Of the thirteen settings that decide where a Google Ads budget goes, six can drain it unnoticed. They're the conversion goal, Search Partners, Display expansion and Presence or interest, plus AI Max and auto apply. Find each campaign's live settings, fix conflicts with the business, then test the uncertain ones on qualified outcomes.
Switching everything off isn't a strategy, because a default or automated feature alone isn't evidence of waste.
Check the account before changing settings
Decide the scope first: which campaign, what it's meant to produce and the period being judged. Also decide whether the numbers are mature enough to judge.
Then write down what each control is set to today. Don't go by what the setup screen showed on launch day. Don't rely on what the person who built the campaign remembers choosing, either. Campaign and ad group settings, inherited account goals, shared budgets and account level automation can override that memory.
Google has also moved several of these settings itself since 2025. Each change is dated in the section it affects.
SourceGoogle Ads Help, 2025 to 2026
This is a settings review for Search led small business accounts. Where Performance Max works differently, the section says so. Performance Max for small business has its own guide. Building an account from scratch is a different job. That's covered in our guide to setting up Google Ads for a small business.
The checklist runs in the same order as the settings below:
what the campaign counts as success
where it can show
who it can reach
how it bids and spends
what can change without anyone approving it
Outcome
The conversion goal this campaign bids on matches an enquiry the business receives
Networks
Search Partners and Display expansion are each a decision, with outcomes segmented by network
Matching
Match types, AI Max search term matching and negative conflicts read together
Reach
Location options, brand controls, audiences, language and devices checked for the combined effect
Spend
Bid strategy objective, target value, average daily budget and the billed limits understood
Time
Schedule in the account time zone, call asset hours and the after hours route agreed
Automation
Auto apply categories, text customisation, URL expansion and account level assets each recorded
Conversion goals and bidding actions
The first setting is the one the other twelve serve. Open the goals this campaign uses, not the account's conversion action list.
Google's primary and secondary actions page says primary actions are used for bidding when their goal is selected. Secondary actions are reported in All conversions and not used for bidding. A secondary action inside a custom goal is used for bidding regardless.
A duplicated import is a tracking fault. So is a form event that fires on the button rather than the success. So is a value nobody can explain. Fix those with our Google Ads conversion tracking setup guide, then come back to the goal.
A successful form event must correspond to an enquiry the business received. A call click is not a completed call. A duration threshold is a proxy for quality and not a test of it. Genuine enquiries, qualified leads and customers are three different counts.
Selecting Maximise Clicks doesn't make measurement unnecessary either, because the clicks still need to become something. Verify: the campaign's goal lists only actions that mean an enquiry. Also check that a labelled test enquiry appears once.
Search Partners and Display expansion
Both settings sit under Networks in a Search campaign. Both widen where the ads can show.
Our audit sample is 180 Google Ads account audits from August 2021 to September 2026. In about 7 in 10 accounts, at least one live Search campaign used Search Partners or Display expansion. In about 2 in 5 accounts, a live Search campaign had both on. The same accounts had no network split in the business's reports. Those are counts of unexamined settings, not of waste.
Google Search Partners
Google's Search Network page says search partners are included by default when a Search campaign is created. They are removed by clearing Include Google search partners under the campaign's Networks setting. The page also says clicks from partner sites may not always reflect highly targeted traffic.
Search Partners is a different thing from the Display Network. Decide by segmenting outcomes by network over a mature period. Each network needs the same conversion definition, with lead quality attached.
The partner network has the lower raw cost per lead and the higher qualified cost per lead. That is a change of denominator, not evidence that partners usually perform worse. One qualified lead is not a sample to make a permanent decision on. Customer value and close rate are still needed before anyone calls either network profitable.
For partner sites, Google's placement report now lists where Search, Shopping and App ads were shown. The list sits in the Content suitability report. It shows impressions for sites with five or more impressions in the period. Sites with any click, view or conversion in that period appear too.
Google's Search Network page still says Google Ads gives no site detail, except for Performance Max. So Google's two pages don't fully agree. The listing is a distribution and suitability report, not a site level profit report. That's because it carries impressions rather than cost or qualified conversions. Verify: network segments compared on qualified outcomes and the Networks setting matching the decision.
Display expansion on Search
Display expansion is the Display Network option in a Search campaign. Google's Display expansion page describes it as a way to get additional conversions using unspent Search budget.
In the same breath, the page says the picture is different for budget consuming strategies. Those include Maximise Conversions, Maximise Clicks and Maximise Conversion Value. There, Display expansion shares the budget with Search when a similar or better CPA can be achieved. The majority is typically still reserved for Search.
