Troubleshooting Common PPC Issues: A 7-Step Diagnostic Guide

Start with a symptom, not a prescription

“Performance is down” is not yet a diagnosis. It might mean fewer impressions, a weaker click-through rate (CTR), a lower conversion rate (CVR), or a higher cost per acquisition (CPA). Those symptoms can point in different directions. Fewer impressions may reflect budget or eligibility; a stable click volume with fewer conversions suggests a post-click or measurement issue; a higher CPA alongside a new mix of expensive queries may be an auction or targeting problem.

Before making an edit, write down what changed, when it changed, and which metric moved first. Compare a fixed recent period with a sensible baseline—for example, the latest 14 complete days against the preceding 14 days. For campaigns with low conversion volume, widen the window to 28 or 56 days. Exclude partial days and account for weekday patterns. A Monday-to-Sunday comparison is generally more interpretable than comparing a busy weekend with a quiet Tuesday.

Do not treat every variance as a meaningful decline. A campaign with four conversions last week and two this week has lost half its recorded volume, but two events are too few to support a confident story about the underlying rate. The right response is to inspect the evidence and its scale, not to panic at the percentage change.

The seven-step diagnostic flow

1. Confirm that conversions are still being measured

Begin with the event itself. Check whether the conversion action is still active, whether its status has changed, and whether the tag or server-side event fires on the intended confirmation page or action. Review recent releases to the website, tag manager, consent platform, checkout, and CRM integration. A form redesign or a change to a thank-you URL can interrupt measurement without creating an obvious error in the ad account.

Compare the platform’s conversion count with a second source, such as a CRM, payment system, or analytics report. They will not necessarily match exactly: attribution windows, modeled conversions, time zones, deduplication, and consent choices can produce legitimate differences. The useful question is whether the gap changed abruptly. If Ads reports 38 leads while the CRM contains 37, that may be normal. If Ads reports 38 while the CRM contains 11—and the discrepancy began on the day a tag was changed—investigate the implementation before touching the campaign.

• Check the conversion action’s recent recording status and diagnostic messages.
• Test the journey from ad landing page through completion, on desktop and mobile.
• Confirm that one completed action is not counted twice through both a browser tag and a server event.
• Compare event timestamps and account time zones before matching daily totals.

For lead generation, distinguish a submitted form from a qualified lead if both are available. Optimizing toward an easy-to-trigger event can make reported CPA look healthy while sales quality deteriorates. Keep the primary optimization event aligned with the business outcome the campaign is supposed to produce.

2. Segment the decline before explaining it

Aggregate campaign totals hide where performance moved. Break the data down by campaign, ad group, keyword or search term, device, geography, network, audience, and landing page. Compare each segment with the same baseline used for the overall account. The aim is not to slice until something looks alarming; it is to find the narrowest segment that explains a material share of the change.

Suppose spend rises 24% and conversions stay flat. If one mobile landing page accounts for nearly all the added spend and its conversion rate fell from 4% to 1.8%, the problem is more likely concentrated there than in every keyword. Conversely, if conversion rates are stable but impression volume has shifted toward broad, high-cost queries, the query mix deserves attention.

Use absolute counts alongside rates. A 50% decline in a segment with two conversions is less persuasive than a 15% decline across several hundred. When a segment has fewer than roughly 20 conversions, treat its rate as noisy and look for corroborating evidence—such as a broken page, a disapproval, or a clear change in search terms—before making a large structural change.

3. Inspect search terms and targeting quality

Search terms reveal what people actually asked for, not merely the keywords an advertiser intended to buy. Sort terms by cost, conversions, and spend without conversions. Look for irrelevant intent, research-only queries, competitor names, job seekers, support requests, and geographic mismatches. A keyword can remain unchanged while the searches it matches shift with query demand and match behavior.

Do not add a negative keyword solely because a query has spent money without converting. A query with $12 of spend and no sale has little evidentiary weight if the typical conversion costs $90. A more useful screening rule is to review terms that have spent at least one target CPA without a conversion; consider stronger action around two or three target CPAs, provided tracking works and the term is clearly misaligned. These are investigation thresholds, not automatic laws. High-margin products, long sales cycles, and small accounts require different tolerances.

