How to Reduce Disputes on Pay-Per-Call Campaigns

Dispute reduction does not mean eliminating legitimate disputes.

It does not mean asking buyers to pay for calls that clearly fell outside the agreed campaign rules. It does not mean asking publishers to accept unexplained deductions, retroactive standards, or a month-end spreadsheet labeled “bad quality.”

It means building a campaign in which both sides know what should happen before traffic starts, preserve useful evidence while calls are flowing, and can explain every material adjustment after the calls occur.

That is the practical standard.

A dependable campaign makes the qualification rules clear before launch, separates consumer-initiated inbound calls from transfers, tracks traffic at the source and sub-source level, documents technical and operational failures quickly, and reconciles buyer credits and publisher payout corrections without rewriting history.

Publishers benefit from that discipline because valid calls are easier to defend. Buyers benefit because invalid calls are easier to identify. Operators benefit because the review can be based on call-level facts instead of bargaining power.

This guide is organized around three stages:

1. Before launch: define the campaign and the dispute process.
2. During operation: preserve evidence and correct recurring problems quickly.
3. After calls occur: review, adjust, and settle in a way both sides can explain later.

This article is educational and operational. It is not legal advice. Call-recording, consent, privacy, advertising, telemarketing, licensing, eligibility, retention, and contract requirements vary by jurisdiction, vertical, traffic method, and campaign. Buyers, publishers, and operators should review their specific practices with qualified counsel.

Dispute reduction is an operating discipline

Most pay-per-call disputes are presented as arguments about individual calls:

• The call was too short.
• The caller was outside the accepted geography.
• The caller was an existing customer.
• The caller had the wrong intent.
• The destination produced dead air.
• The buyer says the call did not convert.
• The publisher says the buyer mishandled it.
• The caller ID did not match.
• A transfer was presented as consumer-initiated inbound traffic.

The call may be the item under review, but the cause often began earlier: an undefined duration clock, an unstaffed destination, mixed call types, a vague duplicate rule, or missing sub-source records. When those gaps survive until settlement, finance becomes the first team forced to interpret campaign rules.

The goal is not a zero-dispute campaign. A zero-dispute claim can be a warning sign if it means nobody reviews invalid calls, nobody reports operational failures, or one party lacks the leverage to challenge an error.

The better goal is a campaign with:

• Fewer preventable disputes.
• Faster detection of recurring problems.
• Specific and timely feedback.
• Consistent evidence standards.
• Clear responsibility for buyer-side and publisher-side failures.
• Traceable adjustments.
• Final records that match invoices and payout reports.

For a broader explanation of the dispute workflow itself, see How Disputes Should Work in a Serious Pay-Per-Call Operation. The focus here is prevention and operating discipline, especially from the publisher side.

Stage 1: Reduce disputes before the campaign launches

The cheapest dispute to resolve is the one the campaign prevents before the first call.

A launch document should do more than list a buyer price, publisher payout, duration threshold, and phone number. It should translate the commercial agreement into rules that routing, QA, account management, and finance can apply consistently.

Put qualification definitions in writing

“Qualified call” is not a complete definition.

A usable definition identifies the required consumer characteristics, accepted intent, excluded intent, geography, call type, schedule, duplicate treatment, existing-customer treatment, duration or conversion rule, and any vertical-specific requirements.

A publisher should be able to answer these questions before sending traffic:

• What product or service must the consumer be seeking?
• Which consumer statements are sufficient evidence of that intent?
• Which intents are specifically excluded?
• Which geography controls: caller-ID area, residence, service location, incident location, or another value?
• Are transfers accepted?
• Are consumers with prior contact excluded?
• How is an existing customer defined?
• What makes a repeat caller a duplicate?
• Which call leg and timestamp determine duration?
• Is settlement duration-based or CPA-based?
• What happens when a buyer-side system fails?

Avoid definitions such as “good call,” “interested caller,” “qualified lead,” or “real intent” unless the campaign adds observable criteria.

A rule should be specific enough that two trained reviewers examining the same evidence are likely to reach the same result.

Define duration start and stop points

A duration threshold is simple only when everyone measures the same interval.

