How to calculate a Google Ads budget before you launch

There is no honest universal Google Ads budget for a local-service business.

A plumber, roofer, cleaner, painter, landscaper, and restoration company do not share a service area, a demand curve, an average job value, a close cycle, a margin, a crew count, or a tolerance for wasted spend. A budget that is too small for one market can be right for a narrow service in another. And a large budget does not rescue bad tracking, slow follow-up, poor service fit, or a business that cannot take more work.

• A budget is an output, not a guess.It comes from what a booked job is worth, your contribution margin, the share you assign to acquisition, how many jobs you can serve, and the cash you will risk.
• Capacity sets the ceiling.Crews, slots, routes, and cash usually limit the economical budget before the auction does.
• Write a test cap before launch.The maximum you authorise before reviewing the evidence, with a review date that fits your sales cycle and booking lag.

The useful question is not “what do other companies spend?” It is: what amount can this business authorise to test a defined outcome, without confusing a planning scenario with a guarantee?

Size your planning budget

The calculator converts job value, contribution margin, the share of contribution assigned to acquisition, capacity, and cash into a planning budget and a written test cap. Treat its output as a scenario to argue with, and rerun it with pessimistic and optimistic inputs before committing the cap.

A responsible budget model should return four separate outputs, and it should never blend them into one confident-sounding suggestion.

If the assumptions are weak, the output should be labelled weak. A model that hides its inputs is not a plan; it is a number with a decimal point.

Start with the business, not the platform

Before choosing a daily budget, write down eight things.

1. The service and geography being advertised.
2. The number of new jobs the business can accept.
3. The average revenue for the defined booked or finished job.
4. The variable costs deducted before contribution is calculated.
5. The expected or observed booked-job acquisition cost.
6. The sales-cycle and conversion-delay assumptions.
7. The cash the owner is willing to authorise for the test.
8. The business outcome Google Ads will improve toward.

A campaign should not receive a larger budget merely because the platform is capable of spending it. Demand, staff, vehicles, crews, appointment slots, response process, service area, and working capital can each cap the number of jobs the business can profitably accept.

Name the outcome before you name the number, because these are different tests.

If the business cannot record booked or paid jobs reliably yet, use a reliable earlier event for platform diagnostics and keep the deeper outcomes in your one-page monthly report. Do not present an unverified event as a deeper one.

The core formulas

Four formulas do the work: maximum allowable cost per booked job from revenue and margin, a target acquisition cost from an owner-chosen planning factor, a planning budget from an evidence target, and estimated contribution after acquisition cost. Every input belongs to the owner; none is a benchmark.

Maximum allowable cost per booked job

maximum allowable CPBJ = average revenue per booked job × contribution margin

At $800.00 average revenue and a 35% contribution margin, the maximum allowable CPBJ is $280.00. That says one booked job contributes $280.00 before the included acquisition cost. It does not calculate the company’s full fixed-cost break-even point, and it does not prove that paying $280.00 is a sensible target. The margin basis has to be defined, and the owner still has to decide how much of that contribution can be assigned to acquisition.

Target acquisition cost

target acquisition cost = maximum allowable CPBJ × owner-selected planning factor

Choosing a factor below 1.0 leaves room for the costs the ceiling ignores: management fees, tracking, refunds, cancellations, overhead allocation, tax, and the return the owner actually wants. In an illustrative case where the owner selects 70%, $280.00 × 0.70 = $196.00. That 70% is one owner’s choice in one example. No published percentage of break-even is the correct target for every business, and any source that offers you one has stopped doing arithmetic and started guessing.

Planning budget from an evidence target

planning media budget = evidence target × target CPBJ

Six booked jobs at a $196.00 target CPBJ is a $1,176.00 planning budget. It does not mean the campaign will produce six booked jobs, and six jobs is not automatically enough evidence to settle a market.

Booked jobs from a lead model

expected booked jobs = planned media spend ÷ expected cost per lead × booked-job rate

At $1,500.00 of spend, a $75.00 expected cost per lead, and a 25% booked-job rate, the model returns 5 booked jobs. That output rests entirely on the cost per lead and the booked-job rate. If both are guesses, the result is a scenario, not a forecast you can defend.

Contribution after acquisition cost

estimated contribution after advertising = booked jobs × average revenue × contribution margin − included acquisition cost

State separately whether management, tracking, landing-page, creative, or software costs are inside the figure. Do not mix a media-only spend number with an all-in acquisition cost and then compare the result to last month.

The eight-step budget method

1. Define the outcome

   A campaign that counts every form submission needs a different test design from one that measures paid jobs. Name the stage (contact, qualified lead, booked job, finished job, or paid job) and record how it will be evidenced.

2. Calculate the owner-defined acquisition ceiling

   Expected revenue multiplied by the owner-stated contribution margin gives the contribution available before acquisition cost. That is a ceiling, not a target.

3. Estimate the test budget

   Multiply the number of outcomes the owner wants to evaluate by the target cost for that outcome. This describes what the owner is willing to invest, not what the campaign will deliver.

