A call duration rule looks simple until money depends on it.

A campaign may say that a call qualifies after a certain number of seconds. The buyer sees a clean threshold. The publisher sees a clean payout condition. The operator sees a field that can be stored in a campaign, target, or bid.

Then real calls start moving.

Does the clock begin when the consumer dials, when the buyer destination rings, when an IVR answers, or when a buyer agent joins? Does hold time count? What happens when a call transfers between legs? Is a call that lands exactly on the threshold qualified? Which duration applies if the buyer changes its settings after the call routes? Can the buyer be billed under one threshold while the publisher is paid under another? What evidence should control when two systems report different durations?

These are not edge questions. They determine whether a call is classified as qualified, whether the buyer is charged, whether the publisher earns a payout, whether a dispute is valid, and whether the invoice can be reconciled later.

The basic principle is straightforward:

A duration rule should be a clearly defined commercial term applied to a clearly defined call measurement and preserved with the call record.

Duration can be a useful qualification signal. It is objective, fast to evaluate, and easier to reconcile than a subjective opinion about whether a conversation was “good.” But duration is still a proxy. It does not prove consumer intent, compliance, agent performance, a sale, or long-term value.

This guide explains how buyers, publishers, and operators should think about duration rules before traffic scales.

What a call duration rule actually does

In a duration-based pay-per-call campaign, a call becomes qualified when the measured duration meets the agreed threshold and any other applicable campaign rules.

A simplified rule might be:

A connected call is qualified when its measured duration is at least the campaign threshold and the call is not excluded by an applicable duplicate, source, geographic, or dispute rule.

The threshold may then affect two separate financial decisions:

  • Buyer billability: whether the buyer should be charged.
  • Publisher payability: whether the publisher should earn a payout.

Those decisions may align, but they should not be collapsed into one status without checking the commercial terms. As explained in the difference between routed, qualified, billable, and payable calls, a call can move through several operational and financial states.

The duration rule answers one narrow question:

Did the measured call last long enough to satisfy the agreed duration condition?

It does not automatically answer:

  • Did the caller match every campaign requirement?
  • Was the traffic source approved?
  • Was the caller a duplicate?
  • Did the buyer make a sale?
  • Was the conversation handled well?
  • Was the call compliant with every applicable law or policy?
  • Should a later dispute or adjustment change the initial result?

A serious operation uses duration as one part of the qualification and settlement process, not as a universal quality score.

Duration is a proxy, not proof of quality

Businesses outside pay-per-call also use call length as a practical signal. Google Ads, for example, allows advertisers to count calls that exceed a minimum duration as conversions when configuring phone-call conversion tracking. Google also explains that businesses can use imported call outcomes or other quality signals when duration alone is not enough. That is a useful reminder: a duration threshold is a measurable proxy for a meaningful interaction, not proof that a business outcome occurred.

Google Ads’ phone-call conversion documentation describes minimum call length as a configurable conversion condition. Its broader call measurement guidance also distinguishes duration-based measurement from systems that evaluate or import richer conversion outcomes.

The same distinction matters in pay-per-call.

A longer call can indicate that:

  • The caller reached a person or useful intake process.
  • The caller had enough interest to continue.
  • The buyer had time to gather information.
  • The conversation progressed beyond a wrong number or immediate rejection.

But a long call can also include:

  • Excessive hold time.
  • An IVR loop.
  • A voicemail greeting.
  • An agent searching for information.
  • A consumer trying repeatedly to reach the right department.
  • A transfer that failed after several minutes.
  • A complaint or service call that did not match the campaign.
  • A caller who stayed connected but was never eligible.

Likewise, a short call is not always bad. A well-prepared caller may schedule an appointment quickly. An urgent home-services caller may confirm availability in less time than a complicated insurance shopper. A returning consumer may need only a brief final step.

Duration is useful because it is consistent and auditable. It becomes dangerous when people treat it as a complete explanation of value.

Start by defining which clock you are measuring

“Call duration” can refer to more than one clock.

