Google Ads usually fails at one of five layers: delivery, traffic quality, measurement, website conversion or business economics. The first job is not to change bids or increase budget. It is to identify which layer is actually failing.
An account getting no impressions needs a different fix from one getting clicks but no leads. A campaign producing leads but no customers may have a sales-quality problem rather than an advertising problem. Diagnose the symptom first, then change the account.
“Not working” is too vague to troubleshoot. Start with the first measurable symptom. If the account is not entering auctions, investigate delivery. If people click but do not convert, verify measurement before blaming traffic. If leads arrive but do not become customers, the problem may sit after the ad click.
Start with delivery and eligibility. Check campaign and ad status, billing, policy review, dates, budget, bid targets, location restrictions, search volume, conversion goals and auction conditions.
The account is entering auctions. Now inspect the relationship between the search query, keyword, ad message and offer. A low click response can also reflect stronger competitors or a weak reason to choose your ad.
Verify measurement first. If conversions are not captured correctly, the account can look worse than it is and Smart Bidding can learn from incomplete signals. Once tracking is verified, inspect search terms, message match, mobile experience, forms, checkout and the offer.
Lead volume and business value are different outcomes. Review search-query quality, lead qualification, the conversion action Google is optimising toward, sales follow-up and whether closed-sale or offline conversion feedback reaches the account.
Break CPA into traffic cost and conversion efficiency. A higher CPA can come from higher CPC, lower conversion rate, weaker search-term mix, a landing-page problem, auction pressure or lower-quality conversions.
Check revenue measurement and economics together. Review conversion value accuracy, CPA, average order value, margin, product mix and whether the reported revenue is commercially useful.
Start with change history and measurement before rebuilding. Look for tracking changes, budget or bid-target changes, demand shifts, competition, conversion delay and seasonality.
The same symptom can have more than one cause, but the cause usually belongs to one of five layers. This model helps you decide which evidence to inspect instead of treating every performance problem as a bidding problem.
Can the campaign enter enough relevant auctions?
Eligibility, budget, bidding, targeting, policy, search volume and auction dynamics.Are you paying for visitors who are likely to become customers?
Search terms, keywords, match types, locations, audiences and devices.Does Google know which actions actually matter to your business?
Conversion actions, Google Tag, GTM, GA4, Enhanced Conversions and conversion value.Can the website turn qualified traffic into an enquiry or purchase?
Landing page, offer, message match, UX, forms and checkout.Can the acquired customer be profitable at the current advertising cost?
CPC, CPA, AOV, margin, LTV, ROAS and profit.This page treats Google Ads as part of a wider acquisition system. If you need the full service structure rather than a troubleshooting guide, see Google Ads services.
The order matters because one bad signal can distort everything after it. If conversions are wrong, bidding decisions are being judged against unreliable data. If traffic is irrelevant, a landing-page test may not fix the problem.
Check whether the account records the actions that matter and whether duplicates, missing tags or incorrect Primary conversion settings are distorting performance.
A campaign optimised for form fills should not be judged as though every form fill is a qualified opportunity. Define the actual outcome before interpreting the account.
Review campaign status, policy, budget, bid targets, targeting and whether there is enough relevant search or product demand.
For Search, review search terms rather than judging the keyword list alone. For automated campaigns, inspect the available query, audience, asset and product-level evidence.
If the ad attracts one expectation and the landing page answers another, the campaign can have healthy clicks and weak conversion.
Check form completion, checkout behaviour, mobile usability, page speed, offer clarity and whether the visitor has enough information to take the next step.
A rising CPA caused by higher CPC needs a different response from a rising CPA caused by lower conversion rate.
The account is only commercially useful if the cost of acquiring a customer fits the margin and customer value the business can support.
Revenue is not profit. Product margin, discounts, shipping, returns and payment costs determine whether attributed revenue is actually useful.
The change should respond to evidence, have a clear reason and leave enough stability for you to measure what happened next.
Clicks without conversions usually point to one of three areas: the conversion is not being measured correctly, the traffic is less qualified than the click volume suggests, or the post-click experience is failing to convert qualified visitors.
I would verify measurement before judging the bidding strategy. If the conversion action is missing, duplicated or defined poorly, you are analysing a distorted account and possibly training automation toward the wrong outcome.
If the problem is specifically measurement, go deeper into ecommerce conversion tracking and measurement. For Search-specific query and campaign structure, see Google Search Ads.
A lead is not automatically a valuable conversion. If Google is rewarded for every form submission equally, the system can optimise toward people likely to submit forms rather than people likely to become customers.
The diagnosis should move beyond the ad platform. Review lead intent, qualification, sales response time and whether downstream outcomes are being fed back into the account where practical.
A campaign can reduce CPL by attracting more low-intent form submissions. The dashboard improves while the sales team spends more time on people who were never likely to buy.
The better optimisation target is not automatically the cheapest lead. It is the signal that best represents a commercially useful lead or customer.
A sudden decline does not automatically mean the campaign structure is broken. Start with measurement and account changes, then separate normal fluctuation from changes in demand, competition, bidding, budget, landing-page behaviour or product availability.
Look for conversion-action changes, tag changes, bid-target changes, budget adjustments, new exclusions, location changes or website updates that line up with the performance shift.
