The call lasted 4 minutes and 12 seconds. Agent and homeowner talked about a leaking water heater, confirmed the address, and scheduled a tech for Thursday morning.
Then the agent hung up and marked the call "Not Qualified."
Nobody caught it for three weeks. By then, the publisher had invoiced for 340 calls, the buyer disputed 60 of them based on disposition codes, and both sides were digging through recordings trying to figure out who owed what. The 4-minute water heater call — an actual appointment, an actual job, actual revenue — sat in a pile marked "reject."
I've seen this exact scenario play out more times than I'd like to admit. (Mostly because I've been on both sides of the dispute table, sometimes in the same month.) Call dispositions sound boring until they're the reason you can't reconcile a five-figure invoice.
What Is a Call Disposition and Why It Matters
A call disposition is a tag applied to a phone call after it ends, classifying the outcome. Qualified. Not interested. Callback requested. Wrong number. Voicemail. Billable. Reject. The specific codes vary by operation, but the concept is universal: every call gets a label that describes what happened.
Why does this matter right now? Pay-per-call billing runs on disposition data. Publishers get paid for calls marked billable. Buyers dispute calls marked qualified that actually hit voicemail. Networks audit disposition accuracy and boot operators who game the system. And increasingly, AI systems are setting dispositions automatically based on transcription and sentiment analysis — which means the rules are getting more complex, not less.
If you're running any volume in pay-per-call campaigns, dispositions aren't optional metadata. They're the spine of your billing reconciliation.
Core Concepts
Before we dig into the mechanics, four things you need to understand.
Disposition vs. Billable Status. These overlap but aren't the same. Billable status is binary: did this call trigger payment? Yes or no. Disposition is categorical: what happened on the call? A call can be billable (crossed the 90-second threshold, passed IVR, matched geography) but have a disposition of "callback" because the homeowner asked to reschedule. Billable status comes from system rules. Disposition comes from human judgment — or AI judgment if you're automating.
Agent Dispositions vs. System Dispositions. Two sources. Agents select dispositions manually after each call, picking from a dropdown or pressing a keypad code. Systems assign dispositions automatically based on rules: call under 60 seconds = "short call," IVR abandoned = "IVR dropout," duplicate caller ID = "repeat." Most setups combine both — system sets a default, agent overrides if reality differs. The problem is when they conflict and nobody notices. This is where VeloCalls' AI conversation intelligence catches mismatches between what the agent tagged and what actually happened on the call.
Disposition Hierarchies. Not all dispositions are equal. Some map directly to billing outcomes (qualified = billable, reject = non-billable). Others are informational (callback requested, transferred to manager). Most platforms let you build hierarchies: top-level categories (billable, non-billable, pending) with subcategories underneath (billable > qualified > appointment set). This structure matters for reporting — you want to slice by "why was this non-billable" without building custom queries every time.
Disposition Windows. How long after a call can the disposition be changed? Most platforms allow edits for 24-72 hours. After that, records lock. This protects billing integrity — you can't go back six months later and flip a thousand "qualified" calls to "reject" to claw back payouts. Networks that allow unlimited edits tend to have... messy financials. For attribution data alongside your call dispositions, JustAnalytics tracks session-level behavior without GDPR consent banner headaches.
The Disposition Process: Step by Step
Here's how a call moves from "ringing" to "dispositioned" in a typical pay-per-call setup.
Step 1: Call Connects and Agent Talks
The caller reaches your agent (or the buyer's agent, depending on the model). Conversation happens. The agent is focused on qualifying the lead, capturing information, scheduling an appointment — not on what code they'll assign afterward.
This is where problems start. The disposition decision happens after the call ends, when the agent is already thinking about the next call. Mental bandwidth is low. Speed matters. They pick a code and move on.
Honestly, I get it. I've listened to agents handle 80+ calls in a shift. By call 60, you're not carefully considering each disposition — you're surviving.
Step 2: Call Ends and Disposition Screen Appears
Most platforms force a disposition selection before the agent can take the next call. A modal pops up: "Select call outcome." Dropdown with 5-10 options. Sometimes a notes field. Sometimes a callback scheduler if that's the outcome.