Depending on the telephony provider, call architecture, and reporting system, operators may encounter:

  • Total time from call creation to termination.
  • Ring time before answer.
  • Connected time after a call is answered.
  • Talk time between the consumer and buyer leg.
  • Time connected to an IVR.
  • Time spent on hold.
  • Parent-leg duration.
  • Child-leg or buyer-leg duration.
  • Recording duration.
  • Transfer-segment duration.
  • Agent conversation duration.

These measurements can differ materially.

Twilio’s official Call resource documentation defines its Call resource duration as the length of the call in seconds and notes that the value is empty for busy, failed, unanswered, or ongoing calls. Twilio also warns that a completed call only proves that a connection was established and audio was transferred; the answering party could have been a person, an IVR, or voicemail.

Twilio’s status-callback documentation further explains that CallDuration is supplied in seconds on terminal call events. That is useful provider evidence, but the commercial agreement still needs to say which call leg and which provider field control qualification.

A buyer and publisher should never assume that two dashboards use the same clock merely because both show a column labeled “duration.”

A practical duration definition should identify

  1. The measured call object. Is the controlling record the inbound consumer leg, buyer leg, bridged call, or another canonical call record?
  2. The start event. Does measurement begin at answer, bridge, agent connection, or another event?
  3. The end event. Does it stop when the caller hangs up, the buyer leg ends, the bridge closes, or the provider marks the call terminal?
  4. The unit. Is the value stored in whole seconds, milliseconds, or another unit?
  5. The comparison rule. Does a call qualify when duration is greater than the threshold or greater than or equal to it?
  6. The source of truth. Which provider event, call-detail record, or canonical platform field controls?
  7. The fallback process. What happens if the duration is missing, delayed, malformed, or inconsistent across systems?

Without those details, “120-second call” can mean different things to different people.

The threshold should match the campaign, not industry folklore

There is no universal duration threshold that makes every call valuable.

The right threshold depends on the actual consumer journey and the commercial model.

A campaign for an urgent local service may reach a meaningful point quickly. A complex insurance, legal, financial, or enrollment conversation may require more time before the buyer can determine fit. A live transfer may arrive after upstream qualification, while a consumer-initiated inbound call may begin with the buyer’s first greeting. An IVR-heavy operation may accumulate connected seconds before an agent speaks.

The threshold should account for:

  • The vertical.
  • The type of call.
  • Whether the call is consumer-initiated inbound or transferred.
  • Whether upstream qualification has already occurred.
  • The buyer’s answer process.
  • IVR and hold-time design.
  • The information agents must collect.
  • Whether the campaign pays for an opportunity or a downstream action.
  • Common wrong-call and service-call patterns.
  • The evidence available for disputes.

A threshold copied from another buyer can fail even when both buyers operate in the same vertical. One buyer may answer directly with licensed agents. Another may place callers into a queue. One may serve a broad appetite. Another may reject many callers after a detailed eligibility check.

The question is not, “What number does everyone use?”

The better question is:

At what point has this buyer usually received a fair, measurable opportunity under this specific call flow?

That requires data and judgment.

What happens when the threshold is too short

A threshold that is too short can make weak interactions billable before the buyer receives a meaningful opportunity.

Common consequences include:

  • Wrong numbers qualifying.
  • Service calls qualifying on a sales campaign.
  • Immediate ineligible calls crossing the line.
  • IVR or greeting time carrying too much weight.
  • Publishers optimizing toward accidental duration rather than caller intent.
  • Buyers disputing calls that technically met the written rule.
  • Invoice totals drifting away from the buyer’s perceived value.

A very short threshold can be appropriate in a narrow flow, but only when the call architecture and qualification work support it.

For example, a properly managed live transfer may arrive with verified campaign information and a warm handoff. The buyer may receive meaningful value quickly because much of the intake occurred before the transfer. That does not mean the same threshold should be applied to cold inbound calls that begin with a menu and no prior screening.

