Where should the next $1,000 of Google Ads go?

TL;DR

• Allocate an additional advertising budget using the added completed jobs and contribution expected from each option. Historical average cost per job describes the work already bought.
• In our invented example, the cheapest historical campaign costs $100 per completed job. Its proposed extra $1,000 buys only three more jobs, leaving −$250 after added media under the stated assumptions.
• With a $1,000 media ceiling and ten spare crew hours, the example selects the drain campaign's $1,000 option: five additional jobs and $1,000 remaining contribution after added media. Fourteen available hours permit a different combination.
• Use the blank decision sheet and open 27-plan comparison. Every forecast input is invented. The calculation does not predict your account's results or change its settings.

Your repair campaign has the lowest cost per completed job. Your agency proposes putting the next $1,000 there. Before approving it, ask what the next $1,000 is expected to buy.

The answer may differ from the campaign's historical average. An owner also needs to know whether the crew can deliver the proposed work and how much money it leaves. This guide puts those questions into one small decision table.

It extends the marketing budget plan, which starts with the business's overall workload and cash. Here, the total possible increase is already agreed. The decision is where to place it, including whether to spend it at all.

What is marginal cost per completed job?

For a proposed increase, divide the additional media cost by the additional completed jobs you expect it to produce:

Marginal media cost per completed job = added media ÷ added completed jobs.

If the extra $1,000 is expected to produce three completed jobs, the figure is about $333.33 per added job. The historical campaign average could still be $100.

Keep the outcome consistent. A platform conversion might represent a form, a qualified inquiry or a completed job. A simulator's conversion estimate needs a documented bridge to the business outcome you are pricing. The lead cost ceiling guide explains how booking, completion and operating costs fit together.

If you expect no extra completed jobs, the ratio is undefined. Write “no additional completed jobs assumed” and inspect the cost. Do not enter a zero-dollar acquisition cost.

The word “marginal” here describes a proposed budget step. It does not turn a forecast or a before-and-after difference into proof of causal lift. For that separate question, see the comparison-group guide.

Start with money each additional job can leave

For this decision, define a contribution amount per completed job before acquisition cost. Start with the amount you expect to collect, then subtract costs required to deliver that extra work. Include parts, subcontractors, card costs and labor or overtime that the additional work actually requires, as appropriate to your shop.

Write the definition beside the number. Payroll already committed for the week and extra overtime answer different cost questions. If a job uses time that could serve another customer, the available-hours constraint still matters even when the payroll check would be unchanged.

Next calculate:

Added contribution after acquisition = added jobs × contribution per job − added media − added fees.

The result is the contribution remaining for fixed overhead and other business needs. It is not company net profit. Avoid mixing invoice revenue in one campaign with contribution in another.

Compare three invented service campaigns

The table uses a hypothetical contractor with three campaigns and a common crew pool. All amounts, job counts and hours are invented teaching inputs. They are not client observations, Google forecasts or typical trade economics.

Historical averages make repair appear the cheapest:

Those totals alone cannot tell us where the next dollar belongs. The owner and agency now write down separate proposed increases:

The $1,000 level for a campaign replaces its $500 level. They are alternative total increases above the same baseline; they cannot be added together.

For repair's larger option, three jobs at $250 leave $750 before the added advertising. Paying $1,000 for that work uses $250 more than the stated contribution. The low historical average does not rescue this proposed step.

The panel option leaves more dollars per added job, but it also needs more crew time. Neither the average acquisition cost nor the largest ticket gives a complete allocation decision.

Ten spare hours change the recommendation

Suppose the owner permits at most $1,000 of extra media and has ten usable crew hours for the additional completed work in the same planning period. These hours are after existing commitments, travel, failed visits and an operating reserve have been allowed for. The example assumes no extra management fee.

There are three choices per campaign: no increase, add $500, or add $1,000. That gives 3 × 3 × 3 = 27 combinations. The open script enumerates all of them, then keeps only those within both the media ceiling and the hours limit.

Consider the choices near the top of the list:

Under these inputs, the highest remaining contribution among feasible plans is the drain campaign's $1,000 increase. It uses the ten available hours and leaves $1,000 after the added media cost.

If the owner has fourteen usable hours, adding $500 to drain and $500 to panel becomes feasible. That combination leaves $1,400. The underlying advertising assumptions have not changed; the operational constraint has.

Do not manufacture four extra hours to make the attractive option fit. If the work needs a specific licensed technician or a different crew, a single pooled-hours calculation may be inadequate. The technician skill matrix handles that earlier capacity check.

Where do the extra-job estimates come from?

Ask the person proposing the increase to document the source, period, conversion definition and uncertainty of each estimate. A credible proposal can still be uncertain. Its assumptions should be inspectable before money is committed.