Check that match types, locations, languages, and audiences still reflect the intended reach. In location settings, distinguish people physically present in a target area from people merely interested in it, where the platform offers that choice. Review whether a campaign’s geography was broadened during an account change. A campaign aimed at local service calls can acquire expensive, irrelevant clicks if location intent is wider than the business can serve.

4. Separate auction pressure from budget and rank limits

Rising CPC does not prove that bids are too low—or too high. Auction conditions, competitor activity, quality signals, device mix, and query composition all affect the price. Review impression share, lost impression share due to budget, and lost impression share due to rank, where available. These metrics help distinguish a campaign that is unable to spend its budget efficiently from one that is constrained by budget or ad rank.

If search lost impression share due to budget is high and the campaign is meeting its CPA target, additional budget may recover valuable volume. If it is losing share due to rank while CPA is already above target, simply raising bids may purchase more of the same unprofitable traffic. First inspect query quality and relevance. A rank problem can be a consequence of weak ad and landing-page experience, not a request to spend more.

Check budget pacing across the full period. A campaign that exhausts its daily budget early may miss later auctions; a recent budget reduction can create a sudden impression decline that resembles falling demand. Conversely, a larger budget can change the mix of traffic the system enters. Compare spend, clicks, CPC, and conversion rate together. The pattern matters more than any single auction metric.

5. Evaluate CTR, ad relevance, and the landing-page experience

CTR is useful when interpreted in context. For a non-brand search campaign, a CTR below 2% can be a reasonable investigation flag, but it is not a universal failure threshold. Brand campaigns often attract substantially higher CTR; broad queries, display inventory, and some social placements may produce lower rates by design. Compare like with like: same network, campaign type, query intent, device, and period.

If CTR falls, check whether the search-term mix broadened, competitors introduced a stronger offer, or the ad no longer answers the query. Review the headline, description, sitelinks, price or promotion claims, and destination URL. A high CTR is not automatically good either. If clicks rise while qualified conversions fall, the copy may be attracting curiosity rather than the audience the offer can serve.

On the landing page, test the complete path on a real phone, not only in a desktop preview. Confirm that the headline matches the ad’s promise, the primary action is visible, the form submits, and no redirect, consent layer, or script error blocks the journey. A page that takes several seconds to become usable on mobile can lose high-intent visitors before they reach the offer. Compare page engagement and conversion rate by device, but remember that analytics engagement is diagnostic evidence, not a substitute for a completed business outcome.

Google’s Quality Score is a keyword-level diagnostic, not a campaign performance target. Its three components—expected CTR, ad relevance, and landing-page experience—can point toward an area to investigate. A Quality Score of 4/10 may be a useful clue; it does not mean the account is failing, nor does increasing it guarantee a lower CPA. Focus on the underlying relevance and user experience rather than trying to optimize a displayed score in isolation.

6. Diagnose ad disapprovals and policy restrictions

When a campaign’s reach drops suddenly, inspect policy and eligibility status before rewriting every ad. In the platform interface, identify which asset or destination is affected, the stated policy reason, and whether the issue is a disapproval, a limited serving restriction, or an account-level problem. Review recent edits to copy, images, claims, landing pages, and tracking redirects. A compliant headline can still lead to a destination that triggers a policy review.

Make the smallest correction that addresses the cited issue. If a claim needs substantiation or qualification, revise the claim and ensure the landing page says the same thing. If the problem is a broken or inaccessible destination, repair that destination rather than replacing compliant copy. Submit an appeal only after checking the policy and confirming the corrected page is live. Repeatedly resubmitting an unchanged asset does not make the underlying issue clearer.

Restrictions can affect volume unevenly. A single disapproved ad may leave other ads eligible, while an account-level verification issue can limit delivery more broadly. Check status at both the asset and account levels, and record the exact message and timestamp. Policy reviews and appeals can take time, so keep a compliant alternative available when campaign continuity matters.

