Call quality is not one number.

It is not the average duration.

It is not the conversion rate.

It is not the percentage of calls a buyer labels “good.”

Those measures can be useful, but none explains the full result. A long call may reflect strong intent, good handling, a long hold, or a wrong department. A high or low conversion rate may reflect the source, the buyer team, follow-up, reporting quality, or several factors at once.

A practical definition is broader:

Call quality is the combined result of consumer intent, lawful and accurate acquisition, source suitability, routing, caller experience, buyer readiness, agent handling, qualification rules, financial treatment, and explainable records.

Every party controls part of that chain. Publishers shape acquisition, expectations, source labels, and change disclosure. Buyers shape capacity, destinations, agents, qualification, and downstream reporting. The operator shapes review, routing controls, records, escalation, disputes, and reconciliation.

A serious process therefore asks at every stage: What must the buyer do? What must the publisher do? What must the operator prove?

This article is educational and operational. It is not legal advice, compliance advice, or a substitute for reviewing a specific campaign with qualified counsel. Telemarketing, advertising, privacy, call-recording, licensing, data-retention, and consumer-protection requirements vary by call type, vertical, technology, state, and operating model. No checklist can promise compliance.

First, keep the call statuses separate

Quality review breaks down quickly when different events are collapsed into one label.

For a fuller foundation, read the difference between a routed call, a qualified call, and a billable call. Write these distinctions into campaign rules and call-level records.

  • Routed: The system selected or attempted a buyer path for the call. Routing does not prove that the destination answered.
  • Connected: A buyer destination answered and a live call leg or bridge was established. A telephony “completed” event may represent an IVR, voicemail, or another endpoint, so provider status alone does not prove quality. See Twilio’s Call resource documentation.
  • Qualified: The call met the campaign’s defined qualification rule, such as an agreed connected-duration threshold, valid geography, accepted call type, nonduplicate status, or another documented condition.
  • Billable: The buyer owes the buyer price under the applicable commercial rule.
  • Payable: The publisher earned the publisher payout under the applicable rule. Billable and payable should not be assumed to be identical.
  • Converted: The buyer reported or the system verified a defined downstream outcome, such as a sale, appointment, enrollment, retained case, or another campaign-specific event.
  • Disputed: An authorized party challenged an outcome within the permitted window and the call is under review.
  • Adjusted: The original financial or status result changed through a documented decision, such as an approved buyer credit, publisher payout correction, or conversion reversal.
  • Settled: The call’s final treatment has been approved for the relevant invoice or payout period. Settled does not necessarily mean cash has cleared; paid is a separate event.

Keep those definitions stable across campaign terms, routing, reports, disputes, invoices, and payout statements.

The 16-stage pay-per-call quality checklist

1. Traffic and source review

Quality starts before a phone rings. A publisher account is not a source: an owned website, paid search, a transfer floor, and partner supply can create very different expectations and risks.

Dependable Calls uses a two-gate model: Dependable Calls decides which reviewed sources are appropriate to offer to a buyer, and the buyer decides which offered sources to enable for a particular campaign, target, or call path. Both gates must be satisfied before curated traffic routes.

Buyer questions

  • What specific source am I reviewing, and can I enable or pause it independently?
  • Is the supply direct, publisher-controlled, partner-supplied, or aggregated?
  • What vertical, geography, language, schedule, and expected volume apply?
  • What changed since the source was last reviewed?

Publisher questions

  • Can I describe the source without hiding behind “inbound” or “network traffic”?
  • Who controls the creative, media, transfer process, and material changes?
  • Will the same stable source and sub-source labels travel with every call?
  • Can I isolate this source from unrelated traffic if quality moves?

Operator questions

  • Is there a complete internal source record tied to the responsible publisher?
  • Has the source been reviewed for the proposed buyer and campaign?
  • Is the buyer being offered only an appropriate, reviewed source?
  • Is buyer enablement enforced in routing rather than only displayed in a portal?

Evidence that should exist: source application, responsible party, traffic method, source/sub-source identifiers, review date, approved scope, material-change history, and buyer enablement record.

Common failure: Several unrelated traffic paths share one source label, so a complaint or performance decline cannot be isolated. Buyers pause the whole relationship, and strong traffic is punished with weak traffic.

