Pacing Ratio Tuning: A Field Guide for Predictive Dialer Operators

What the Pacing Ratio Is, Precisely

The pacing ratio is a multiplier applied to the base dial rate derived from agent availability. At a 1.0 ratio, the system dials one call per available agent — functionally identical to power mode. At a 2.0 ratio, it dials two simultaneous calls per available agent seat. At 3.0, three.

Most platforms implement this as lines-per-agent (LPA) or dials-per-agent (DPA). If your platform uses a different label, it is the same concept: how many calls does the system attempt for each agent it expects to be available when a call connects?

The ratio is calculated against expected availability, not current availability. If 15 agents are logged in and the system projects that 12 will be free in 15 seconds (based on handle time), a 2.0 ratio fires 24 simultaneous calls. If 13 of those connect, the system needs 13 available agents. It has 12. One call is abandoned.

The Variables That Determine Your Starting Ratio

Before you set a pacing ratio, you need three numbers:

1. Connect rate. What percentage of dials result in a live human answer? This varies significantly by list type (cold purchased, warm inbound web lead, prior customer), time of day, and day of week. Run 200 calls in power mode at the start of any new list to measure this before committing to a ratio.

2. Average handle time (AHT). How long does an average connected call last, including wrap-up? If your agents work a 7-minute average call and 90-second wrap-up, your effective AHT is 8.5 minutes.

3. Agent count. How many agents are live and logged into the campaign simultaneously? This is your capacity ceiling.

A starting ratio formula used by experienced operators: 1 + (1 - connect_rate) / connect_rate * 0.5. On a 25% connect rate list, that gives 1 + 0.75/0.25 * 0.5 = 2.5. On a 40% connect rate, it gives 1 + 0.60/0.40 * 0.5 = 1.75. This is a starting point, not a target; adjust based on observed abandon rate within the first 20 minutes.

The Four Phases of a Session and How Ratio Should Change

Phase 1: Warm-up (first 20–30 minutes). Connect rates on a fresh list are typically 5–10 percentage points higher than the session average. Ratios should be set conservatively (1.5–1.8) during this phase. The model does not have enough data to predict reliably and the risk of over-dialing is highest.

Phase 2: Steady state. Connect rates have stabilized and the model has 30+ minutes of handle time data. This is where you run your target ratio. Monitor rolling 5-minute abandon rate; if it stays below 1.5%, there is headroom to push the ratio 0.2–0.3 higher.

Phase 3: List penetration drop-off. As the easy-to-reach contacts are exhausted, connect rates drop. Most operators see a 30–40% reduction in connect rate from session start to 75% through a fresh list. When connect rates start dropping, drop the ratio proactively — before the model catches up. A connect rate that drops from 28% to 18% over 30 minutes will cause the model to over-dial significantly if ratio stays fixed.

Phase 4: End of session / warm wind-down. In the last 15–20 minutes of a dialing session, drop to a 1.3–1.5 ratio or switch to power mode. You do not want to generate abandoned calls in the final minutes of a session when you cannot recover the agent capacity.

The Abandon Rate Feedback Loop

Pacing ratio adjustments should be driven by the rolling abandon rate, not the session abandon rate. Here is a practical threshold ladder:

Do not make more than one adjustment per 10-minute window. The model needs time to reflect ratio changes in actual call behavior.

The Handle Time Variance Problem

A single outlier call — an agent who spends 25 minutes on a call the model predicted would take 8 — can cascade across 15 minutes of predicted availability. During that 25-minute call, the model kept predicting that agent would free up at minute 8, kept queuing calls for minute 9, and kept generating abandoned calls.

Platforms that cap individual call duration to prevent handle time outliers (via a hard wrap-up timer or a supervisor alert at 130% of AHT) significantly improve model stability. If your platform supports this, configure it. The forced wrap-up time is not penalizing your best-performing agents — it is protecting the campaign from statistical noise.

What STIR/SHAKEN Does to Connect Rates and Why That Changes Your Ratio

In the US and Canada, calls that do not carry a valid STIR/SHAKEN attestation are increasingly being labeled as "Spam Likely" or silently suppressed by carrier analytics platforms. Depending on your carrier gateway and origination path, unlabeled calls can see connect rates 15–25 percentage points below the same campaign with full A-attestation.

If your termination carrier does not support STIR/SHAKEN signing, your measured connect rate is artificially depressed — and your pacing ratio is calibrated against a suppressed baseline. Fixing the attestation problem on your carrier path will increase your connect rate, which means your existing ratio will over-dial until you recalibrate.

UnlimCall's outbound network handles STIR/SHAKEN for US and Canada campaigns on the network side. If you are moving from a non-STIR/SHAKEN carrier, expect connect rates to improve and plan to recalibrate pacing ratios after the first session on the new network.