The spare budget behaviour doesn't apply to shared budgets or campaign experiments. So Display expansion doesn't always spend only money Search could never have spent. Google's page also quotes up to 15% more conversions and an average of 14% more conversion value. Those are Google's promotional averages, not a forecast for any account.
For an intentional Search only baseline, untick the Display Network in the Networks setting. For an intentional Display role, decide whether a separate campaign serves the business better. It would bring its own budget and creative control. Our guide to Google Ads campaign types compares the options.
Display isn't inherently an error. Turning it off doesn't save a specific percentage either. Verify: the inventory the campaign served on and the outcomes by network.
AI Max for Search
Phrase and exact match do not mean literal only matching. Google's keyword matching page defines all three types by meaning. So the control over unsuitable searches is the search terms report and the negative list together. Our guides to Google Ads match types and Google Ads negative keywords cover both in depth.
The newer control is AI Max. Google's AI Max page describes it as a feature suite for Search campaigns. It's neither a campaign type nor Performance Max. Its search term matching adds broad and keywordless reach to your keywords. As the match types guide explains, the search terms report labels those matches as AI Max.
The defaults matter more than the description. Google's AI Max setup page says new Search campaigns have AI Max turned on by default. Search term matching must be on at the campaign level before an ad group can use it. Each ad group can then switch it on or off. Text customisation and final URL expansion are enabled by default, with the second depending on the first.
Campaign: [name and type]
AI Max state: [on or off, and whether it arrived by default, upgrade or choice]
Search term matching: [campaign setting]
Ad group exceptions: [ad groups with matching turned off]
Text customisation: [on or off]
Final URL expansion and URL rules: [on or off, exclusions, inclusions]
Date checked: [date and who checked]
Existing campaigns need their actual state checked. Google's text customisation page says campaigns using text customisation are automatically upgraded to AI Max from September 2026. Those defaults and upgrades are not the same thing as enrolling in auto applied recommendations. So neither "all automation is opt in" nor "the last manager must have switched this on" is a safe assumption.
Write down the AI Max state and whether it arrived by default, upgrade or choice. Note any ad groups with matching turned off and the date you checked.
Google's setup page says these features need a conversion based Smart Bidding strategy to work well. So before expanding reach through any of them, check four things:
a reliable conversion signal
a budget that can carry the extra traffic
destinations that suit it
someone who will read the search terms
Narrower matching is not always cheaper. Verify: the recorded state matches the served search terms and their source column.
Location, brand, audience, language and devices
These five controls decide who can see the ads. Each can be sensible on its own and still choke reach in combination. So check them separately, then together.
Location options and service area
Google's advanced location options page says Presence reaches people in or regularly in the targeted locations. Presence or interest, the default and Google's recommended option, adds people who have shown interest in them. Because location signals vary, 100% accuracy isn't guaranteed. Neither option is a live GPS fence.
Take the local service accounts in our audit sample. About 3 in 5 had at least one live Search campaign using Presence or interest.
The service location can differ from where the enquirer is standing. Someone interstate arranging work on a Melbourne property is a valid enquiry for a Melbourne plumber. Match the geography to fulfilment rather than to a rule that every interstate user is wasted spend.
Review actual lead addresses, the location reports and the exclusions. A Melbourne word in a query doesn't prove presence. Over narrow areas and conflicting exclusions suppress viable demand as surely as a state wide target wastes it.
Our guide to Google Ads location targeting covers radius targeting, exclusions and the other geography controls. Verify: where enquiries and jobs came from, against the location options and the targeted areas.
Brand inclusions and exclusions
Google's brand settings page says brand inclusions limit a Search campaign to queries associated with the selected brands. Exclusions stop it serving for queries associated with brands you want to avoid. Where both exist, exclusions take precedence.
Brand lists serve an explicit purpose: a brand only campaign, or one kept clear of brand demand. They aren't a brand safety switch. Nor are they a complete new customer filter, because a non brand campaign isn't automatically incremental.
From 27 May 2025, Google began moving Search brand settings into the campaign's AI Max panel. An ad group's brand inclusions override the campaign's. So check the scope of each list, not just whether one exists. A reseller should take care not to exclude the product brands it sells.
The negative keywords guide linked above covers how brand exclusions sit beside negatives, including in Performance Max. Verify: the brand lists attached and the queries the campaign matched.