For deeper buyer and publisher preparation, see what buyers should ask before accepting publisher call traffic and how publishers should prepare traffic for buyer review.

2. Creative, landing-page, and consumer-experience review

The buyer receives the expectations created upstream. A technically valid call can still be poor quality when advertising implies a government program, guarantee, free service, existing relationship, or unsupported product. Review the whole consumer experience, not only the final landing page.

The FTC’s Telemarketing Sales Rule guidance emphasizes accurate disclosures, prohibits misrepresentations, and explains that some inbound calls made in response to advertising may still fall within telemarketing rules. The FCC’s current 47 CFR § 64.1200 addresses telephone solicitations, consent, identification, calling restrictions, and related protections. Qualified counsel should determine which rules apply to a particular consumer journey.

Buyer questions

  • Does the promise made upstream match the product, service, and agent script?
  • Are agents prepared to identify the company and explain the offer accurately?
  • Would a reasonable caller understand why this buyer is answering?

Publisher questions

  • Are advertisements, landing pages, disclosures, calls to action, and transfer scripts current?
  • Are material claims supportable and appropriately qualified?
  • Can I preserve the exact version that produced a disputed or complained-about call?

Operator questions

  • Were the relevant artifacts reviewed before activation?
  • Are changes versioned and subject to re-review?
  • Is there a process to stop traffic when the live experience no longer matches the reviewed package?

Evidence that should exist: dated screenshots, URLs, ad copy, disclosures, consent language where relevant, transfer scripts, call-to-action text, version history, and approval notes.

Common failure: The publisher updates a winning creative after approval, the new message changes caller expectations, and the traffic keeps using the old source reputation.

3. Call-type identification

“Call” is too broad to configure a campaign safely.

Consumer-initiated inbound calls, warm transfers, outbound-origin transfers, scheduled callbacks, and other call paths create different expectations, evidence, and handling requirements. The call type should be explicit before routing begins.

Buyer questions

  • Is my team prepared for this specific call type?
  • Will the caller arrive directly, after screening, or after speaking with another agent?
  • Does the destination script fit the handoff?

Publisher questions

  • What action caused the consumer to enter the call?
  • If transferred, what was said before the handoff and was the handoff clear?
  • Does the call type ever change by source, hour, or sub-source?

Operator questions

  • Is call type a required source and routing attribute?
  • Are different call types prevented from borrowing one another’s approval?
  • Are recordings or sample calls available where lawful and appropriate?

Evidence that should exist: call-type declaration, consumer-journey diagram, transfer method, screening questions, sample handoffs, and call-level type labels.

Common failure: A buyer expects self-initiated inbound callers but receives outbound-origin transfers. Agents interpret caller confusion as fraud when the real problem is an undisclosed call type.

4. Campaign and target configuration

A campaign describes the commercial and operating rule. A target describes the destination that can receive the call. Both must fit the source.

Buyer questions

  • Which target should receive this source and call type?
  • Is the destination correct, tested, and owned by an accountable team?
  • Are buyer price, qualification, conversion, dispute, and reporting rules approved?

Publisher questions

  • Do I understand the accepted vertical, geography, hours, exclusions, payout, and qualification standard?
  • Are the publisher-facing terms clear before I send traffic?
  • Will I receive a usable rejection reason when a call does not route or qualify?

Operator questions

  • Do source, campaign, target, and buyer enablement all agree?
  • Are configuration changes authorized, timestamped, and auditable?
  • Are buyer destinations protected from publisher-facing systems?

Evidence that should exist: approved campaign terms, target ownership, tested destination, source-to-target mapping, pricing/payout versions, qualification rule, conversion definition, and audit history.

Common failure: The campaign sheet says one thing, the routing configuration says another, and finance uses a third version. Every later dispute becomes a debate about which rule was real.

5. Geographic, schedule, cap, and concurrency controls

A buyer can want more calls and still be unavailable for the next call.

Geography decides where the caller must be. Schedule decides when the target is open. Caps limit volume or spend over a period. Concurrency limits simultaneous live demand. These controls protect the caller and the buyer from routing into capacity that does not exist.