The forcing function matters. Without it, agents skip dispositioning, leaving calls in "pending" limbo that requires supervisor cleanup. With it, you get 100% disposition coverage — but not necessarily 100% accuracy.
Step 3: Agent Selects Disposition
The agent picks a code. Common options in home services pay-per-call:
- Qualified / Appointment Set — lead meets criteria, next step scheduled
- Not Interested — talked to homeowner, they declined
- Wrong Number — caller wasn't trying to reach this service
- Voicemail — no live pickup, left message or didn't
- Callback Requested — interested but can't talk now
- Out of Area — homeowner's location doesn't match buyer's service area
- Duplicate — same caller from recent call
- Test Call — internal QA or publisher testing
That's eight codes. Most operations don't need more than that. Some get fancy with sub-codes (Not Interested > Price Too High, Not Interested > Already Hired Someone, Not Interested > Just Researching). Useful for product teams analyzing objections, less useful for billing reconciliation.
Step 4: System Validates or Overrides
Here's where automation kicks in. Your platform checks the agent's selection against system data.
Agent marked the call "Qualified" but call duration was 45 seconds? System flags it — human review required, or automatic override to "Short Call." Agent marked it "Out of Area" but caller's ANI area code is in the buyer's service zone? Flag for review.
Good platforms surface these conflicts. Bad platforms let them through and you find out during invoice disputes.
Step 5: Disposition Feeds Billing
The disposition record flows downstream. Calls marked "Qualified" or "Appointment Set" typically map to billable. Calls marked "Voicemail," "Wrong Number," or "Short Call" map to non-billable.
This mapping is configured per campaign, per buyer, per network. Don't assume. A disposition of "Callback Requested" might be billable in one campaign (caller engaged, just bad timing) and non-billable in another (buyer only wants closed appointments). Check your IO.
If you're tracking paid search alongside your call campaigns, ClickzProtect catches click fraud before fake sessions generate fake calls that need dispositioning anyway.
Step 6: Reporting and Reconciliation
At invoice time, both sides pull disposition reports. Publisher shows 500 calls, 420 marked billable. Buyer shows 380 billable based on their disposition review. The delta — 40 calls — becomes a dispute.
You pull recordings. You compare dispositions. You argue about whether a 95-second call where the homeowner said "let me think about it" is qualified or not qualified. This is the least fun part of pay-per-call, and disposition accuracy is what makes it faster or slower.
Operators running high volume across multiple publishers use dashboards to track disposition consistency. Publisher A runs 78% qualified rate, Publisher B runs 52%. Either Publisher B has worse traffic or their agents are dispositioning differently. Dig in.
Advanced Tips
Calibrate across agents monthly. Pull a random sample of 20-30 calls per agent. Review recordings. Compare the disposition they assigned vs. what you would have assigned. Drift happens — new agents miscategorize, experienced agents develop shortcuts. Monthly calibration catches it before it shows up in your invoice disputes. For engineering teams building custom calibration tooling, DevOS streamlines the development environment setup.
This is non-negotiable, in my opinion. Skip it and you're flying blind.
Build disposition rules into your IVR. If a caller presses "1 for sales" and then abandons before agent connect, the system should auto-disposition as "IVR Abandon" without agent input. Same for calls under 30 seconds — system disposition "Short Call" regardless of what the agent might pick. The fewer manual decisions per call, the more consistent your data.
Use AI transcription to validate agent dispositions. This is where things are heading in mid-2026. Platforms with AI conversation intelligence — VeloCalls, Invoca, some Ringba setups — can analyze transcripts and flag mismatches. Agent marked "Not Interested" but transcript shows "yes, I'd like someone to come out Thursday"? Flagged for review. Not perfect (AI still misreads sarcasm and handles accents unevenly) but catches obvious errors.
Separate intent dispositions from outcome dispositions. Some operations track both. Intent: what did the caller want? (Service call, estimate, emergency, general question.) Outcome: what happened on the call? (Appointment set, referred out, callback, no sale.) Two-axis tracking gives you better data — you can see that 80% of "emergency" intent calls convert vs. 30% of "estimate" intent calls. Most platforms support custom fields for this, even if their default disposition list doesn't.