4. Use scenarios instead of one forecast

   Run pessimistic, base, and optimistic inputs. A single forecast hides the fact that click cost and booking rate are both uncertain.

5. Check the capacity constraint

   Count crews, appointment slots, emergency coverage, travel time, parts, estimate capacity, call answering, follow-up ownership, backlog, and seasonality. A plan can be economically attractive and operationally wrong.

6. Implement the platform budget correctly

   Translate the owner’s cap into an average daily budget, and understand how Google paces against it. The platform limit and the owner’s all-in exposure are two different numbers.

7. Set the decision date before launch

   Write the decision rule before seeing any results: start, end, service, geography, outcome, target acquisition cost, maximum media spend, maximum all-in exposure, capacity, tracking source, minimum evidence, and the continue, adjust, stop, or extend conditions.

8. Review the evidence in stages

   At the review date, find the stage that failed rather than judging the channel as a whole. The right next action depends entirely on which stage broke.

There is no universal 14-, 30-, or 90-day review rule. The date should account for the sales cycle, conversion delay, estimate and booking timing, completion timing, payment timing, seasonality, available search volume, capacity, and which of those outcomes is actually being measured.

Use scenarios instead of one forecast

Assume a $1,500.00 planning budget and these stage rates.

The chain behind those columns is simple.

clicks = budget ÷ CPC · qualified leads = clicks × click-to-qualified rate · booked jobs = qualified leads × qualified-to-booked rate · finished jobs = booked jobs × completion rate

Three columns are more honest than one, and they make the argument concrete: the pessimistic case is the one the business has to be able to survive. If a decimal booked job appears in the output, label it an expected value rather than a promise of a partial job.

The economic inputs deserve the same treatment, and every row here belongs to the owner rather than to a benchmark.

Sensitivity analysis: the number that matters most may not be CPC

Most budget conversations are arguments about click cost. Run the arithmetic and the click price often turns out to be the less powerful lever.

Take the base case: $1,500.00 of spend, a $12.00 CPC, a 12% click-to-qualified rate, and a 35% qualified-to-booked rate.

1,500 ÷ 12 = 125 clicks · 125 × 12% = 15 qualified leads · 15 × 35% = 5.25 booked jobs · 1,500 ÷ 5.25 = $285.71 planning cost per booked job

Now hold the CPC still and move only the booking rate.

The same spend and the same auction price produce a planning cost per booked job that ranges from $200.00 to $500.00. Who answers the phone, how fast the callback happens, and how the estimate is presented can move the economics as much as the auction does. That is why a CPC-only budget conversation is incomplete, and why “the clicks are too expensive” is a conclusion that has to be earned rather than assumed.

Capacity is part of the budget decision

The maximum economical budget is usually set by crews, slots, routes, and cash, not by the auction. Multiply the new jobs the business can actually serve by an acceptable utilisation factor, and treat that as the ceiling the media plan may not exceed.

maximum practical test outcomes = available incremental capacity × acceptable utilisation factor

Ask how many new jobs the business can take each week, which service lines have capacity now, whether the service area supports the needed route density, whether calls and forms can be answered during the advertised hours, whether estimates, dispatches, crews, and materials are available, whether extra work will delay existing clients, and whether the business has cash to fund labour and materials before payment arrives.

If the business can profitably complete six extra jobs next month, a plan built around twenty new bookings is not a growth plan. It is a scheduling problem with a media invoice attached.

Budget pacing

An average daily budget is a delivery-planning input. It is not a daily invoice, and it is not a promise of identical spend every day. Four numbers answer four different questions, and owners routinely compare the wrong two.

Average daily budget

The average amount Google uses to plan campaign delivery.

Owner question: what average did I authorise?

Served cost

The cost of the clicks or impressions actually delivered.

Owner question: what did Google serve?

Billed cost

The amount charged after the applicable limits are applied.

Owner question: what will I actually pay?

Daily spending limit

The maximum billable amount on a single day for most campaigns.

Owner question: what is my day-level exposure?

Monthly spending limit

The maximum billable amount over a month for most campaigns.

Owner question: what is my month-level exposure?

Media budget

The amount assigned to Google Ads clicks or impressions, excluding everything else.

Not the same as the all-in cap.

All-in acquisition cap

The maximum the owner authorises for the defined test, including management, tracking, and landing-page costs.

This is the number the business actually risks.

a fixed test budget and an end date

The period used to decide when contribution or profit recovers the acquisition cost.

Set it before launch, or the verdict becomes a matter of mood.

Google documents the mechanics plainly. For most campaigns, a campaign may spend up to twice its average daily budget on a given day, and the advertiser will pay no more than 30.4 times the average daily budget over a month.

daily spending limit = average daily budget × 2 · monthly spending limit = average daily budget × 30.4

At a $50.00 average daily budget, that is a $100.00 daily spending limit and an approximate $1,520.00 monthly spending limit. Those are platform limits for most campaigns. They are not a suggestion to spend $1,520.00, not a promise the campaign will receive that volume, and not evidence that the work produced will be profitable.