Buyer questions

  • Which states, ZIP codes, service areas, products, and languages are truly supported?
  • Which hours are staffed in the target’s time zone?
  • What daily, hourly, or budget cap reflects a controlled test?
  • How many simultaneous calls can the team handle without queue deterioration?

Publisher questions

  • Can I restrict traffic to the approved scope before sending?
  • Do I understand what happens when a cap closes or a schedule changes?
  • Will I receive timely notice of planned capacity changes?

Operator questions

  • Are geographic rules normalized and tested?
  • Do schedules use the correct time zone and daylight-saving behavior?
  • Are cap and concurrency counters enforced atomically and observable?
  • Does the route fail closed when eligibility cannot be verified?

Evidence that should exist: geography matrix, schedule and time zone, cap definitions, concurrency limit, fallback policy, test results, and change log.

Common failure: A buyer advertises a large daily capacity but has no burst control. Ten calls arrive in two minutes, agents miss half of them, and the source is blamed for low connection and conversion.

Read how caps, schedules, and concurrency shape call flow for the control-by-control explanation.

6. Caller-ID and duplicate handling

Caller identity can support routing continuity, duplicate detection, and investigation, but it must be handled carefully.

A caller-ID match may help connect an accepted pre-call opportunity to the arriving call. A duplicate policy may prevent the same opportunity from being bought repeatedly. Neither control should be a vague excuse for rejection.

Buyer questions

  • What counts as a duplicate: same caller, same campaign, same buyer, same source, or same conversion?
  • What lookback window applies?
  • Are legitimate repeat callers treated differently from accidental resubmissions?

Publisher questions

  • Will the same normalized caller identifier appear in the pre-call request and live call?
  • How are blocked, unavailable, international, or malformed identifiers handled?
  • Can I see a clear duplicate or mismatch reason without receiving unnecessary consumer data?

Operator questions

  • Is phone data normalized consistently?
  • Is duplicate scope defined before launch?
  • Can the system distinguish repeated API requests from repeated consumer calls?
  • Are privacy and access controls applied to caller data?

Evidence that should exist: identifier-normalization rules, match requirements, duplicate key, scope, lookback window, exception handling, and call-level rejection reason.

Common failure: “Duplicate” means something different to operations, finance, the buyer, and the publisher. The same label is used for repeated pings, repeat callers, and repeated conversions.

See why caller-ID matching matters in call routing and why duplicate policies matter in call campaigns.

7. Routing and connection quality

Judge source quality separately from route quality. Routing includes target selection, reservation validity, destination health, ring time, bridging, audio, fallback, and failure reasons.

Buyer questions

  • Is the destination answering consistently and quickly enough for this call type?
  • Are IVR, queue, voicemail, and after-hours behavior intentional?
  • Can technical failures be distinguished from caller hang-ups?

Publisher questions

  • Did the route accept the opportunity before the call was sent?
  • How long is a route or reservation valid?
  • What response explains a rejection, timeout, or failed connection?

Operator questions

  • Can the call be traced from source through routing decision and buyer leg?
  • Are reservations single-use and protected from replay?
  • Are destination errors, carrier errors, no-answer events, and caller abandonment separated?
  • Is audio degradation or one-way audio included in investigation procedures?

Evidence that should exist: routing decision, eligible and rejected targets with reasons, reservation timestamps, call-leg identifiers, provider events, ring and answer timestamps, bridge result, and destination-health record.

Common failure: The buyer destination stops answering, but the report shows only short calls. The publisher’s source is paused even though the failure occurred after routing.

8. Agent and intake readiness

The buyer’s call center is part of quality. Strong traffic can look weak when agents are unavailable or unprepared; weak traffic can look temporarily strong when only the best agents receive a small test.

ISO’s ISO 18295-1 customer contact centre standard is one useful quality-management reference for customer contact centers. A pay-per-call campaign does not need to claim ISO certification to borrow the underlying discipline: defined processes, competent staff, monitoring, and continual improvement.

Buyer questions

  • Are agents trained on the source, call type, consumer promise, qualification, and escalation path?
  • Is staffing sufficient for expected arrival patterns?
  • Are dispositions consistent and reviewed?
  • Can agents identify wrong-number, service, language, eligibility, and complaint scenarios correctly?