When buyers repeatedly dispute technically qualified calls as “too short to matter,” the operation may have a threshold-design problem rather than a dispute problem.

What happens when the threshold is too long

A threshold that is too long shifts too much operational risk toward the publisher.

The publisher can deliver a legitimate consumer, the buyer can answer, and the agent can conduct a meaningful conversation—yet the call may fail to earn because the buyer ended the call just before the threshold.

Common consequences include:

  • Publishers paying for real media without receiving credit for valid conversations.
  • Buyer agent behavior materially affecting publisher earnings.
  • Strong sources appearing weak because the threshold does not match the call flow.
  • Publishers favoring tactics that extend calls rather than improve fit.
  • Increased payout disputes.
  • Supply becoming harder to retain.

A long threshold may be justified when the buyer needs substantial time to establish eligibility or value. But the buyer should be prepared to explain why that duration represents a fair qualification point.

A threshold should not become an indirect way to make publishers absorb poor answer rates, long queues, confused agents, or inefficient intake.

Buyer operations can change the measured duration

Publishers generate the call, but buyers often control much of what happens after connection.

Buyer-side factors that can affect duration include:

  • Answer speed.
  • IVR design.
  • Queue length.
  • Hold time.
  • Agent availability.
  • Agent scripting.
  • Licensing or department transfers.
  • System latency.
  • Destination failures.
  • Whether agents end calls quickly when a consumer falls outside appetite.
  • Whether callers are asked to repeat information already collected upstream.

That means duration performance should not be interpreted as a pure source metric.

Suppose one source sends the same type of caller to two targets. Target A answers directly with trained agents. Target B uses a long menu, has inconsistent staffing, and transfers callers between departments. The resulting duration distributions may differ even though the source did not change.

Buyers should review duration alongside:

  • Connection rate.
  • Time to answer.
  • Abandonment.
  • Agent availability.
  • Target and destination performance.
  • Source identity.
  • Call type.
  • Disposition or conversion outcomes.
  • Dispute reasons.

Otherwise, the buyer may blame supply for a problem created inside its own call path.

Publishers need the complete rule before sending traffic

A publisher cannot price, route, or optimize traffic responsibly without knowing what earns.

At minimum, the publisher should understand:

  • The qualifying duration.
  • Which call clock controls.
  • Whether the threshold is inclusive.
  • Whether IVR, hold, or transfer time counts.
  • Applicable duplicate rules.
  • Geographic and schedule requirements.
  • Source restrictions.
  • Whether the payout rule differs from the buyer billing rule.
  • How duration exceptions are handled.
  • What reporting will be available.
  • How disputes affect finalized payouts.

The publisher should also know whether the duration is fixed for the campaign or returned dynamically through RTB.

In an RTB flow, the winning bid may include both an amount and a billable-duration term. That allows different buyers or targets to express different economics for the same opportunity. It also creates an important recordkeeping requirement: the winning duration term should be preserved with the call.

A publisher should not have to reconstruct the winning terms from today’s campaign settings weeks after the call occurred.

Duration terms should be snapshotted at routing time

Campaign settings change.

Buyers adjust thresholds. Operators correct configurations. Targets move between test and active states. Commercial terms are renegotiated. RTB responses can vary from one opportunity to the next.

If the platform evaluates an old call against the current setting, it can rewrite history.

Consider this clearly hypothetical example:

  • On Monday, a target accepts calls at a 90-second billable duration.
  • A call routes under that term and completes at 105 seconds.
  • On Wednesday, the target changes to 120 seconds.
  • The invoice is generated on Friday.

The Monday call should be evaluated against the 90-second term that applied when it routed, not the 120-second setting that exists on Friday.

The same principle applies to:

  • Winning bid amount.
  • Buyer price.
  • Publisher payout.
  • Currency.
  • Buyer target.
  • Source identity.
  • Duplicate policy version when applicable.
  • Qualification method.

This is why route-time financial snapshots matter. They preserve the commercial facts attached to the decision.