Audience targeting and observation
Google's Targeting and Observation page says Targeting restricts the ad group's reach and Observation doesn't. It recommends Observation for all Search campaigns. First party audience segments added under Observation are used as signals for Smart Bidding.
An audience left on Targeting by accident suppresses demand quietly. Read audiences together with geography, brands and schedules, because the intersection is what the campaign actually reaches.
In Performance Max, Google's audience signals page says ads may show to relevant audiences outside your signals. So a signal is guidance, not a boundary. Any real exclusions live in that campaign's own settings. Verify: the Targeting or Observation flag on each audience, and reach after any change.
Language matching and supported languages
Google's language targeting page says Search campaigns lose the campaign level language targeting setting from September 2026. Search ads then match automatically on the language of the ads. Google also checks that the user understands the language of the creative and landing page. No action is required.
The same page says the setting no longer applies to Performance Max ads on the Search Network. But the selected languages still apply to its other channels. Shopping ads in Performance Max aren't affected by language settings.
The check that remains is operational. The ad, the landing page and the enquiry handling have to support whichever languages the business actually serves. Automatic matching does not translate the website or the person who answers the phone.
Verify: the languages in the served ads and pages match the languages the business can handle. Also confirm that channel specific settings are still correct where they apply.
Devices and supported adjustments
Investigate device performance, enquiry quality and the actual mobile journey before excluding anything. A broken mobile form is a repair job and not evidence that mobile users are poor leads. A sparse segment is unknown rather than failed.
The control is narrower than it looks. Google's bid adjustments page says manual bid adjustments on automated Smart Bidding strategies won't be supported. Target CPA device adjustments modify the CPA target rather than the bids. Maximise Conversions, Maximise Conversion Value, Target ROAS and Target Impression Share support only a minus 100% device adjustment.
The same page says Smart Bidding still considers targeting and exclusions. It also says ads strictly follow the days and times set in the ad schedule. Use Google's strategy specific table, not a universal percentage. Verify: the adjustment the strategy honours and the device split of qualified outcomes afterwards.
The combined restriction
A campaign can also become too narrow through the intersection of individually sensible controls. When reach drops, inspect these together:
audiences
brands
geography
devices
hours
negative conflicts
Separate verified business exclusions from performance guesses. Keep the least restrictive setup that still matches the business and the measurement.
Campaign scope and business boundary: [what the campaign is for and the real service limits]
Location option and areas: [Presence or Presence or interest, targeted and excluded areas]
Brand and audience controls: [inclusions, exclusions, Targeting or Observation]
Language and device behaviour: [languages served and the device control the strategy honours]
Combined restriction to verify: [the intersection of all of the above and how reach was checked]
Bidding targets, budgets and schedules
Three things get confused here: the bid strategy, its target and the budget. The schedule then decides when any of it can be spent.
Bidding targets
Check five things:
the strategy's objective
the conversion data it receives
the target value
any portfolio or shared budget relationship
the total spend the business authorised
A tight target restricts opportunities. A stale permissive target matters just as much.
Nobody has to run a Manual CPC phase or reach a conversion threshold before a strategy is allowed. Choosing the strategy itself is covered in our guide to Google Ads Smart Bidding.
Google's target bidding update describes a change starting 17 August 2026. It says target based strategies now deliver more consistently against the target set. That includes budget limited campaigns that had been beating their targets. Their results will move towards the target. Google won't automatically adjust targets or budgets.
A campaign running at half its target CPA is the one to review now. Verify: the target against the business economics and mature results.
Budget limits and pacing
Google's spending limits page sets two caps on billed spend for most campaigns. On any one day, the cap is twice the average daily budget. Over a month, it's 30.4 times that budget.
Google's pacing page covers campaigns whose schedule skips whole days. Since 1 June 2026, they still pace towards that monthly limit across their active days. But the daily limit can cap a short schedule.
Budget edits, mid month starts and setting the budget in the first place belong elsewhere. Our guide to how much to spend on Google Ads covers them. What a budget buys at current click prices is in how much Google Ads costs in Australia. Verify: billed spend against the daily and monthly limits.
Ad schedules and call availability
Google's ad schedule page says schedules run on the account's time zone. Ads won't show outside the times listed. Customers in another state need the hours converted, daylight saving included. The location targeting guide covers that conversion.
Campaign schedules govern ad eligibility. Call assets can carry their own hours. Google's assets page says supported assets can be scheduled. Its example is call assets limited to the hours the business can take calls. Scheduling the asset doesn't stop someone calling the number on the website.