Publisher questions

  • What buyer experience should the caller expect after connection?
  • Is feedback specific enough to improve the source rather than merely label it “bad”?
  • Are material buyer-side script or destination changes communicated?

Operator questions

  • Has the buyer confirmed launch readiness?
  • Are answer rate, time to answer, hold time, disconnects, and agent dispositions visible?
  • Can performance be segmented by buyer target or team?

Evidence that should exist: training acknowledgment, current script, staffing plan, escalation contacts, disposition dictionary, launch test, and agent QA sample.

Common failure: The buyer activates traffic during training, meetings, or a shift change. Calls connect but receive weak handling, and the resulting conversion rate is treated as a permanent source score.

9. Qualification and conversion rules

Define qualification and conversion before data arrives. Duration is not a complete quality judgment, and CPA outcomes can be late or affected by buyer follow-up. Specify the event, source of truth, timing, exceptions, and corrections.

Buyer questions

  • What exactly makes a call qualified?
  • Which event makes it billable?
  • What exactly counts as a conversion, and who reports it?
  • How quickly will downstream outcomes be returned?

Publisher questions

  • What makes the call payable?
  • Is payout determined from the same event as buyer billing or a separate rule?
  • Can I see enough evidence to understand nonpayable outcomes?

Operator questions

  • Are qualification, billable, payable, and converted statuses modeled separately?
  • Is the applicable rule version preserved with the call?
  • Are late conversions, reversals, and missing dispositions handled explicitly?
  • Is manual override authority limited and audited?

Evidence that should exist: rule definition, event source, timing window, version, required fields, conversion intake record, reversal reason, and manual-review history.

Common failure: “Qualified” is used to mean duration-met in one report, buyer-approved in another, and converted in a third. The campaign appears profitable until invoice and payout populations are compared.

10. Call recordings and QA

Recordings provide evidence but create legal, privacy, security, retention, and access questions. Technical capability does not establish permission. Consent rules vary, sensitive data may appear, and access should be limited by purpose.

The NIST Privacy Framework provides a useful governance lens for identifying privacy risk, defining controls, limiting access, and managing data over its lifecycle.

Buyer questions

  • Is recording lawful for the call path and jurisdictions involved?
  • Who may access recordings, for what purpose, and for how long?
  • Does the QA form measure agent behavior, caller experience, source fit, and technical quality separately?

Publisher questions

  • Will I receive feedback or redacted evidence without unnecessary consumer or buyer information?
  • Are sample calls representative rather than selectively chosen?
  • Can I challenge a QA interpretation with the underlying rule?

Operator questions

  • Is recording status known at the call level?
  • Are access, download, sharing, retention, and deletion controlled?
  • Does QA distinguish facts from reviewer judgment?
  • Are recordings one input rather than an automatic compliance or payment decision?

Evidence that should exist: recording-consent analysis, notice language where applicable, access log, retention rule, QA form, reviewer notes, redaction policy, and sample-selection method.

Common failure: A recording link is copied into email or chat, exposing caller data and a protected destination. The evidence process creates a larger risk than the original quality question.

11. Source-level reporting

Blended reporting hides both good and bad traffic.

Quality reporting should preserve the source as a stable unit of analysis and show enough of the chain to distinguish acquisition, routing, handling, qualification, conversion, disputes, and financial outcomes.

Buyer questions

  • Can I compare sources under the same target, hours, and rules?
  • Can I separate source performance from target performance?
  • Are sample size and date range visible?

Publisher questions

  • Can I see routed, connected, qualified, payable, disputed, adjusted, and settled outcomes by source?
  • Can I identify whether the problem is source-specific, geography-specific, or buyer-path-specific?
  • Are rejection reasons useful enough to act on?

Operator questions

  • Do buyer and publisher views expose only appropriate fields?
  • Are metrics computed from consistent populations and definitions?
  • Can a call-level drilldown explain the aggregate?
  • Are historical results preserved when a source changes?

Evidence that should exist: source identifier, time range, metric definitions, numerator/denominator populations, target context, rejection reasons, dispute rate, and call-level support.

Common failure: The buyer sees a publisher average, the publisher sees a campaign average, and the operator cannot reconcile either number to the same calls.