Without a snapshot, finance teams may be forced to compare completed calls against mutable settings. That is how defensible records turn into memory-based arguments.

Buyer billable duration and publisher payable duration may differ

In the simplest campaign, one threshold controls both sides.

A call reaches the threshold, the buyer is billed, the publisher is paid, and the exchange retains the agreed spread.

More complex arrangements may use different buyer and publisher conditions. That can happen when:

  • The exchange buys supply under one commercial structure and sells it under another.
  • A publisher agreement includes additional validation or duplicate terms.
  • The buyer’s bid supplies a dynamic billable duration while the publisher has a separate payout schedule.
  • A source is being tested under a limited guarantee or special arrangement.
  • A buyer and publisher use different qualification products, such as duration on one side and CPA on the other.

Different thresholds are not automatically improper. Hidden thresholds are the problem.

The operator must track buyer billability and publisher payability separately. The publisher should understand its payout condition. The buyer should understand its billing condition. Internal settlement records should be able to explain the difference without exposing confidential pricing or partner information.

This is one reason buyer price and publisher payout must remain separate concepts.

Exact-boundary calls need an explicit rule

A surprising number of disputes can come from one comparison operator.

Suppose the threshold is 120 seconds.

Does a call at exactly 120 seconds qualify?

Most operations intend “at least 120 seconds,” which means duration >= 120. But a system implemented as duration > 120 would require 121 seconds when duration is stored as whole seconds.

The commercial agreement, platform logic, reports, and settlement tests should all use the same interpretation.

The same applies to rounding.

If one provider reports milliseconds, another reports whole seconds, and a spreadsheet rounds minutes, a boundary call can appear as:

  • 119.6 seconds.
  • 120 seconds.
  • 2.0 minutes.

Those values may look equivalent to a person reviewing a dashboard but produce different automated results.

The safest approach is to store the canonical measurement at the finest reliable unit used by the settlement logic, define rounding rules explicitly, and show partners a consistent value.

Completed does not necessarily mean qualified

Telephony status and commercial qualification are different layers.

As Twilio notes in its call-status documentation, a completed call indicates that a connection was established and audio moved. A human, IVR, or voicemail may have answered.

A completed call may still:

  • Fall below the duration threshold.
  • Reach the wrong department.
  • Be a duplicate.
  • Come from an ineligible source.
  • Miss a geographic requirement.
  • Be subject to a defined dispute.
  • Fail a CPA condition.

Likewise, a terminal provider event may be necessary before final duration is known, but it should not be treated as a financial decision by itself.

A clean system separates:

  1. Provider call status.
  2. Routing status.
  3. Connection status.
  4. Measured duration.
  5. Qualification result.
  6. Buyer billability.
  7. Publisher payability.
  8. Dispute or adjustment status.
  9. Invoice and payout settlement.

That separation makes reporting more useful and prevents one provider field from carrying more meaning than it should.

RTB makes duration rules more flexible—and more demanding

In a fixed campaign, one target may have a configured bid amount and qualifying duration.

In buyer-side RTB, multiple eligible buyers can return different combinations of:

  • Bid amount.
  • Billable duration.
  • Bid expiration.
  • Destination.
  • Acceptance or rejection.

That flexibility allows the routing decision to consider more than price. A higher bid with a much longer duration may not create the best publisher economics. A lower bid with a shorter threshold may produce a more reliable payable rate. The routing policy must decide what it optimizes, and the publisher-facing response should communicate the terms it is allowed to see without exposing protected buyer destination information.

RTB duration rules also create several operational obligations:

  • Parse the buyer response consistently.
  • Validate the duration field.
  • Reject malformed or unsafe values.
  • Preserve the parser and bid terms used.
  • Associate the winning duration with the reservation and call.
  • Apply the same term at settlement.
  • Keep publisher-visible data scoped appropriately.
  • Prevent later configuration changes from altering the result.

The auction may happen in milliseconds. The financial explanation may be needed weeks later.