A call can contain several time periods:

• Initial inbound ringing.
• Publisher IVR.
• Routing or reservation time.
• Buyer destination ringing.
• Buyer IVR.
• Queue time.
• Hold time.
• Transfer introduction.
• Two-way agent conversation.
• Post-call wrap-up recorded by another system.

Before launch, identify:

• The authoritative call leg.
• The event that starts the clock.
• The event that stops the clock.
• Whether IVR, queue, hold, or transfer-introduction time counts.
• Whether multiple legs are combined.
• How reconnects are treated.
• How a platform timeout differs from a consumer hang-up.
• Whether duration is a payment rule, a quality signal, or both.

Twilio’s Call resource documentation illustrates why raw telephony status needs interpretation. A call record can include status, start time, end time, duration, direction, and related call-leg identifiers. Twilio also explains that a “completed” call means a connection was established and audio transferred; the answering endpoint could still have been a person, IVR, or voicemail.

That technical record is valuable evidence, but it is not the complete qualification decision.

Publishers should preserve the exact duration method used at launch. Buyers should not change the clock after traffic has run. A detailed treatment appears in Why Call Duration Rules Matter for Buyers and Publishers.

Separate consumer-initiated inbound calls from transfers

Call type should be a required campaign field, not a vague sales description.

For a consumer-initiated inbound call, the consumer generally responds to an advertisement or other consumer-facing experience and places the call. For a transfer, another party may speak with the consumer and then connect the consumer to the buyer.

Transfers can be attended, unattended, warm, cold, screened, or unscreened. They may include an introduction, qualification questions, or a handoff script.

The campaign should define:

• Which call types are accepted.
• Whether mixed traffic is permitted.
• Whether prior agent contact is allowed.
• What the transferring party may say.
• What qualification must occur before transfer.
• Whether the consumer must affirm interest in speaking with the buyer.
• Which caller ID should be presented.
• Which recording or event evidence is expected.
• How transfer time affects duration.

A transfer represented as consumer-initiated inbound traffic creates a predictable dispute because the buyer approved one consumer journey and received another.

See Consumer-Initiated Inbound Calls vs Transfers for the operating distinction.

Define geographic and product eligibility

“National” and “local” are rarely enough.

A campaign may accept certain states, counties, ZIP codes, service areas, licenses, incident locations, or property locations. The relevant geography can differ by vertical.

The controlling geography depends on the vertical: property location for many home services, residence and licensed sales path for insurance, or incident and jurisdictional facts for legal intake. Identify which value controls and how it is verified. Caller-ID area code alone should not silently decide eligibility.

Product eligibility needs similar precision. Broad categories may include services a buyer does not handle. List accepted and excluded categories, and state how mixed-intent calls are treated.

Define existing-customer rules narrowly

“Existing customer” can mean:

• An active customer for the same product.
• Any active customer of the company.
• A former customer.
• A person found anywhere in the CRM.
• A member of the same household.
• A prospect already assigned to an agent.
• A caller who previously requested information but never bought.

Those are different rules.

The campaign should define:

• The relevant customer status.
• Product scope.
• Lookback period.
• Household treatment.
• Evidence required.
• Whether suppression occurs before routing.
• What happens when the publisher has no access to the buyer’s private customer data.

A publisher cannot screen against a buyer’s CRM unless a lawful and operational suppression method exists. If the buyer retains that information, the agreement should allocate the risk clearly rather than assuming the publisher could have known.

Write the duplicate policy as a complete rule

“No duplicates” is incomplete.

A duplicate policy should identify:

• The matched identifier.
• How the identifier is normalized.
• The lookback period.
• Whether the rule is buyer-wide, campaign-wide, target-specific, source-specific, or product-specific.
• The event that starts the lookback.
• Whether rejected or unconnected calls count.
• Whether a new consumer need resets eligibility.
• Whether a new enrollment or service period changes treatment.
• Whether the publisher receives a pre-call suppression response.
• How caller-ID privacy or shared household numbers are handled.

The first call should also be identifiable. A repeat call cannot be reviewed fairly without the record it allegedly duplicates.