Watch for disposition gaming. Publishers sometimes push agents to mark more calls "qualified" to inflate billable counts. Buyers sometimes mark too many "reject" to reduce payouts. Neither side is innocent here — and if you think your side is, you're probably not looking hard enough. Network-level disposition audits (random sample, third-party review) keep both sides honest. If you're the network, build this into your compliance process. For email outreach campaigns that drive inbound calls, JustEmails handles deliverability so your sequences actually land.
Common Mistakes
Using too many disposition codes. I've seen setups with 30+ codes. Agents freeze, pick randomly, or always choose "Other." Consolidate to 5-8 codes that map to actionable categories. Every code should answer a question you actually care about.
Not validating dispositions against call duration. A 22-second call marked "Qualified - Appointment Set" is almost certainly wrong. Build validation rules that flag impossible combinations. Takes 30 minutes to configure, saves hours of dispute resolution.
Letting pending dispositions pile up. Calls without dispositions are billing black holes. Some platforms age these into "Unresolved" after 48 hours and auto-mark non-billable. Others let them sit forever. If you're running the latter, build an alert for pending counts and clear them daily.
Assuming dispositions transfer between platforms. Migrating from Ringba to CallRail? From CallRail to VeloCalls? Your historical disposition data doesn't follow. Export it before migration, and map your old codes to new codes in documentation. Otherwise you're rebuilding three years of training data from scratch. Related: check out our migration guides if you're switching platforms.
Not training new agents on disposition meaning. "What does 'qualified' mean here?" should have a documented answer. Not "use your judgment" — actual criteria. Duration over 90 seconds AND homeowner confirmed AND service need identified AND in service area. Write it down. Test on it. Revisit quarterly.
Look, I know documentation is annoying. I've put off writing disposition SOPs for months at a time. But the first invoice dispute that takes four hours to resolve because nobody wrote down what "qualified" means? That's your punishment.
Frequently Asked Questions
Who sets the call disposition — the agent or the system?
Both, depending on your setup. Agents manually select dispositions after each call (qualified, not interested, callback requested). Automated systems can set dispositions based on rules — duration threshold crossed, IVR path completed, or AI analysis of the conversation. Most operations use a hybrid: system sets a default based on call data, agent overrides if needed.
Can a call disposition be changed after it's set?
Yes, within limits. Most platforms allow disposition edits for 24-72 hours after the call ends. After that, records lock for billing integrity. Some networks require supervisor approval for changes that affect payout — like flipping a "reject" to "qualified" after the fact. Audit trails track who changed what and when.
What's the difference between a disposition and a call outcome?
They're often used interchangeably, but technically a disposition is the tag you assign, while the outcome is what actually happened. A call where the homeowner said "yes, schedule the appointment" has an outcome of "appointment set." The disposition you tag it with — "qualified" or "appointment" or "hot lead" — is your label for that outcome. Consistent labeling matters more than perfect terminology.
How many disposition codes should I use?
Five to eight for most operations. Too few and you lose granularity — lumping "not interested" with "wrong number" hides problems. Too many and agents skip thinking, pick randomly, or always choose "other." Start with the minimum that lets you distinguish billable from non-billable, then add codes for patterns you need to act on. If a code isn't driving a decision, cut it.
Resources
- What Is a Billable Call? — the payout trigger that dispositions map to
- How Pay-Per-Call Billing Works — invoicing, disputes, and reconciliation
- Pay-Per-Call Glossary — 40+ terms including disposition codes, duplicate windows, and qualification gates
- JustAnalytics — session attribution for multi-channel campaigns
- ClickzProtect — click fraud detection for paid search
Try VeloCalls for Your Vertical
AI calling + pay-per-call platform built for HVAC, plumbing, roofing, PI lawyers, Medicare brokers, and insurance. Smart routing, real-time bidding, visual IVR builder, AI conversation intelligence. Per-minute pricing — Managed starts at 4¢/min, BYOC at 2¢/min, both drop as you scale.