Google's simulator guide, checked October 29, 2026, explains that its tools use auction information to estimate performance under different settings. It also discusses sparse conversion data and delayed outcomes. A platform estimate is an input to review; it does not supply your crew capacity, completed-job margin or proof of incremental business.

Ask whether the expected conversions correspond to completed jobs. If the account counts qualified inquiries, document the assumed path from those inquiries to completed work. Keep that assumption visible instead of relabeling every estimated conversion as a paid job.

Use the same planning period across campaigns. Compare outcomes after an appropriate follow-up interval, and preserve unresolved jobs. Google's conversion-window guide, checked October 29, explains that the chosen window controls how long after an interaction a conversion can be recorded. A shop's operational time to completion is another part of the review.

Test a worse outcome before approving the increase

The downside fixture changes the assumed extra jobs while keeping the reserved hours and media costs. It deliberately leaves the crew reservation intact even if fewer jobs materialize.

For the $1,000 drain option, suppose only three additional jobs complete. Three at $400 leave $1,200 before media, or $200 after the $1,000 increase. That is materially different from the base assumption's $1,000.

This is a stress scenario chosen for teaching. It is not a confidence interval, a lower bound estimated from data or a claim about the probability of three jobs.

Also check any new fee. An additional $200 management or production charge would reduce the base plan's remaining contribution from $1,000 to $800. Under the downside assumption, the same fee would consume the entire $200 remainder.

The script charges the supplied additional fee once whenever a plan spends extra media. Existing fees that are unchanged belong in the wider business budget. If fees vary by campaign or option, incorporate them deliberately before using this simplified model; the supplied single-fee input cannot represent that contract.

Use the worksheet with your agency

Download the blank decision sheet. For each campaign, write down the additional media, expected completed jobs, contribution definition and crew time required. Include a no-increase option and a downside scenario. Then check the feasible combinations.

The offline comparison bundle includes the sample, downside assumptions, original source and tests. With Node.js 20 or later, extract it and run node core.mjs. Use node core.mjs downside.json for the stress scenario. The README explains the schema and the output.

Amounts use integer cents and crew time uses integer minutes. The tool chooses one level per campaign and compares every supplied combination. It never invents a spending level between the choices. Equal-surplus plans favor lower media spend, then fewer crew minutes; a final deterministic ordering settles an otherwise identical tie.

The public tests check the ten- and fourteen-hour decisions, an extra fee, a smaller budget, an unprofitable increase, invalid inputs and the downside case. Independent arithmetic and a separate enumeration were also used to check the worked example. Passing those checks establishes the calculation, not the truth of any forecast input.

What this doesn't cover

The model treats campaign options as separate and additive. It does not estimate overlap between campaigns, displacement of organic or returning-customer work, auction effects, service-specific staffing, cash timing or statistical uncertainty. If campaigns compete for the same jobs, summing their estimates can overstate the work available.

It assumes the current business plan remains in place. It compares increases above that baseline rather than choosing what to cut. It also excludes the financing and cash-flow effects of delivering work before customers pay; see the advertising cash-gap guide for that decision.

Lonta AI has a commercial interest as a marketing agency. This is an open planning method with invented inputs. We did not run these spending changes in a customer account, validate a Google forecast or observe additional paid jobs.

Record the next decision before approving the increase

Before approving an increase, record the selected option, its assumptions, total added cost, hours reserved and review date. Decide what evidence would justify continuing, revising or stopping it. The test stop-rules guide provides that decision record.

After the outcome window, compare the proposal with actual completed work and money collected. Record what changed in demand, tracking or operations. A forecast miss should produce a better next decision, rather than a rewritten original forecast.

For a wider plan connecting job economics, capacity and media, use the Lonta strategy review.

FAQ

Should the cheapest campaign always receive more budget? Check the expected cost and contribution of the proposed increase. The campaign's historical average describes earlier work and can differ from the next step.

Must the full extra budget be spent? The media amount is a ceiling. The comparison includes no increase and smaller increases. Under weak assumptions or added fees, leaving money unspent can be the best supplied option.

Does a simulator forecast prove the advertising will create more jobs? It provides a platform estimate under stated conditions. The link from platform conversions to completed work, capacity constraints and causal lift still require separate evidence.

Can the method compare HVAC, plumbing and electrical campaigns? Only after their outcome definitions, period and contribution basis are comparable. Different skill or scheduling constraints may require separate capacity pools beyond this example.

Sources

• Google Ads: bid, budget and target simulators, checked October 29, 2026.
• Google Ads: conversion windows, checked October 29, 2026.
• Original Lonta AI 27-plan example, source and tests. All business inputs invented; code MIT, original commentary and fixtures CC BY 4.0. Platform documentation retains its own rights.