Read why source-level reporting matters for publishers.

12. Complaints and compliance escalation

A complaint is neither automatic proof nor ordinary performance feedback. It needs a separate path because it may involve confusion, misrepresentation, unwanted contact, recording, identity, sensitive data, or agent conduct.

Buyer questions

  • Can agents identify and escalate a complaint immediately?
  • Is the consumer’s concern recorded accurately rather than reduced to a generic disposition?
  • Who decides whether the buyer path should pause?

Publisher questions

  • Can I identify the source, creative, call type, and responsible upstream party?
  • Can I preserve relevant evidence without contacting or exposing the consumer improperly?
  • What material changes must stop while the issue is reviewed?

Operator questions

  • Is there an escalation owner and severity framework?
  • Can the affected source or target be paused without disabling unrelated traffic?
  • Are legal, compliance, privacy, and security questions routed to qualified professionals?
  • Are facts preserved before records expire or change?

Evidence that should exist: complaint timestamp, call identifier, source, target, consumer statement, involved artifacts, preservation action, access log, decision owner, and resolution.

Common failure: A serious complaint is placed in the normal dispute queue and reviewed days later as a billing question.

13. Disputes and adjustments

A dispute should test the call against the agreed rule and evidence, not reopen the bargain after the result. Define who may submit, permitted reasons, evidence, timing, holds, reviewer authority, decisions, and adjustments.

Buyer questions

  • Is the reason tied to a campaign rule?
  • Was the dispute submitted within the agreed window?
  • Does the evidence distinguish source failure from buyer handling or technical failure?

Publisher questions

  • Can I see the call, rule, reason, and non-sensitive evidence needed to respond?
  • Does an open dispute hold payout, reverse payout, or merely flag the call?
  • Can a decision be appealed or corrected if material evidence was missing?

Operator questions

  • Is the reviewer independent enough to apply the rule fairly?
  • Are buyer price and publisher payout reviewed separately?
  • Does an approved decision create an explicit adjustment rather than edit history?
  • Can the final reason be reported consistently?

Evidence that should exist: dispute reason code, narrative, submitted evidence, rule version, reviewer, timestamps, decision, financial impact, and adjustment entry.

Common failure: The original call row is silently changed. Later, nobody can explain why yesterday’s invoice and today’s report differ.

See how disputes should work in a serious pay-per-call operation.

14. Buyer invoices and publisher payouts

Buyer invoices and publisher payout reports are related but not mirror images. Buyer price is charged to the buyer; publisher payout is paid to the publisher. Each needs its own rule and call population.

Buyer questions

  • Can every line or summary be traced to billable calls?
  • Are credits and adjustments shown clearly?
  • Do taxes, fees, payment terms, and period boundaries match the agreement?

Publisher questions

  • Can every earning be traced to payable calls?
  • Are holds, disputes, adjustments, and nonpayable reasons clear?
  • Are payout terms and period boundaries consistent?

Operator questions

  • Are invoice and payout batches generated from controlled call-level records?
  • Are money values stored and calculated consistently without floating-point drift?
  • Are duplicate entries and cross-period adjustments prevented or detectable?
  • Are invoice and payout access appropriately scoped?

Evidence that should exist: batch identifier, period, call population, buyer price or publisher payout, adjustment references, totals, payment terms, and generation timestamp.

Common failure: Finance exports one spreadsheet for billing and another for payouts from different filters. Both look reasonable, but the call populations do not reconcile.

15. Reconciliation and post-launch review

The period is not complete until call records, disputes, invoices, payouts, and settlement agree. Reconciliation also exposes definition drift, missing events, duplicates, delayed conversions, and manual adjustments.

Buyer questions

  • Does the invoice match my accepted billable population and approved credits?
  • Are unexplained differences investigated at the call level?
  • Did buyer-side handling or reporting create avoidable losses?

Publisher questions

  • Does the payout report match payable calls and approved adjustments?
  • Are missing calls, holds, and rejected calls explainable?
  • Did a source or sub-source change during the period?

Operator questions

  • Do routed, connected, qualified, billable, payable, converted, disputed, adjusted, and settled counts reconcile?
  • Can totals be reproduced from immutable or append-only financial records?
  • Are late events assigned to the correct period?
  • Is every unexplained variance owned and resolved?

