What ChatGPT ads cost, and how to budget for them.

There is no honest universal price list

Nobody can responsibly give every home service advertiser one fixed cost per click for a developing channel.

Auction prices, buying options, inventory, reporting, and eligibility can change as an advertising product develops. Your market, service mix, schedule, offer, and campaign settings also shape what you pay. A roofing replacement campaign spanning several counties is not economically comparable to an emergency drain campaign limited to a few ZIP codes. Even when two advertisers buy similar traffic, their cost per qualified opportunity can be very different.

Treat a confident, context-free CPC quote as a warning sign. Ask what inventory, geography, period, billing model, and definition of a click support it. If those details are absent, the number is not a budget. It is sales theater. AdsOnGPT is independent and cannot set or guarantee platform pricing, placement, or delivery. A useful plan therefore starts with the most you can afford to pay for a business outcome, not a speculative market average.

Budget for the whole acquisition system

Media is only one line item. A cheap click that reaches a weak page or an unanswered phone is expensive in practice.

• Media

  The amount paid for eligible delivery under the objective you choose, currently CPM, CPC, or oCPC. CPM bills per 1,000 impressions, while CPC and oCPC both bill per valid click, so oCPC optimizes toward a conversion without charging per conversion. Keep media separate so its efficiency remains visible.

• Creative

  Offer development, copy variants, approvals, and refreshes. Strong creative requires trade knowledge and truthful operational details.

• Landing page

  Strategy, copy, design, development, speed work, form routing, phone behavior, and ongoing revisions for each meaningful offer.

• Measurement

  Analytics configuration, UTMs, call tracking, CRM fields, consent controls, testing, and reconciliation with booked-job records.

• Management

  Campaign setup, pacing, search-term or context review where available, testing, reporting, and coordination with dispatch and sales.

• Operations

  Call handling, estimate scheduling, sales follow-up, and capacity. These may be existing costs, but the campaign depends on them.

Work backward from a booked job

Use contribution economics to set guardrails. The following is an ILLUSTRATIVE worked example, not a forecast, benchmark, or recommended bid.

Run the model conservatively and show a range, not one magic number. Test what happens if close rate drops, margin compresses, or the mix shifts toward smaller work. Also distinguish a submitted lead from a qualified lead. Spam, out-of-area calls, job seekers, existing-customer service requests, and services you do not perform should not make the economics look healthier.

1. 1

   Enter an illustrative average job value: $8,000

   Replace $8,000 with collected revenue for the exact service being advertised. Do not blend a maintenance visit with a replacement job if they have materially different economics.

2. 2

   Apply an illustrative gross margin: 40%

   $8,000 × 40% = $3,200 gross profit before advertising and other overhead. Substitute your actual gross margin after direct labor, equipment, materials, permits, and commissions as your accounting defines them.

3. 3

   Choose an illustrative acquisition allowance: 20% of gross profit

   $3,200 × 20% = $640 maximum advertising cost per booked job under this example. This policy choice must reflect overhead, cash flow, capacity, repeat value, and risk tolerance; it is not an industry rule.

4. 4

   Apply an illustrative qualified-lead close rate: 25%

   At one booked job per four qualified leads, $640 × 25% = $160 affordable cost per qualified lead. Use CRM data for the advertised service, and define “qualified” before doing the math.

5. 5

   Apply an illustrative landing-page conversion rate: 10%

   If one in ten paid visitors becomes a qualified lead, $160 × 10% = $16 affordable cost per click. Replace 10% with observed traffic-to-qualified-lead performance. Early estimates should be treated as assumptions and revised with evidence.

Fund a readable test, not a ceremonial one

One figure is documented rather than estimated. The current minimum daily budget is $25 for accounts billed in USD, with different documented minimums for other billing currencies. That is the floor the platform requires before a campaign can run, not a recommendation and not a realistic test budget for most home service advertisers. Budgets can be set as a daily amount or a campaign total, and a daily budget behaves as an average across a seven day period, so daily spend will not exceed twice the selected amount and spend across the period will not exceed seven times it. Confirm the current minimum for your billing currency inside your own Ads Manager account.

A first budget should be large enough to generate repeated opportunities for the conversion action you care about, while remaining an amount the business can lose without distress. Start with the work-backwards model: define the maximum qualified-lead cost, decide how many qualified leads would let you inspect patterns across calls and forms, then add the fixed setup costs. That produces a planning range. It does not guarantee the platform will spend it or produce those leads.

A tiny budget spread across many services, counties, offers, and devices creates fragments. One click in each cell teaches almost nothing. Narrow the first test to a service with clear economics, a service area dispatch can support, and one primary offer. Set a calendar window that includes ordinary weekdays and operational conditions; avoid declaring victory or failure from a single weather event or holiday.

Write the decision rules before launch. Define spend caps, lead-quality criteria, the minimum follow-up standard, and who can pause the campaign. A readable signal includes call recordings or dispositions, not just dashboard totals. If tracking fails or phones go unanswered, stop and repair the system rather than paying to learn the wrong lesson.

Know what moves cost per lead

CPL is the result of an entire chain. Diagnose the chain before blaming the bid.

• Offer strength

  A clear, credible next step can earn more qualified responses than vague “contact us” language. Discounts are not the only offer; scheduling clarity or a useful estimate process may matter.

• Service-area breadth

  A wider footprint can add locations you serve less efficiently or cannot schedule promptly. Match targeting, page claims, and dispatch reality.

• Seasonality

  Demand, urgency, crew availability, and competitive behavior change with weather and planning cycles. Compare periods with context rather than treating every week as equal.

• Speed to lead

  A qualified inquiry loses value when it sits unanswered. Measure answer rates, callback lag, and appointment-setting outcomes by time of day.

• Landing-page quality

  Message match, speed, proof, mobile usability, and a low-friction contact path determine whether paid attention becomes a real inquiry.

• Lead definition

  If reporting counts every ring and submission, CPL may look low while sales sees little value. Optimize to a shared, documented qualification standard.

Decide whether the test earns another dollar

Continue when tracking is trustworthy, the campaign is reaching the intended service and geography, leads are genuinely eligible, and booked-job economics are moving toward the guardrail you set. You do not need to pretend early data is conclusive. You do need a specific hypothesis for the next spend block (for example, a clearer replacement offer or faster call routing) and a limit on what it may cost to test.

Pause when the business cannot respond, the landing experience is broken, qualification data is missing, or spend has reached the preset ceiling without enough evidence of a viable path. Stop when verified lead quality and downstream outcomes remain outside your economics after meaningful fixes, or when the channel conflicts with capacity and cash flow. Sunk spend is not a reason to continue.

Compare the channel using booked jobs, gross profit contribution, cancellation behavior, and sales effort, not clicks alone. Preserve a written record of assumptions and changes so the next test starts smarter. If the case remains uncertain, a constrained follow-up test is more responsible than calling an ambiguous result either a failure or a success.