Publishers should not accept a duplicate deduction supported only by “already in system.” Buyers should not be expected to pay repeatedly for the same opportunity when a clear duplicate policy was provided and the publisher could apply it.

A fuller framework is available in Why Duplicate Policies Matter in Call Campaigns.

Set caller-ID requirements

Caller ID can support duplicate checks, routing, attribution, fraud review, and call matching. It can also fail for legitimate technical reasons or be altered by a transfer configuration.

Before launch, define:

• Whether the original consumer ANI must be preserved.
• Whether proxy or tracking numbers are allowed.
• How transfers present caller ID.
• Whether anonymous or restricted callers are accepted.
• Which formatting standard is expected.
• How mismatches are reviewed.
• Whether a mismatch creates an automatic rejection or manual review.
• What evidence can substitute when caller ID is unavailable.

A caller-ID mismatch should not be treated as automatic proof of fraud. It is a reason to investigate the call path.

Align caps, schedules, and concurrency with real buyer capacity

When a buyer target is open, it is declaring availability under the configured rules.

The campaign should document:

• Business hours and time zone.
• Holiday schedule.
• Daily, weekly, or monthly caps.
• Call-count and budget caps.
• Concurrency.
• Queue tolerance.
• Overflow behavior.
• After-hours handling.
• Destination failover.
• Pause authority.
• How quickly changes take effect.

A publisher should not be charged with “bad quality” when the buyer accepted calls after closing, routed callers into an unstaffed queue, or allowed concurrency beyond agent capacity.

A buyer should not be forced to accept traffic sent outside a clear schedule or after a cap that the publisher was responsible for enforcing.

The rules need to match the routing system, not merely appear in an onboarding document.

Decide how recordings and consent will be handled

Recordings can resolve disputes involving intent, transfer classification, dead air, hold time, agent behavior, required statements, and disconnects. They also create legal and privacy obligations.

Before launch, determine:

• Whether recording is permitted and expected.
• Which party controls recording.
• Which call legs are recorded.
• How required notice or consent is handled.
• Who may access recordings.
• How access is logged.
• How long recordings remain available.
• Whether transcripts are created.
• How sensitive information is protected.
• What evidence is used when no recording is lawfully available.

Twilio’s Recording resource documentation specifically warns that users must comply with applicable recording and consent laws and recommends consulting legal counsel.

A campaign should never assume that because recording is technically possible, it is automatically lawful or appropriate. It should also avoid making “no recording” an automatic invalid-call rule when the agreement did not assign recording responsibility or when another reliable evidence path exists.

See Call Recordings, Consent, and QA: What Operators Need to Think Through.

Define CPA conversion and reversal rules

A CPA campaign shifts the financial decision from call duration to a later outcome.

The launch terms should state:

• The exact conversion event.
• Who reports it.
• The identifier that links the event to the call.
• The reporting deadline.
• The locking or finalization point.
• The allowed reversal reasons.
• The reversal deadline.
• The evidence required.
• How duplicate conversions are handled.
• How post-payout reversals are handled.
• What happens when the buyer’s CRM or integration fails.

“Did not close” is not an appropriate dispute reason on a duration-based campaign. On a CPA campaign, “conversion” still cannot mean whatever a buyer decides at month-end. The event and reversal rules need to be defined before traffic.

Establish dispute deadlines and evidence requirements

A dispute policy should answer:

• How many days are available to file.
• Which timestamp starts the window.
• Which reason codes are accepted.
• What evidence is required for each reason.
• Whether batch submissions are allowed.
• Who reviews.
• Which outcomes are available.
• Whether partial approval is possible.
• How pending amounts are treated.
• How appeals or reconsideration work.
• How late-discovered technical or compliance issues are handled.
• How rule changes apply prospectively.

Evidence requirements should match the reason.

A wrong-geo dispute may require the accepted geographic rule and the consumer’s relevant location. A duplicate dispute requires the matching earlier call and applicable lookback. A short-call dispute requires authoritative duration events. A transfer-classification dispute may require call-path or recording evidence. A CPA reversal requires the agreed conversion record and reversal basis.