7. Change one variable, document it, and wait for evidence

After locating the likely cause, choose one intervention. If irrelevant queries dominate spend, add a small, well-justified set of negatives. If a mobile page fails, fix the page. If tracking broke, repair and validate it. Avoid simultaneously changing bids, match types, budgets, landing-page copy, and conversion settings. A recovery after five edits tells you that something changed; it does not tell you which change helped.

Log the edit, its rationale, the segment affected, and the date. Set a review window appropriate to volume and conversion delay. A high-volume retail campaign may show a useful directional signal within several days; a B2B campaign with a long sales cycle may need weeks before lead quality matures. Do not evaluate a conversion action solely on the day of the click if conversions commonly arrive later.

For formal creative or landing-page comparisons, avoid declaring a winner from a handful of conversions. A simple two-sided test at a 5% significance level and 80% power often requires substantial traffic. As an illustration, distinguishing a conversion rate increase from 3% to 4% can require roughly 5,300 visits per arm under a basic two-proportion calculation. Actual needs depend on the chosen method, baseline rate, test allocation, and minimum effect worth detecting. If the account cannot reach that volume, report the result as directional rather than statistically conclusive.

Keep tests isolated where possible. If the platform’s automated bidding system is learning from concurrent changes, a short-term CPA swing may reflect reallocation rather than a stable treatment effect. Use an experiment feature or a controlled split when available, preserve the same audience and time window, and resist stopping as soon as one variant leads. Repeatedly checking results and stopping at a convenient moment increases the chance of calling noise a win.

Reading common PPC symptoms

A compact troubleshooting worksheet

Use the following record for each investigation. It makes the diagnosis reviewable by another marketer and discourages a vague “we optimized the campaign” account of what happened.

• Symptom: Name the metric and its absolute change, such as CPA moving from $72 to $96.
• Comparison window: Record dates, account time zone, and any conversion-lag adjustment.
• Measurement check: Note platform conversions, CRM or transaction count, and known discrepancies.
• Largest affected segment: Identify campaign, query, device, region, or landing page and its share of spend or lost conversions.
• Evidence for the cause: Include query examples, status messages, page tests, or auction metrics.
• Single intervention: State exactly what changed and what remained untouched.
• Decision date: Set a review date that accounts for conversion volume and lag.
• Outcome: Report the result, including uncertainty and any trade-off in volume, quality, or cost.

Worked example: A home-services search campaign’s reported CPA rises from $84 to $118 over 14 days. Spend increases 18%, but clicks rise only 3%. The first comparison with the CRM reveals that recorded form leads fell on the same date the mobile quote form was updated. A test submission on an iPhone reaches the confirmation screen, but the conversion event no longer fires. After the event is repaired and verified, Ads records 21 conversions over the next comparable period; the CRM records 20. The apparent CPA returns toward its prior level. Raising bids or adding negatives would have treated a measurement break as an auction problem.

When not to change the campaign

Sometimes the best diagnostic action is to wait. Hold changes when the signal is based on very few conversions, the comparison period is incomplete, a major promotion has distorted demand, or conversion lag has not matured. Waiting does not mean ignoring a broken form, a policy restriction, or obviously irrelevant spend. It means separating a concrete fault from ordinary variance.

Also be cautious when multiple campaigns share budgets, audiences, or automated bidding strategies. A change in one place can redistribute spend elsewhere, making a neighboring campaign appear to deteriorate. Record account-level edits alongside campaign-level ones, and inspect change history before assigning a cause. The timeline often contains a more useful clue than the latest dashboard snapshot.

Make the next decision easier

Reliable PPC troubleshooting is less about finding a clever bid adjustment than preserving a chain of evidence: measurement is intact, the decline is localized, the likely cause is supported by data, and the intervention can be evaluated on its own. A search CTR below 2% may justify inspection; it cannot, by itself, tell you to rewrite the ads. A low Quality Score can direct attention to relevance; it cannot establish that a campaign is unprofitable. A high CPA can be a real business problem—or a conversion tag that stopped firing on Tuesday.

Keep a change log, use fixed comparison windows, and make the smallest defensible edit. That discipline protects both budget and learning. More importantly, it leaves the account in a state where the next person can tell what happened, rather than inheriting a cluster of simultaneous changes and a story built after the fact.