Review search demand, Auction Insights, CPC changes, search-term mix, seasonality and conversion delay before concluding that the campaign itself has failed.
Performance Max can underdeliver or underperform for different reasons. Treating every weak period as proof that PMax “does not work” skips the more useful questions about goals, signals, eligibility, budget, products and the time allowed for the campaign to stabilise.
Check whether the campaign is optimising toward the right conversion actions and values.
Targets that are too restrictive can reduce the campaign's ability to enter useful auctions.
Campaign, asset or product-level disapprovals can limit delivery before optimisation even becomes relevant.
For ecommerce, feed quality, product eligibility, availability and pricing can influence what the campaign can serve.
Review whether the campaign has usable creative coverage and whether traffic is being directed to appropriate pages.
Do not judge an automated campaign from a few isolated days after major changes. Compare performance over a meaningful conversion cycle.
If the issue is product-led, the deeper topic belongs on Google Shopping Ads and the ecommerce Google Ads architecture.
An ecommerce campaign can generate purchases and report positive ROAS while still losing money after product margin, discounts, shipping, payment costs, returns and customer acquisition cost are considered. Platform revenue is only one part of the commercial equation.
At a ₹50 CPC and 2% conversion rate, the acquisition cost is ₹2,500. If the average order value is ₹4,000 and gross margin is 35%, the gross profit before advertising is only ₹1,400.
The campaign can report revenue while the business still loses money on the first order. That is not solved by celebrating attributed revenue. You either improve acquisition efficiency, improve conversion, increase order economics, improve customer value, or change the product and campaign mix.
For the full ecommerce campaign structure, see Google Ads for ecommerce.
Weak performance often creates pressure to “do something.” The wrong response is to change several important variables before you know which one is causing the problem. That makes the account harder to understand and can reset the evidence you were trying to read.
If the account is buying poor traffic or measuring the wrong conversion, more budget simply scales the existing problem.
A short-term fluctuation is not enough evidence to introduce another major variable.
When everything changes at once, you lose the ability to connect a result to a specific intervention.
Smart Bidding can only learn from the signals you provide. The wrong conversion goal can make the system efficient at the wrong outcome.
Recommendations still need to be judged against your business goal, margin, conversion quality and campaign context.
Use the matrix to decide what you should inspect first. It is not a substitute for account context, but it prevents common shortcuts such as assuming every low-impression problem needs more budget or every low-ROAS problem is caused by bidding.
| Symptom | Likely Layer | Check First | Do Not Assume |
|---|---|---|---|
| No impressions | Delivery | Eligibility, targeting and bidding | The budget is automatically too low |
| Low impressions | Delivery | Search demand, budget and bid targets | Increasing bids is always the answer |
| Impressions but few clicks | Traffic / message | Queries and ad relevance | The landing page is the first problem |
| Clicks but no conversions | Measurement / conversion | Conversion tracking | The keyword is automatically bad |
| High CTR but low conversion rate | Intent / landing page | Query-to-page match | A high CTR means good traffic |
| Leads but poor sales | Lead quality | Conversion definition and lead intent | More leads will solve it |
| CPA increasing | Traffic / conversion / auction | CPC and conversion rate separately | The bid strategy alone is responsible |
| Low ROAS | Economics / measurement | Revenue accuracy and margin | Any universal ROAS target applies |
| Sudden performance decline | Change / market / measurement | Change history and tracking | The campaign needs rebuilding |
| PMax not spending | Delivery / automation | Goals, targets, eligibility and budget | PMax simply does not work |
An audit becomes useful when you cannot confidently identify whether the problem sits in measurement, traffic, campaign setup, landing-page conversion or economics. If the cause is already isolated and you have enough data to test a fix, you may not need an audit.
Start with campaign eligibility, status, policy, targeting, bidding, budget and search demand. A campaign that cannot enter enough relevant auctions needs a delivery diagnosis before you change ads or landing pages.
Verify conversion tracking first. If measurement is correct, inspect search terms, ad-to-landing-page message match, mobile experience, form or checkout friction and the offer itself.
Check change history, conversion tracking, conversion delay, budget changes, bid targets, demand, competition, search-term mix, landing-page changes and seasonality before rebuilding the campaign.
The issue can sit in traffic quality, conversion measurement, landing-page relevance, offer strength or the lead path itself. Diagnose those layers in that order instead of increasing budget immediately.
Review search intent, the conversion action Google is optimising toward, lead qualification, sales follow-up and whether downstream customer-quality signals are being fed back into the account.
Not until you know what is limiting performance. Increasing budget can help a healthy, budget-constrained campaign, but it can also scale irrelevant traffic, weak conversion or incorrect measurement.
Check conversion goals, tracking, budget, bid targets, eligibility, assets, Merchant Center data where applicable, product availability and whether the campaign has had enough stable time to evaluate performance.
Follow the sequence: delivery, traffic quality, measurement, website conversion and economics. If the ads deliver relevant traffic and tracking is correct but qualified visitors do not convert, the investigation should move beyond the campaign into the landing page, offer or conversion experience.
If you have checked several layers and still cannot identify why performance is weak, an audit should isolate whether the problem sits in measurement, traffic, campaign setup, landing-page conversion or economics.