Define adjustment and settlement procedures

The launch terms should identify how an approved dispute affects:

• Buyer price.
• Publisher payout.
• Operator margin.
• Pending invoice batches.
• Pending payout batches.
• Finalized invoices.
• Paid publisher statements.
• Credit memos.
• Future offsets.
• Tax and accounting records where applicable.
• Final settlement reporting.

Buyer price and publisher payout are related, but they are not the same field. A buyer credit does not automatically explain what happened to the publisher payout. Each financial effect needs a record.

Stage 2: Reduce disputes while calls are flowing

A clear launch document is necessary, but it does not monitor the campaign.

During live operation, the priority is to preserve enough information to identify patterns early and to correct problems before they become a settlement-wide argument.

Track source and sub-source consistently

Campaign-level reporting is often too broad.

A publisher may use several domains, media buyers, placements, keywords, call-only ads, landing pages, creatives, transfer teams, or distribution partners. If all traffic appears under one publisher name, a problem in one sub-source can contaminate the reputation and economics of the entire relationship.

At minimum, retain stable identifiers for:

• Publisher.
• Source.
• Sub-source.
• Campaign.
• Creative or landing-page version.
• Tracking number or call path.
• Call type.
• Buyer target.
• Call ID.
• Relevant routing decision.
• Rule version.

The purpose is not unrestricted buyer access to every proprietary publisher relationship. It is scoped accountability. The operator should be able to isolate a recurring pattern and give the publisher enough information to investigate it.

Why Source-Level Reporting Matters for Publishers explains why that granularity protects good sources as much as it identifies weak ones.

Run test calls before and after material changes

Test calls should cover the real consumer path.

Test:

• The displayed number.
• Publisher IVR.
• Routing decision.
• Buyer destination.
• Caller-ID presentation.
• Buyer IVR.
• Queue.
• Agent ringing.
• Recording behavior.
• CRM record creation.
• Disposition capture.
• Disconnect behavior.
• After-hours path.
• Failover.
• Cap and concurrency enforcement.

Repeat tests after changing destinations, schedules, IVRs, transfer scripts, tracking numbers, routing rules, or recording settings.

Document the test time, source, expected result, actual result, and any remediation. A test call that only proves the phone rings is not enough.

Monitor destination health

A publisher can generate a valid consumer response and still deliver a poor experience because the destination fails.

Watch for:

• Busy signals.
• No answer.
• Repeated carrier failures.
• Dead air.
• IVR loops.
• Excessive queue time.
• Long holds.
• Voicemail.
• Unexpected prompts.
• Early buyer disconnects.
• Transfer bridge failures.
• Sudden changes in answer rate or connected duration.

A telephony platform’s “completed” status does not prove that a trained agent handled the call. Review the event sequence and destination behavior.

When a buyer destination is failing, pause or reroute according to the campaign rules. Do not continue sending calls merely to collect more proof.

Require specific rejection reasons

“Bad quality” is not a useful reason code.

Useful reasons describe the rule or failure:

• Wrong geography.
• Wrong service.
• Existing customer.
• Duplicate within the defined lookback.
• Short under the agreed clock.
• Transfer not approved.
• Caller-ID mismatch.
• No answer.
• IVR or bridge failure.
• Dead air.
• Buyer-side disconnect.
• Publisher-side disconnect.
• CPA conversion reversed.
• Recording unavailable.
• Other with required notes.

Reason codes should direct review, not predetermine the outcome.

A publisher should request timely, call-specific, source-level feedback. A buyer should provide enough detail to reproduce the issue without exposing unnecessary private consumer or internal business data.

Improve agent disposition quality

Agent dispositions can be valuable evidence when they capture observable facts.

Weak dispositions include:

• Bad lead.
• Junk.
• No sale.
• Not interested.
• Unqualified.
• Publisher issue.

Better notes include:

• Caller requested billing support for an existing same-product account.
• Caller needed service in a ZIP outside the approved service area.
• Caller asked for employment information.
• Caller requested a product excluded by the campaign.
• Call reached voicemail after the buyer IVR.
• Caller disconnected after four minutes on hold.
• Agent could not hear the caller after the bridge connected.

An agent disposition is one piece of evidence, not an automatic override of route events, campaign rules, or recordings. Managers should audit disposition consistency.