A closed office doesn't mean no conversion opportunity. That's true for any business with a form, a booking tool or an honest callback promise. Test the after hours route. Then read answered call rate, response delay and qualified outcomes by hour and day before cutting anything.
Restrict calls or campaigns when capacity warrants it. If whole days come out, remember the June 2026 pacing change. Our guide to call tracking for Google Ads covers measuring which calls were answered and useful. Verify: test the account time zone, the schedule, the call asset hours and the after hours route.
Auto applied recommendations
Google's auto applied recommendations page says you enable the feature by selecting recommendation types at account level. You can select them in bundles or individually. Auto applying recommendations won't increase your budget. The History tab summarises recently applied changes. The Add responsive search ads recommendation stopped being suggested or applied automatically from 26 January 2026.
The types that can be applied include:
target and bidding changes
broad match keywords
audiences
search partners
Display expansion
So the budget stays put while the way it is spent moves. Review the selected categories, the queue, the history and who approved the enrolment.
In our audit sample, about 1 in 3 accounts had at least one auto apply category switched on. In about 1 in 4 accounts, the owner, asked at the audit, didn't know any category was enrolled.
SourceElev8d Google Ads audit sample
Scaled to 100 accounts and rounded, that's 33 enrolled. In 25 of them, the owner didn't know. Keep this feature separate from AI Max, account level automated assets, scripts, automated rules and manager account activity. Turning off auto apply doesn't turn those off or undo earlier changes.
Judge each recommendation by the campaign's purpose and not by the promise of a higher optimisation score. Verify: the auto apply selections, the History tab and the account's change history.
Generated assets and landing page URLs
Two automations shape the ad the person sees, each with its own switch. Google's text customisation page says the feature was formerly known as automatically created assets. It generates headlines and descriptions from the domain, landing page, existing ads and ad group keywords. They're refreshed at least every 48 hours.
Google's page points to the ad level asset report and its Added by column for review. Keep the website's claims current, because the generator reads them.
Account level automated assets are different again. Google's assets page says some are added automatically when Google predicts they'll improve performance. Specific types can be turned off in the account level automated assets advanced settings. Our guide to Google Ads assets covers each type.
Destinations have their own control. Google's final URL expansion page for Search says the feature sends traffic to the domain's most relevant URLs. It takes URL exclusions at the campaign level and URL inclusions at the ad group level.
The setup page's matrix adds two facts that catch people out. With inclusions on and expansion off, the ad group uses only the specified URLs as dynamic landing pages. When URL inclusions are enabled, pinned assets aren't used.
Pinning isn't a universal protection. Google's final URL expansion page spells out when pinned assets always serve. For that, both final URL expansion and URL inclusions have to be off. Turning off a parent feature changes which subordinate rules operate. So record the actual state and verify the result.
Google's AI Max page also warns about tracking templates that aren't set up compatibly. They can send people to a broken expanded landing page with a 404 error.
A saved exclusion isn't proof the journey works. Read the generated message and the destination that actually served. Don't assume every homepage is unsuitable or every blog URL is waste.
Write the message you want with our guide on how to write Google Ads. Choose the destination with our Google Ads landing page vs homepage guide. Verify: the asset report and the landing pages report. Finish with a test click through the tracking template to the enquiry.
Choose settings for your campaign readiness
Readiness isn't account age. It's data quality, fulfilment capacity, spend tolerance and the capacity to review. A new campaign doesn't mean every control off. A mature one doesn't mean every control on.
Situation | Sensible starting position | What justifies a change |
|---|---|---|
Tracking or delivery not yet tested | Fix measurement and the enquiry route, keep exposure proportionate to what can be verified | An end to end test and a defined useful outcome |
New campaign with limited evidence | An intentional reachable market, chosen networks and a review scope the owner can manage | Qualified outcomes and enough evidence to evaluate an expansion |
Mature campaign with reliable outcomes | Retain useful automation and inventory, review the actual constraints | A documented hypothesis, available budget and a bounded evaluation |
Business cannot handle the resulting demand | Align hours, service areas, offers and routing with capacity | A functioning handling process and commercially useful outcomes |
Website or platform controls have changed | Check assets, destinations, tracking and the effective settings again | Verified behaviour after the change, not a saved setting assumption |
Correct a verified conflict today. Anything else is a test that needs outcomes you can measure.
Run a settings check and verify changes
A short first pass on the highest spending campaign flags obvious mismatches. It can't establish every cause or calculate a reliable saving. An enabled feature isn't quantified waste. Verified unsuitable spend is counted once, even when three settings overlap on it.
