HELP & REFERENCE
Find a clear answer.
Bots, IVT, crawlers and suspicious visits
- Are bots visiting my website?Look at the visitor's browser and network evidence, the classification reasons and the recorded journey. A suspicious label warrants investigation, not automatic blocking.
- What counts as invalid traffic (IVT)?IVT concerns advertising interactions that should not be treated as valid. Automated traffic and paid-click IVT overlap, but not every crawler request is an invalid paid click.
- Are bounced visits evidence of bots?No. People leave quickly because of intent, landing-page mismatch, poor performance or missing tracking. Compare cohorts and observed paths before attributing short visits to automation.
- Is this Googlebot, an AI crawler or someone pretending to be one?Inspect available crawler classification evidence and request patterns. A user-agent string alone cannot verify a crawler's identity or show whether a page was indexed.
- Is someone scraping our products or prices?Repeated product-page traversal can suggest automated browsing; it cannot alone establish what data was extracted or by whom. Compare visit behavior with authorized crawlers and normal shopping.
- Do multiple visits from one IP mean one person?No. NAT, mobile networks, offices, and proxies can place many people behind one address. Treat IP-level links as context, not person-level identification.
- Why did Button flag this visitor?Inspect the recorded reasons, available browser/network evidence and observed journey. Missing or conflicting signals sometimes leave the classification unresolved.
- Can suspicious-looking visitors still become customers?Yes. A proxy flag or unusual browser is not a commercial-intent test. In one ecommerce investigation, residential-proxy indicators alone did not justify excluding the whole group.
- Is form spam the same as bot traffic?Not necessarily. A flagged visit does not prove a submitted form is spam. Use server-verified form protection when the question concerns submissions.
- Should I block VPNs or residential proxies?Not solely based on the network type. Compare observed progression, examples and false-positive risk before choosing a rule.
- Can browser fingerprinting detect advanced automation?Browser inconsistencies can contribute evidence, but sophisticated automation can imitate normal browsers. No single fingerprint establishes bot identity.
- What if the evidence is inconclusive?Keep observing, check collection coverage, investigate a bounded segment and defer intervention until the risk is better understood.
Campaigns, paid-click quality and the customer journey
- Why are Google Ads clicks rising without sales?Investigate tracking, targeting, landing-page experience and traffic quality. Compare recorded journeys, then validate sales with your order or CRM records.
- How do I check for Meta Ads click fraud?Find paid visits with supported Meta identifiers, inspect automation evidence and compare subsequent journeys. Button does not determine Meta billing credits.
- Which campaigns should I investigate first?Compare tagged sources and campaigns against the site baseline. Use both counts and rates; an alarming percentage from a tiny sample can be misleading.
- Are these bots or just the wrong audience?Low engagement is not proof of automation. Review technical evidence and what the visitor did after the click before blaming fraud.
- Why did my campaign performance suddenly change?Compare like-for-like observation windows, traffic mix, site changes and measurement coverage. Investigate whether the traffic itself changed or the reporting changed.
- What changed since last week?Compare equivalent time windows, volumes, goal-path progression, classifier versions and collection gaps. A change in observed rates is a finding, not a complete cause.
- Why do paid visitors stop at the landing page?Check the actual arrival and subsequent path, site friction, page content and automation evidence. A landing-only visit is a starting point, not a fraud verdict.
- Are platform conversions the same as actual customers?No. A tagged conversion event, goal-page view, submitted form and completed order measure different things. Independent business records establish actual outcomes.
- Could blocking a suspicious segment lose buyers?Yes. Look for counterexamples and comparable visitor behavior before using any exclusion or tag rule.
- What can I tell a client about suspicious traffic?Report the affected campaign, observed pattern, sample size, supporting evidence, competing explanations and a clear next check. Do not label a suspected-click share as recovered money.
- Should I pause a campaign because of suspicious visits?Traffic-quality findings alone rarely settle the budget question. Investigate measurement and landing paths, then combine findings with campaign economics.
- Is PPC fraud protection the same as all ad fraud protection?No. Brand impersonation, creative theft and other ad fraud categories are separate from investigating visits that reach your site.
- Do you provide an automatic click-blocking API?Button provides traffic collection and supported integrations, but its investigation and on-site tag controls should not be confused with generic IP exclusions or refunds.
- Does Performance Max send real leads?You can inspect tagged visits and observed goal paths where identifiers are available. Confirm lead quality and purchase outcomes in CRM or commerce records.