Review recordings or event records when lawfully available

When recordings are lawfully available, review only the portion necessary to answer the dispute question. Protect access and avoid sharing raw recordings more broadly than needed.

When recordings are unavailable, use other evidence:

• Call-leg events.
• Routing logs.
• Timestamps.
• Destination responses.
• IVR events.
• Source records.
• CRM records.
• Agent notes.
• Transfer events.
• Test-call results.
• Incident logs.

Escalate recurring problems quickly

One wrong-service call may be an ordinary exception. Ten similar calls from one sub-source require investigation.

Define escalation thresholds for:

• Repeated wrong geography.
• Repeated existing customers.
• Duplicate spikes.
• Transfer misclassification.
• Caller-ID mismatches.
• Sudden short-call clusters.
• Dead air.
• Destination failures.
• Long holds.
• Missing recordings.
• CPA reporting failures.
• Creative complaints.

The first response may be to reduce volume, pause a sub-source, retest a destination, review a creative, or retrain agents. Fast escalation limits buyer exposure while preserving unaffected sources.

Monitor creatives and landing pages

The consumer journey can drift after approval.

A publisher should retain current versions of:

• Ads.
• Landing pages.
• Forms.
• Disclosures.
• Calls to action.
• Transfer scripts.
• Pre-call questions.
• Brand references.
• Offer language.

Material changes should trigger review before more traffic runs.

Google Ads’ current Misrepresentation policy emphasizes clear, honest ads and destinations and prohibits misleading identity, qualifications, claims, and unavailable offers. Platform policy is not a substitute for legal review, but it reinforces the operational need to compare the promise in the creative with the experience delivered to the caller.

If consumers repeatedly arrive with an intent the buyer does not accept, inspect the ad and landing page before blaming every caller.

Document technical incidents

A technical incident record should include:

• Start and end time.
• Affected sources and targets.
• Call IDs.
• Symptoms.
• Carrier or platform responses.
• Configuration changes.
• Test results.
• Mitigation.
• Calls potentially affected.
• Financial review required.
• Final resolution.

The same symptom can have different owners. Dead air may originate in the publisher IVR, routing bridge, carrier, buyer IVR, contact-center platform, or agent endpoint. A timeline narrows responsibility.

Keep a campaign change log

Every material rule change should record:

• What changed.
• Why.
• Who approved it.
• Effective timestamp.
• Affected sources and targets.
• Whether the change is prospective.
• Whether tests passed.
• Whether publishers were notified.
• Whether finance rules changed.

A dispute should be reviewed against the rule version active when the call occurred.

A screenshot of current settings does not prove what the settings were last week.

Recurring dispute categories and the evidence they need

The table is a starting point. The campaign terms decide what evidence is authoritative.

Stage 3: Review and settle after calls occur

After calls occur, dispute prevention becomes evidence control and financial reconciliation.

The operator should avoid two extremes:

• Approving every buyer challenge to preserve the buyer relationship.
• Rejecting every challenge to protect publisher payout.

The correct question is whether the call met the agreed rule and what financial treatment follows.

Build a call-level evidence packet

A review packet should contain only the information needed for the decision:

• Call ID.
• Campaign, source, sub-source, and target.
• Call date and time.
• Call type.
• Caller-ID value or protected reference.
• Route and call-leg events.
• Relevant duration values.
• Qualification result.
• Dispute reason.
• Buyer notes.
• Publisher response.
• Recording or transcript status when lawfully available.
• Applicable rule version.
• Financial entries affected.
• Review decision.
• Reviewer and timestamp.

Do not send raw caller information, recording URLs, or unrelated private data through casual email or chat.

Use consistent reason codes

Reason codes support trend analysis and consistent evidence requirements.

They should be specific enough to separate:

• Traffic mismatch.
• Buyer handling failure.
• Technical failure.
• Duplicate or existing-customer rule.
• Call-type mismatch.
• Conversion reversal.
• Missing evidence.
• Other exceptions.

“Other” should require notes and should be reviewed for recurring new categories.

Reason codes should never become automatic deductions without review. A code describes the allegation.