Served cost is not billed cost

Google explains that served cost is the cost of the clicks or impressions received, while billed cost is the amount the advertiser is responsible for after adjustments. Google says served cost can exceed the daily or monthly limit in some circumstances, but that for most campaigns the billed cost will not exceed the applicable spending limits, and that the billed-cost report can be used to compare the two.

Thursday’s served cost runs past the daily spending limit while the billed cost stops at it. A quiet Monday and an expensive Tuesday are not a billing error. The check that matters is billed cost against the applicable limit, not whether each day equals the average.

Two more documented behaviours catch owners out. Google says that when a campaign uses ad scheduling, the monthly spending limit remains 30.4 times the average daily budget, and the system paces toward that full monthly limit regardless of how many days the campaign is scheduled to be active. Google also says that if the average daily budget changes mid-month, the rest of the month paces to the new daily budget multiplied by the remaining calendar days rather than the remaining scheduled days, and that a campaign created mid-month paces on the remaining calendar days too.

So restricting ads to phone-coverage hours does not by itself lower the monthly exposure. If the business wants lower exposure, change the budget or use a separately budgeted campaign.

Local Services Ads use a different control entirely. Google says LSA advertisers set an average weekly budget based on the average number of leads they want to receive, and pay for valid leads under that product’s billing rules. Do not transfer a Search budget formula to LSA without translating the event being purchased: the Local Services Ads budget guide works through that model separately.

A pacing worksheet

Fill this in weekly during a test. It takes four minutes and it prevents the most common budget argument, which is really an argument about which number someone was looking at.

When the account looks like it overspent, work through the causes in order: is the figure served cost or billed cost, is the date range a day or a month, did the budget change mid-month, was the campaign created mid-month, is ad scheduling active, is the budget shared or separate, is the report showing one campaign or the account, and is the amount inside the documented limit?

Pacing can also look wrong when the real problem is elsewhere: weak search intent, the wrong service area, inaccurate or duplicated conversions, missed calls, weak qualification, a low booking rate, unavailable capacity, delayed CRM imports, or a campaign improved toward a shallow event. Do not lower the budget to hide a measurement problem.

Set the test cap before launch

A test cap is the maximum amount the owner authorises before reviewing the evidence. It is not the target budget and it is not the platform limit.

Write down the service and location, the media budget and the all-in cap, the main conversion outcome, the evidence target, the expected conversion lag, the maximum allowable acquisition cost, the date or evidence condition for review, and the conditions for continue, adjust, stop, or extend.

Then set a fixed test budget and an end date: the period over which contribution has to recover the acquisition cost for the test to count as a success. A first test that has not reached the end of its fixed test window has not failed. It is simply not finished, and the honest verdict is “not yet” rather than “no”.

Do not raise the cap because the campaign has spent money. Raise it because the evidence supports more demand and the business can serve it.

Budget and conversion lag

A click, form, call, qualified lead, estimate, booked job, finished service, and payment can all happen on different dates.

Google says conversions occurring after the selected conversion window are not recorded, and suggests setting the window long enough to cover the observed conversion distribution. A same-day emergency call, a recurring cleaning plan, a roof replacement, a commercial painting contract, and a water-damage restoration project do not share a lag pattern.

Where the identifiers, CRM, and consent model support it, Google’s offline conversion tools can connect later outcomes such as a contract signed or a sale finished back to the earlier ad interaction. State which stage the budget model is measuring, and whether the most recent period is still immature.

Read the result stage by stage

At the review date, classify the result before deciding anything.

Most budget decisions go wrong at the point where a symptom is read as a diagnosis.

A small test can produce a noisy result. That does not make it useless, but it does limit the confidence. If the period produced one or two booked jobs, report that fact without turning it into a market conclusion.

Scaling after the test

Do not treat a fixed percentage increase as a scaling law. Scale when the intended outcome is being recorded, the acquisition cost sits inside the owner’s target or its variance is understood, the business can handle more work, tracking and attribution are stable, the outcome is not inflated by duplicates or a temporary condition, and the owner has a defined downside limit if the next period is weaker.

A smaller increase is appropriate when the evidence is thin. A pause or a repair is more sensible when the business is full, calls are being missed, or the outcome data is incomplete.

Budget does not fix bad measurement

Before trusting any output from the model, verify that form submissions represent successful submissions, that calls are distinguished from phone-link clicks, that duplicate GA4 and Google Ads conversions are controlled, that main and secondary conversion settings match the decision, that booked jobs can be reconciled against a CRM, calendar, call system, or owner record, that conversion windows cover the actual lag, and that service area, capacity, missed calls, and response handling are all visible.

A larger budget applied to false conversions does not produce a better answer. It produces a larger measurement problem, faster. Work through the conversion tracking audit before increasing spend on a signal you have not validated.

What to do next

A budget you can defend is the input, not the verdict. The fixed test period guide shows how to cap the first months of spend and set the date you will judge the number by.