Record each of these:
the objective
the selected goal
the networks
the reach controls
the budget
the target
the schedule
the automation
the destinations
Mark the unknowns as needing verification rather than guessing. Our audit scorecard extends this pass to the wider account. But because it can't see the account, its score isn't measured profitability.
Before launch or a major change, verify the effective settings and run an enquiry test. In the early weeks, check reporting and unexpected changes often. Then judge performance only after enough relevant data and the conversion delay. Our guide to the first 30 days of Google Ads sets out the timing.
Review again after website updates, ownership changes and platform migrations, of which 2026 has supplied several. Correct verified harmful errors at once.
For an uncertain performance choice, record:
one main hypothesis
the baseline
the change
the owner
the review window
the rollback condition
Scope, current state and evidence: [campaign, the setting as active and what showed it]
Hypothesis, change, owner, date and previous value: [what should improve, the edit, who, when and the value before]
Spend boundary, review window and rollback condition: [allowance, dates and what triggers reversal]
Qualified result, competing explanations and next action: [outcomes after lag, what else changed and the decision]
The record only helps if the test is read fairly. Four mistakes turn a settings test into a guess.
Changing several settings in one go
Nobody can tell which edit moved the result
Change one setting per campaign where practical, and log each edit with its date
Trusting the saved setting
An ad group override, a parent feature or a Google upgrade can change what actually serves
Read the search terms, networks, served assets and destinations after the change
Judging on raw leads in the first week
Google's conversion reporting dates each conversion to the click, so recent days fill in later, and raw counts hide quality
Wait out the conversion delay, then compare qualified outcomes
Reading a before and after as proof
Seasonality, demand shifts and other edits move results too
Compare like periods, note what else changed and keep the rollback condition
When a settings review doesn't explain the results, the problem sits somewhere else in the chain. Our guide to why Google Ads isn't working routes the diagnosis.
What we recommend at Elev8d
Start with the goal. If the campaign bids on a duplicated import or a button click, the bidding chases the wrong thing. No other setting fixes that, so we correct it first on the Google Ads accounts we run.
Then we make the network call. For a small Search account, we turn Search Partners and Display expansion off. That gives a clean Search baseline. Once qualified leads are counted, we test both back one at a time, each with a spend limit.
For a business that serves a local area, we set Presence rather than Presence or interest. The exception is when customers often arrange the job from somewhere else.
Next comes automation. We switch off any auto apply category nobody can justify and write down the AI Max state. AI Max search term matching stays on only where there's a reliable conversion signal. Even then, someone has to read the search terms every week.
A mature account with reliable outcomes is different. There, switching things off throws away what works. So we review the actual constraints. Then we test expansion deliberately, the way our guide on how to scale Google Ads sets out.
We'll look at the live settings and tell you which ones we'd change first.
FAQ: questions that need a little more context
Does a higher optimisation score mean the settings are right?
No, it doesn't. Google's optimisation score page calls the score an estimate of how well the account is set to perform. It also says an account can reach 100% by applying or dismissing all recommendations. So a high score can mean the list was cleared, not that the settings suit the business. Google also says the score isn't used by Quality Score.
Will changing a setting reset Smart Bidding's learning?
Only some changes trigger it. Google's learning period page lists three. One is a new or reactivated bid strategy. Another is adding or removing campaigns, ad groups or keywords. The third is a significant change to the conversion goals the strategy optimises for. The page says budget, target and targeting changes never trigger it.
After a trigger, performance can take up to one to two conversion cycles to stabilise. That can be faster with more conversion data.
A network or location change isn't on that list. But it still changes the traffic the strategy bids on. Google also says its algorithms keep learning after the Learning status clears. So judge any change on lagged, qualified results rather than the status label.
Can I see who changed a setting, and undo it?
Mostly. Google's change history page covers the last two years and shows who made each change. It includes changes made through automated rules and the API.
Most types of change made in the last 30 days can be undone. But an undo can fail if a related item has since been removed. Past 30 days, putting a setting back is a manual job. That's why the change record keeps the previous value.
Next steps
Record the active state of all thirteen controls
Start with the highest spending campaign. Record each setting at the level where it lives, with the date and who checked.
Fix the conflicts, then pick one hypothesis
Correct anything that contradicts the business today. Choose one uncertain setting to test, with a limit, a window and a rollback.
Verify that serving changed
Read the search terms, networks, locations, served assets and destinations after the change. Then compare qualified outcomes after the lag.
