Reduce Google Ads costs without cutting services

A cost-cutting mandate usually arrives as a figure: ten per cent, twenty per cent. But the figure does not say which costs are meant, who controls them or how you can tell whether the reduction had an effect or was merely shifted.

Which costs are you actually reducing?

A cost category is a type of cost with its own recipient, its own evidence and its own approval path. A reduction programme without this allocation produces apparent savings: a lower fee looks like a saving even though the same work is now incurred in-house. So assign every item to exactly one cost category.

Two rules follow from this. First: a shift between two cost categories is not a saving as long as the shifted work and the associated costs are not quantified; the worksheet column “shifted work and costs” records both. Second: a reduction in one cost category proves nothing about the other three.

Five levers that reduce operating costs

Most of these levers act outside the account — precisely where cost programmes often look too late.

The fourth lever is maitiq’s own: the recurring account analysis runs daily as a workflow by default, and every proposal carries the proposed action, the reason and the data basis; old and new values are given where they are meaningful. Frequency, output format and traceability are therefore verifiable instead of estimated.

Why the fee formula is part of the decision when cutting costs

Every fee formula creates a different incentive in a reduction case. That is not an accusation but a structural property you should know before the conversation.

• Percentage of the media budget: If the contractual calculation basis falls, the fee falls with it — subject to any floors, caps or tiers. That is a property of the formula; it says nothing about a provider’s intentions. So record in the contract who may propose a budget cut and how it is assessed.
• Fixed fee or retainer: A media reduction does not change the fee. A fixed fee does not have to lead to silent scope cuts, however: check that the agreed scope of services is still delivered in full.
• Hourly or daily rate: Fewer billed hours can reflect genuine efficiency without reduced coverage. Check that coverage is unchanged instead of assuming it fell; without that check you risk cutting attention rather than costs.
• Performance-based components: The target measure decides the invoice. An unclear conversion definition immediately becomes expensive here.

For every formula, ask for the same four details: calculation basis, exceptions, behaviour when budgets change, and the services that would be dropped at a lower fee. The comparison of pricing models sets out the models; here only their behaviour in a reduction case matters.

maitiq does not use a percentage-of-media-spend management fee: the fee is not tied to the media budget — maitiq does not earn more when your advertising spend rises. Because the quote is per account, the agreed scope of services remains the benchmark.

Why efficiency ratios such as CPL and CPA need a defined denominator

Efficiency ratios such as CPL or CPA are fractions. If their denominator is unclear, a lower numerator is not an improvement, but merely a smaller number. Total spend and fees, by contrast, are absolute amounts: a lower total cost can be a real saving even while efficiency remains unknown. Four documented platform properties affect almost every cost programme.

• Primary or secondary. Google uses primary conversion actions for bid optimisation and reports them in the “Conversions” column. Secondary actions serve observation, appear in “All conversions” and are not used for bidding unless they are part of a custom goal specifically included in a campaign.
• Counting method. One setting counts every conversion after an ad interaction; the other counts one per ad interaction and conversion action — not necessarily one per click. The same events produce different costs per conversion depending on the setting.
• Visibility in the search terms report. The report shows search terms that were used by a significant number of people and omits terms with too little activity for data protection reasons. Absence from the report is no evidence of absence in the account.
• Gaps in the change history. The change history covers two years at account, campaign and ad group level and names the user or tool. Not all account-level changes and not all changes by Google staff appear there.

So define the business goals and conversion actions that are relevant to your decisions, and document when they count as valid. Secondary actions and custom goals remain part of the picture when they are included in a campaign.

Repetitive work is a cost in its own right, not a side effect

A large part of ongoing costs arises from recurring work: the same analyses, the same reports, the same follow-up questions. Record this work cleanly once — task, frequency, duration, who performs it, output format — before you talk about tools or providers.

Automation initially changes only who performs this work. In your own worksheet that is a shift between cost categories and is subject to the same rule as any other measure: without quantified shifted work and associated costs and without a verified cycle, no saving is booked.

As an offer in the account management category, by contrast, coverage is negotiable: maitiq halves the management costs of a traditional Google Ads agency at a comparable scope of services. That is a statement by maitiq, not independent evidence; the comparison covers the service, not your Google media budget.

What is verifiable, by contrast, is traceability. Google’s auto-applied recommendations provide the test pattern: only a limited selection can be applied automatically, budget increases are not part of it in this iteration, the setting can be switched on or off at any time, and applied recommendations can be read in the change history. Ask for exactly these four details — scope, exclusions, switchability and auditability — for every automation in your account.

When does the waste diagnosis take over?

The reduction programme organises and steers; the row-by-row diagnosis of media waste is not part of it. The most important rule in the whole topic applies to that: a row without a conversion is first a row without a conversion, not waste. Measurement maturity, campaign role, relevance, materiality and possible conflicts with relevant searches or existing keywords come first.

The detailed method is in the article Identifying Google Ads waste. Take its findings as inputs into the following worksheet: as rows in the media spend category, with an evidence state and the named shifted work and costs.

The prioritisation worksheet

This worksheet is a manual template in two stages: copy it into your own worksheet or register and work through the two stages in order. It does not assign an estimated saving and calculates no weights or ranking; it answers the narrower and more honest question of which measure may be assessed at all, and in what order you work through the admitted measures. Whatever fails stage 1 is not ranked but clarified.