- What can PPC managers do about suspicious paid traffic?Start with evidence, then review targeting, protect forms, audit conversion goals, investigate invalid clicks and measure verified business outcomes.
- How can I see what visitors do after clicking my ads?Use available click identifiers and recorded journeys to inspect visited pages and observed goal progression. Coverage depends on instrumentation.
Conversion signal feedback and configured tag controls
- Can bots trigger conversion pixels?Yes. Automation may reach a conversion-tagged page and fire configured events. Inspect the tag trigger and the visitor evidence first.
- Can poor-quality conversions affect Smart Bidding?Optimization depends on configured conversion signals. Button can inspect visit evidence but cannot establish what a specific event did to bidding.
- Does filtering tags recover ad spend?No. Withholding an on-site tag does not undo a paid click or issue credits.
- Could filtering hide legitimate buyers?Yes. An overly broad classification rule may suppress genuine prospects' signals. Inspect examples and scope controls carefully.
- How can I tell whether changing conversion feedback helped?Check both eligible tag events and mature external business outcomes. Lower pixel counts alone do not demonstrate improvement.
- Can Smart Responder change my Google Ads bids?No. It can control supported tag loading on configured website paths, not the bids or exclusion lists inside an advertising account.
- Does a thank-you page visit prove a sale?No. Goal-page activity is not an independently verified purchase or qualified lead.
- What happens if the original pixel still loads?A separate or duplicate loader may fire outside the managed integration. Check all tag sources on the page before relying on a control.
- Should I turn on signal suppression immediately?Start in observation mode. Review classification evidence, counterexamples and the paths and tags the rule would affect.
- Will fewer conversions mean Smart Bidding improved?No. Fewer reported events can simply mean fewer tags fired; improvement needs separate evidence of lead quality or sales.
An evidence-first guide to wasted spend, click-fraud investigations, website safeguards and conversion quality, with named PPC practitioners and primary sources.
- Search terms and negativesAction: review search terms and apply negative keywords, match types and landing-page changes where the evidence shows irrelevant intent. Compare cost and downstream behavior, not just bounce rate.
- Locations, placements and networksAction: audit geographic presence settings, excluded regions, placements and partner networks; check whether campaign type supports each control. Exclude identifiable poor-quality inventory where its actual outcomes justify the lost reach.
- Audiences and IP exclusionsAction: consider narrowly targeted audience restrictions or IP exclusions only after identifying a repeatable problem and estimating false-positive costs. Rotating IPs and shared networks make IP lists unreliable as a general solution.
- Budget and campaign segmentationAction: investigate branded versus nonbranded demand, channel mix, product margins, conversion volume and campaign learning before splitting, consolidating or pausing campaigns.
- Form verification and proportionate defensesAction: add server-side verification to high-value forms; consider WAF rules, rate limiting, bot scoring, honeypots and proportionate challenges according to observed abuse. Evaluate conversion friction and accessibility.
- Primary conversions and duplicate tagsAction: inventory conversion actions, triggers and duplicate tags; decide which outcomes should drive bidding, which should remain secondary and how server-verified business outcomes will be reported.
- Qualified leads and real outcomesAction: connect click identifiers and consent-appropriate attribution to CRM qualification, verified calls, accepted leads, orders and revenue. Investigate duplicates and spam before reporting form submissions as success.
- Conversion value adjustmentsAction: where the advertising platform supports it, use verified downstream outcomes to restate conversion values or retract conversions. Preserve identifiers, eligibility windows and audit logs; don't convert unconfirmed suspicion into an automatic retraction.
- Investigate invalid clicks and creditsAction: review the platform's invalid-click and credit reporting, then document bounded suspicious periods, click IDs, campaign context, network and browser evidence, and affected lead outcomes before escalating.
- Controlled mitigation and rollbackAction: record baseline rates and sample sizes, choose one scoped intervention, measure observation coverage, cost, qualified outcomes, false positives and side effects, and retain rollback criteria.
- Quick mitigation checklistAd platform: negative keywords; search-term analysis; location and presence settings; placement and search-partner exclusions; audience restrictions; device and schedule review; campaign segmentation; budget reallocation.
- How Button supports the managerInvestigate: connect paid-click identifiers when present to recorded visitor journeys and available browser/network classification evidence. Show competing explanations and counterexamples.
Button Solutions Knowledge Base Evidence before action.