Know when to review individually and when to review a batch

Individual review is appropriate when:

• The reason depends on call content.
• The evidence differs by call.
• A publisher disputes the classification.
• The financial amount is material.
• The issue may involve compliance or privacy.
• The pattern is not yet established.

Batch review may be appropriate when:

• One documented destination outage affected a defined time window.
• A routing rule was misconfigured for a specific geography.
• A duplicate file was applied incorrectly to a known group.
• A technical incident produced the same verified failure.
• A conversion import failed for a bounded set of calls.

A batch should still contain call-level identifiers and explain why the same finding applies. “All calls from this source were bad” is not a batch review method.

Separate buyer-side handling failures from publisher-side source failures

Publisher-side failures may include:

• Unapproved traffic type.
• Misleading creative.
• Wrong geography.
• Incorrect pre-qualification.
• Hidden sub-source.
• Repeated duplicates contrary to policy.
• Caller-ID manipulation.
• Transfer misclassification.
• Material source change without notice.

Buyer-side failures may include:

• Closed or unstaffed destination.
• Broken IVR.
• Excessive hold.
• Agent disconnect.
• Inaccurate disposition.
• CRM outage.
• Conversion reporting failure.
• Schedule or concurrency misconfiguration.
• Applying a rule that was not disclosed.

Some incidents involve both sides. The review should not force a single owner when the evidence supports shared responsibility.

Allow partial adjustments when the evidence supports them

A dispute does not always require an all-or-nothing result.

A partial adjustment may be appropriate when:

• Only part of the disputed amount is supported.
• A batch contains both valid and invalid calls.
• Buyer price and publisher payout have different contractual treatment.
• A technical failure affected only one call leg or time period.
• The parties agree to a limited commercial accommodation without reclassifying every call.

The record should distinguish:

• The original amount.
• The disputed amount.
• The approved adjustment.
• The rejected portion.
• The reason.
• The effect on buyer billing.
• The effect on publisher payout.

Do not use an unexplained “courtesy credit” to hide the actual operating issue.

Use credit memos and payout corrections visibly

When an invoice has already been issued, an approved buyer adjustment may require a credit memo or a clearly labeled offset. When a publisher payout report changes, the correction should reference the affected calls and reason.

A clean record shows:

• Original invoice or payout batch.
• Adjustment document.
• Call IDs.
• Reason code.
• Approval date.
• Amount.
• Remaining balance.
• Settlement status.

The buyer should not see a lower total with no explanation. The publisher should not see a reduced payout with only a generic quality note.

What Publishers Should Expect in a Payout Report explains the publisher-side reporting standard.

Reconcile the operating record to finance

At settlement, compare:

• Calls marked billable.
• Calls marked payable.
• CPA conversions.
• Open disputes.
• Approved adjustments.
• Rejected disputes.
• Buyer credits.
• Publisher payout corrections.
• Invoice totals.
• Payout totals.
• Ledger or accounting entries.
• Final balances.

The operating system and finance system should not maintain competing truths.

A call can be valid for one state and unresolved for another. For example, it may be connected and qualified but still pending a CPA outcome. It may be billable before a later approved adjustment. Preserve the history rather than overwriting the original event.

Close with a final settlement record

A final settlement record should allow a future reviewer to answer:

• What was originally billed and payable?
• Which calls were disputed?
• Why?
• What evidence was reviewed?
• Which amounts changed?
• Who approved the changes?
• When did they become final?
• Which invoice, credit memo, payout report, or offset reflects them?
• Are any issues still open?

How publishers protect legitimate payment rights

Dispute reduction should not require publishers to surrender valid payment rights.

Publishers should insist on several basic protections.

Request timely feedback

A problem reported while traffic is live can be investigated. A vague deduction delivered weeks later may arrive after logs, recordings, creatives, or sub-source records are harder to retrieve.

The campaign should include a dispute deadline and faster feedback for recurring problems.

Request call-specific and source-level detail

A useful notice identifies the call, reason, rule, and evidence.

For a pattern, it should identify the affected source or sub-source and time window.

“Quality is down” may justify a conversation. It does not support a payout deduction by itself.