Stage 1 — Readiness check

Address all seven fields. A field may be marked “not applicable” with a justification — that is different from unknown and from zero. unknown is valid and routes the row into clarification rather than implementation when the information is material to the decision. Record “no shifted work and no additional costs” only once you have checked it; never fill missing data with zero.

Stage 2 — Order

You order admitted measures using four criteria. This page calculates no weights, thresholds or scores: you record which priorities you choose and why.

The result is an order, not a figure in francs. It says which action to address first — not how much you will save.

An illustrative starting order with a factual basis

The following starting order is illustrative and follows dependencies, not expected savings. Independent checks can run in parallel; the fee review and the control of uncontrolled exposure do not have to wait for the conversion data to be clarified:

1. Compare the scope of services and the fee formula. This lever can be checked even without robust account data.
2. Limit uncontrolled exposure. Without approval rules and a change trail, costs arise faster than an analysis can find them.
3. Clarify the efficiency denominator. An open conversion definition, counting method or time reference distorts efficiency comparisons such as CPL or CPA; it is not a prerequisite for fee and coordination questions.
4. Assign repetitive work. Only once it is clear which analysis is needed how often can its coverage be shifted.
5. Investigate documented irrelevant demand. Otherwise you risk cutting useful activity based on unreliable measurement.

This is a dependency logic, not a statement about the size of a possible saving.

The controlled reduction cycle

Every admitted measure runs through the same cycle of six fields. Define the baseline, the authorisation and the observation and verification plan before implementation; record the actual observed result, the verification and the subsequent decision afterwards.

There is no universally valid figure for the observation window. What fits volume, sales cycle and risk is appropriate — fixed in advance, not afterwards. Also record neutral and negative results.

Which levers reduce costs without cutting off valuable demand?

Start with the cause: a high CPC calls for a different answer than many unqualified leads. Four checks help with prioritisation:

Illustrative calculation example: CHF 3,000 in advertising spend with 60 leads gives a CPL of CHF 50. If only ten leads are qualified, a qualified lead costs CHF 300. After a change, the same CHF 3,000 produces only 40 leads, but 15 of them qualified: CPL CHF 75, qualified CPL CHF 200. The higher cost per form submission would not be a step backwards here. For efficiency comparisons, the figures need the same definitions, comparable periods and complete feedback. For an exact exclusion based on a verified qualitative mismatch or an explicit business exclusion, by contrast, a mature conversion window is not a prerequisite; a “strategy decision” here means documented business policy, not any change to a numerical bidding strategy. The example is not a customer result.

maitiq creates a recurring review process for this instead of a one-off cost-cutting exercise. Changes are justified and their effect is subsequently reviewed. Lower management effort and less wasted media budget remain separate goals: a fee reduction does not yet prove a better campaign.

How maitiq operates this lever

maitiq runs the cycle continuously for recurring account work: budget pacing and budget reallocation, target CPA adjustment, search term and exclusion maintenance, keyword and ad group preparation, ad review, and separately agreed preparation and coordination of CRM measurement, without automatically generated proposals and without upload to the Data Manager. Selected recurring workflows run daily by default; the cadence is configurable, and not every listed activity or import is daily. You receive proposals with the proposed action, the reason and the data basis; old and new values are given where they are meaningful. You approve manually, or you enable a bounded policy that can approve and trigger a separately authorised, logged implementation; an exact exclusion can be authorised for that scoped action with a single decision. The starting point is a read-only audit: it supplies findings from connected or uploaded available data and reports coverage and gaps. It does not automatically populate every field in this manual worksheet and cannot fill missing conversion data, contracts or shifted internal costs; shifted work and costs, approval and the agreed scope of services come from your company.

When you should not cut

• The observation period is shorter than your sales cycle.
• The conversion measurement is demonstrably incomplete for the intended efficiency comparison, or was changed recently.
• Other changes are running at the same time, which makes the result impossible to separate. A coordinated, authorised group of changes is not automatically an error, but the inability to separate the effects must be recorded.
• The measure is irreversible, but the evidence is only partial.
• The only reason is a metric without a business reference. The optimisation score, for example, is an estimate of how well an account is set up. It says nothing about your cost position.
• The measure is a cut to the average daily budget that you treat as an immediate spending stop. A budget reduction can be a valid control, but the nominal daily amount is not a hard stop: for most campaigns with an average daily budget, Google states a daily limit of twice that budget and a monthly limit of 30.4 times it, and it bills within those limits. The 30.4 factor assumes the budget stays unchanged through the month; later changes follow separate rules, and campaign total budgets and account budgets work differently. The budget and fee article provides the full distinction.

Missing evidence is neither a success nor proof of waste. Document the open point and the next safe check. “Not deciding yet” is a permissible result of the worksheet.

Frequently asked questions

Sources and classification

The sources are Google Ads help pages: they document the search terms report and conversion counting, negative keywords, budget and spending limits, primary and secondary conversion actions, change history, auto-applied recommendations, the budget report and optimisation score. They are platform documentation, not market-price or success evidence. Information on how maitiq works is available at maitiq.com.

Have your specific case reviewed by maitiq.