Both have to work.

Disputes should not redefine the duration rule after the fact

Duration-based buying works best when the primary rule is objective.

A buyer should not be able to say, “The call did not convert, so it should not count,” when the campaign pays for qualified duration rather than a CPA outcome. A publisher should not be able to say, “The call was real, so it should count,” when the call did not meet the written duration or other qualification rules.

A dispute process can still address defined problems such as:

  • Duplicate callers.
  • Wrong call type.
  • Ineligible geography.
  • Technical routing failure.
  • Misrepresented source.
  • Fraud or manipulation supported by evidence.
  • A call record that used the wrong duration field.
  • A mismatch between provider evidence and the canonical record.

But the dispute should apply the agreed rules, not invent a new commercial model after the call.

A strong dispute record should preserve:

  • Original measured duration.
  • Applicable threshold.
  • Initial qualification result.
  • Dispute reason.
  • Evidence reviewed.
  • Decision.
  • Resulting buyer credit or publisher adjustment.

For a broader framework, see how disputes should work in a serious pay-per-call operation.

Duration manipulation is a real incentive problem

Any metric tied to payment can influence behavior.

When duration alone controls payout, participants may be tempted to optimize for seconds rather than value. Risk patterns can include:

  • Unnecessary scripts designed to keep callers connected.
  • Repeated questions.
  • Artificial holds.
  • Delayed transfers.
  • IVR padding.
  • Calls that remain bridged after the useful conversation ends.
  • Sources selected for long average duration even when downstream results are weak.

The answer is not to abandon duration. It is to avoid treating duration as the only signal.

Buyers and operators should compare duration with:

  • Source-level conversion or disposition outcomes.
  • Complaint patterns.
  • Duplicate rates.
  • Short-call and long-call distributions.
  • Agent or target performance.
  • Recording or QA findings where lawfully available and appropriately governed.
  • Dispute concentration.
  • Revenue and payout reconciliation.

The goal is not maximum average duration. The goal is a qualification rule that creates a fair proxy for a valuable call opportunity.

The reporting should make the decision explainable

A call-level report should provide enough information to reconstruct qualification without exposing protected data.

Useful fields may include:

  • Call identifier.
  • Date and time.
  • Campaign or vertical.
  • Source or approved source label.
  • Buyer target identifier or scoped label.
  • Provider status.
  • Connection status.
  • Canonical duration in seconds.
  • Applicable buyer duration threshold.
  • Applicable publisher duration threshold, when different and appropriate to show.
  • Qualification method.
  • Qualification result.
  • Duplicate result.
  • Dispute status.
  • Buyer billing status.
  • Publisher payout status.
  • Adjustment reason.
  • Invoice or payout batch reference.

The exact portal view should remain scoped. Publishers should not receive hidden buyer destinations or confidential buyer details. Buyers should not receive private publisher information beyond what the operating model permits.

Scoped transparency means each party can understand its result without turning the exchange into an unrestricted directory.

Three hypothetical examples

The following examples are simplified and hypothetical. They illustrate the mechanics, not standard pricing or recommended thresholds.

Example 1: The call meets duration but does not convert

A consumer-initiated inbound call reaches a buyer, connects, and lasts longer than the campaign’s duration threshold. The agent does not close a sale.

Under a duration-based campaign, the call may still be qualified and billable because the buyer received the agreed call opportunity. The lack of a sale is a buyer conversion outcome, not automatically a qualification failure.

Under a CPA campaign, the same call may not be billable unless the agreed acquisition event occurs.

This is why buyers and publishers should understand duration-based call buying versus CPA call buying.

Example 2: The call lasts a long time because of hold time

A caller connects to an IVR, waits in a queue, and speaks with an agent only briefly before hanging up. The total provider duration exceeds the threshold.

If the written rule uses total connected duration, the call may technically qualify. But repeated calls with the same pattern suggest the buyer should review queue design and whether the chosen clock represents a fair opportunity.