Evidence that should exist: reconciliation report, source totals, invoice totals, payout totals, dispute/adjustment totals, variance log, reviewer, and period-close approval.

Common failure: A campaign is called “profitable” from a routing dashboard before disputes, conversion reversals, credits, and publisher adjustments are included.

For a deeper finance framework, read why pay-per-call needs better financial reconciliation.

16. Scaling readiness

Scaling is a new operating decision, not a reward for one metric. Scale only when source identity is stable, buyer capacity is real, reporting is trustworthy, disputes are understood, and finances reconcile. Scale may change caps, hours, geography, targets, or concurrency—not simply “send everything.”

Buyer questions

  • Can my staffing, destination, budget, and follow-up process absorb more calls?
  • Does performance remain acceptable by source, target, hour, and geography?
  • Are conversion reports complete enough to support the decision?

Publisher questions

  • Can the source increase without changing media, creative, upstream supply, or caller experience?
  • Can I keep source labels and controls stable at higher volume?
  • What happens if performance moves after the increase?

Operator questions

  • Has the test produced a meaningful sample under representative conditions?
  • Are connection, qualification, complaint, dispute, and reconciliation patterns understood?
  • Is the increase bounded, observable, and reversible?
  • Does a material source or buyer change require a new test?

Evidence that should exist: baseline period, sample size, source and target performance, quality sample, dispute and complaint review, capacity confirmation, reconciliation result, approved increase, and rollback threshold.

Common failure: A promising source moves from a small morning test to full-day national volume. The buyer’s capacity, geography, and agent mix change at the same time, making the original benchmark irrelevant.

See how to evaluate a pay-per-call source before scaling.

Hypothetical example: why one quality score fails

Consider two hypothetical sources.

Source A produces long calls and a strong connected-duration rate. QA review finds that many callers expected customer service for an existing account, not a new product. Agents spend several minutes identifying the mismatch. Duration looks strong; consumer intent and source suitability are weak.

Source B produces shorter calls and a lower raw conversion rate during the first two days. Call review shows accurate consumer expectations, but the buyer routed the source to a new-agent queue with long holds. After moving the source to a prepared team, answer time improves and conversion rises.

A single score might reward Source A and reject Source B. An end-to-end review reaches a different conclusion:

  • Source A needs acquisition and consumer-experience correction.
  • Source B needs buyer-path and staffing correction.
  • Neither result can be explained by duration alone.
  • The buyer, publisher, and operator each own part of the remedy.

The checklist identifies the stage that produced the result.

Pre-launch buyer checklist

Before enabling a source, the buyer should be able to check every item below.

  • We reviewed the specific source, not merely the publisher account.
  • We understand the consumer journey and call type.
  • The source fits our actual product, geography, language, and licensing or service scope.
  • Our destination is tested and assigned to an accountable team.
  • Our schedule, cap, concurrency, and budget reflect real capacity.
  • Agents are trained on the source, caller expectation, script, dispositions, and complaint path.
  • Qualification, billable, conversion, duplicate, dispute, and reporting rules are written.
  • We can return timely and consistent dispositions or conversions.
  • We know what would trigger an immediate pause.
  • We approved a bounded test before discussing scale.

Pre-launch publisher checklist

Before sending a call, the publisher should be able to check every item below.

  • The source has a stable identity and responsible owner.
  • Traffic type, consumer journey, geography, language, hours, and expected volume are documented.
  • Creative, landing pages, disclosures, scripts, and material claims are current.
  • Required source and sub-source labels are included consistently.
  • The live call can match the accepted pre-call opportunity where required.
  • Duplicate, qualification, payable, dispute, and payout rules are understood.
  • Material source changes will be disclosed before traffic continues.
  • Callers will receive an experience consistent with the reviewed materials.
  • We can isolate, reduce, or pause the source quickly.
  • We are prepared to start with a controlled test.

First-week monitoring checklist

Use the first week to find mismatches while the test remains small.

