a review. Everything in Local Services Ads runs on those reviews, and they cost nothing.
2. Open your campaign settings and fix the four defaults from the settings section above: location targeting, Display Network, auto-applied recommendations, and the hours your ads run.
3. Confirm calls are being tracked with a duration floor, so voicemails and misdials stop counting as leads.
4. Read one month of search terms and write down what you would block. That exercise alone tells you more about your account than any report.
5. Check that paid traffic lands on a page built for the search rather than on your homepage.
6. Work out your close rate on the leads you already get, and divide your cost per lead by it. That is your real cost per job.
7. If you sell ice dam or eavestrough work, put each on its own budget calendar rather than one always-on campaign.

If you would rather hand the whole thing to someone who does this every day, here is what that looks like and what it costs. And if you are going to do it yourself, do it before the spring step-up. The season does not wait for the account to be ready. Fix it in the quiet months.

Common questions about how to get roofing leads