Changing the rule prospectively may be appropriate. Rewriting it after invoices are generated is not.

Example 3: The threshold changes after routing

A buyer submits an RTB bid with a specific billable duration. The call routes under that bid. The buyer later changes its endpoint configuration.

The completed call should be evaluated against the winning route-time term, assuming that term was valid and accepted—not the later configuration.

That is the difference between a settlement record and a mutable settings lookup.

A buyer checklist for setting a duration rule

Before launching or changing a duration-based campaign, a buyer should be able to answer:

  • What exact call type are we buying?
  • Which call leg and duration field control?
  • When does the clock begin and end?
  • Does IVR or hold time count?
  • Is the threshold inclusive?
  • Why does this threshold represent a fair opportunity?
  • How does the rule differ for inbound calls and live transfers?
  • Are agents consistently available?
  • Are destination and queue performance measured separately from source performance?
  • What other qualification rules apply?
  • Which disputes can change the result?
  • How will exact-boundary calls be handled?
  • Will the route-time term be preserved?
  • Can invoice lines be traced back to the call and threshold?
  • How will changes be communicated before new traffic runs?

A buyer that cannot answer these questions is not ready to argue about call duration after the invoice arrives.

A publisher checklist before accepting a campaign

A publisher should ask:

  • What is the payable duration?
  • Which duration measurement controls?
  • Is the threshold fixed or dynamic?
  • Is the threshold returned in the RTB response?
  • Does hold or IVR time count?
  • What happens at the exact boundary?
  • What source and call-type restrictions apply?
  • What duplicate window applies?
  • Can the buyer dispute a call that met duration?
  • Which dispute reasons are valid?
  • How soon is duration finalized?
  • What call-level reporting will be available?
  • Can payout rows be traced to individual calls?
  • Can campaign changes affect previously routed calls?
  • Who resolves provider-duration mismatches?

A high advertised payout is not attractive if the qualification rule is vague or impossible to audit.

What a serious operator should test

Duration settlement needs more than a field in a form.

A serious implementation should test:

  • Calls below the threshold.
  • Calls exactly at the threshold.
  • Calls above the threshold.
  • Zero or missing duration.
  • Busy, failed, and no-answer outcomes.
  • Duplicate terminal callbacks.
  • Concurrent callbacks.
  • Configuration changes after routing.
  • Fixed and dynamic duration terms.
  • Buyer-only qualification.
  • Publisher-only qualification where commercial terms differ.
  • Neither side qualifying.
  • Dispute adjustments.
  • Ledger or batch retries.
  • Provider-event reordering.
  • Rounding and unit conversions.
  • Missing financial snapshots.

Settlement should also be idempotent. A provider retry should not charge the buyer twice or create a duplicate publisher payout.

This is part of finance-grade pay-per-call reconciliation: each financial entry should be tied to one explainable call event and remain safe when systems retry.

How Dependable Calls is approaching duration rules

Dependable Calls is being built around the idea that the duration term should travel with the routing and financial record.

The current implementation includes buyer bids that can carry a billable-duration term, fixed-bid targets with a configured qualifying duration, completed-call duration capture, and settlement logic designed to apply route-time financial facts rather than mutable current settings.

That implementation has test coverage and controlled validation behind it, but it remains subject to continued live campaign validation and hardening. A coded duration field is not the same as a mature operating process.

The operating standard is broader:

  • Define the clock.
  • Define the threshold.
  • Preserve the winning term.
  • Separate buyer billing from publisher payout.
  • Record the qualification result.
  • Handle retries safely.
  • Keep disputes evidence-based.
  • Reconcile calls to invoices and payouts.
  • Give each partner the scoped information needed to understand the outcome.

A duration rule should reduce ambiguity, not create a new kind of it.

When buyers, publishers, and operators agree on what is measured and preserve that rule with the call, duration-based campaigns become easier to scale, dispute, invoice, pay, and explain.

If you buy calls, generate inbound call traffic, or refer businesses that do either, start a conversation with Dependable Calls.