  • Review route acceptance, connection, no-answer, failure, and abandonment daily.
  • Compare performance by source, target, hour, geography, call type, and agent team.
  • Listen to or review a lawful, representative QA sample rather than only the longest or shortest calls.
  • Check caller expectations against the reviewed creative and scripts.
  • Confirm source labels, call types, and duplicate reasons are stable.
  • Confirm buyer dispositions and conversions arrive on schedule.
  • Separate source problems from routing, destination, and agent problems.
  • Review complaints and unusual calls immediately.
  • Compare billable and payable populations before the first period close.
  • Record decisions, owners, and changes instead of managing the test through memory.

Source-scaling checklist

Increase volume only when each item is satisfied or the limitation is explicitly accepted.

  • The source identity and acquisition method remained stable during the test.
  • The consumer experience matches the reviewed package.
  • Routing and destination health are reliable.
  • Buyer staffing can absorb a defined increase.
  • Qualification and conversion data are complete enough to interpret.
  • Results are understood by source, target, hour, and geography.
  • Complaint and dispute patterns are acceptable and explainable.
  • The first invoice and payout populations reconcile.
  • The increase has a cap, monitoring window, owner, and rollback threshold.
  • Scale does not silently introduce new sub-sources, creatives, geographies, or call types.

Campaign pause or investigation checklist

Pause the affected scope—not automatically every relationship—when one or more serious triggers appear.

  • Consumer complaints suggest confusion, deception, unwanted contact, recording concerns, or misuse of data.
  • The live source no longer matches reviewed materials.
  • Source or sub-source identity cannot be traced.
  • Call type is misrepresented or changes without approval.
  • Caller-ID mismatch, duplicate, or reservation failures spike unexpectedly.
  • Connection quality or destination health deteriorates.
  • Buyer handling changes materially or agents are not prepared.
  • Qualification, conversion, billing, or payout definitions conflict.
  • Reports cannot be reconciled to call-level evidence.
  • Disputes or adjustments rise without a clear cause.
  • Required records are missing, inaccessible, or exposed improperly.
  • A legal, regulatory, privacy, security, or licensing question requires qualified review.

Preserve the source version, configuration, call and routing records, QA evidence, complaints, disputes, and financial state. Do not erase the original outcome to clean the report.

Compact master checklist

Use this as the final launch-and-scale gate.

Source and consumer

  • Specific source identified and reviewed
  • Responsible publisher and control boundary known
  • Creative, lander, script, disclosures, and call type aligned
  • Material-change and version process defined
  • Buyer offered the source; buyer enabled it for the intended path

Routing and capacity

  • Campaign and target rules agree
  • Destination tested
  • Geography and schedule validated
  • Cap and concurrency reflect live capacity
  • Caller-ID, duplicate, reservation, and fallback rules defined
  • Routing and connection failures are explainable

Buyer handling and outcomes

  • Agents trained and staffed
  • Dispositions defined
  • Qualification, billable, payable, and conversion rules separated
  • Conversion reporting has an owner and deadline
  • QA measures consumer, source, routing, and agent factors separately

Records, complaints, and finance

  • Source-level reporting reconciles to call-level records
  • Recording, privacy, access, and retention requirements reviewed
  • Complaint and compliance escalation path active
  • Dispute window, evidence, review, and adjustment rules active
  • Buyer invoice and publisher payout populations are traceable
  • Reconciliation completed before scale

Scaling

  • Test sample is representative
  • Source and buyer conditions remained stable
  • Complaint and dispute patterns are understood
  • Capacity increase is bounded and reversible
  • Pause owner and rollback threshold are documented

What Dependable Calls is building toward

Dependable Calls is being built as a controlled, operator-led exchange for serious buyers and publishers. The model centers on reviewed supply, source accountability, buyer routing control, explicit statuses, and finance-grade reconciliation—not unrestricted discovery or default routing.

The current implementation supports source and campaign domains, routing eligibility, capacity controls, caller-ID matching, duplicate and dispute workflows, reporting, and financial records. These remain subject to live validation, production certification, operational adoption, and continued hardening. Code does not prove every workflow is exposed, mature, or used live.

The standard is not a promise that every call will comply, qualify, convert, or settle without disagreement. It is disciplined questions, controls, evidence, and explainable decisions.

Want cleaner call supply or a more serious buyer-review process? Start a conversation with Dependable Calls.