Preserve your own records

Publishers should not rely entirely on buyer reports.

Preserve:

• Source and sub-source mapping.
• Call IDs.
• Tracking numbers.
• Call type.
• Routing responses.
• Timestamps.
• Caller-ID handling.
• Creative versions.
• Landing-page versions.
• Transfer scripts.
• Test-call records.
• Change logs.
• Publisher payout reports.
• Dispute responses.

Keep only what is lawful and necessary, and protect sensitive data.

Challenge retroactive rules

A buyer should not apply a new geography restriction, duration threshold, duplicate lookback, or existing-customer definition to calls that ran under earlier rules. Legitimate changes need an effective timestamp.

Separate non-conversion from invalidity

On a duration-based campaign, a buyer’s failure to close does not automatically make the call invalid.

The buyer may have had weak handling, poor capacity, an unavailable product, or an attribution failure. The call should be judged against the agreed qualification rule.

On a CPA campaign, the publisher should still require a defined conversion event, timely reporting, and specific reversal rules.

Escalate patterns without conceding the whole source

A publisher can acknowledge a problem in one sub-source without accepting that every source is invalid.

Offer a corrective action:

• Pause the affected sub-source.
• Replace a creative.
• Correct geography targeting.
• Retrain a transfer team.
• Fix caller-ID presentation.
• Add suppression.
• Reduce volume while reviewing.
• Run new test calls.

Scoped remediation is more credible than denial and fairer than a publisher-wide deduction.

Joint campaign dispute-prevention checklist

Before launch

• Qualification rules are written and testable.
• Call type is classified as consumer-initiated inbound, transfer, or another approved type.
• Duration start, stop, authoritative leg, and exclusions are defined.
• Accepted and excluded products or services are listed.
• The controlling geography value is defined.
• Existing-customer rules are specific.
• Duplicate identifier, scope, lookback, and triggering event are defined.
• Caller-ID requirements are documented.
• Caps, schedules, time zones, holidays, and concurrency match buyer capacity.
• Recording, consent, access, and alternative-evidence expectations are reviewed.
• CPA conversion and reversal rules are defined where applicable.
• Dispute reasons, deadlines, evidence, outcomes, and escalation are documented.
• Buyer credit and publisher payout adjustment procedures are documented.
• Destination and end-to-end test calls pass.

During operation

• Source and sub-source identifiers are preserved.
• Test calls are repeated after material changes.
• Destination health is monitored.
• Rejection reasons are specific.
• Agent dispositions describe observable facts.
• Recordings or event records are reviewed only when lawfully available and necessary.
• Recurring problems trigger rapid escalation.
• Creatives, landing pages, and transfer scripts are monitored.
• Technical incidents receive a timeline and affected-call list.
• Campaign rule changes have effective timestamps.
• Feedback reaches the affected party while the issue can still be corrected.

After calls occur

• Each dispute identifies the exact call and rule.
• Reason codes are applied consistently.
• Evidence standards match the dispute category.
• Batch reviews still contain call-level identifiers.
• Buyer-side and publisher-side failures are separated.
• Partial approval is available when supported.
• Credit memos and payout corrections reference the original records.
• Open disputes are excluded or clearly identified during settlement.
• Invoice, payout, adjustment, and operating records reconcile.
• Final decisions preserve the original event history and the later adjustment.

Dependable campaigns create explainable outcomes

The central lesson is straightforward.

Make the rules clear before traffic starts.

Preserve evidence while calls are flowing.

Reconcile adjustments in a way both sides can explain later.

That does not eliminate legitimate disputes. It makes them narrower, faster, and fairer. It helps buyers avoid paying for calls that clearly missed the agreed standard. It helps publishers defend calls that met the standard. It gives campaign operators a record they can use instead of relying on memory, pressure, or vague quality labels.

Dependable Calls is being built around controlled source enablement, call-level operating records, structured dispute review, and finance-grade reconciliation. The current implementation supports reasoned dispute states and adjustment workflows, subject to live validation and continued hardening.

Have buyer-ready traffic and want clearer campaign rules, source-level feedback, and explainable payout reporting? Apply to become a Dependable Calls publisher.