Running your old dialer and VICIdial in parallel during migration

A parallel run means both dialers are live at the same time, each handling a portion of your traffic. It is the safest way to validate VICIdial in production before you commit to a full cutover. The catch is that two dialers sharing a lead database will double-dial contacts unless you manage the split carefully.

Why run in parallel at all

A direct cutover bets your entire operation on a system your team has not used in production. A parallel run lets you compare VICIdial's answer rate, Wrap-up time, and agent comfort against your existing numbers before you pull the plug. If VICIdial shows a problem - a mis-configured SIP trunk, an unexpected Disposition behavior, or a Predictive dialing level that needs tuning - you catch it on 20% of traffic, not 100%.

Splitting the lead list without duplicates

The most reliable split is a hard partition: export your full Lead list and divide it into two non-overlapping files by row range or by a consistent hash on the phone number. Load the smaller portion into VICIdial and keep the remainder in your existing platform. Neither system sees the same record.

Avoid splitting by campaign name or date unless you are certain those boundaries are mutually exclusive. Shared lists with status-based filtering are tempting but produce race conditions: a Lead marked "called" on one system may not yet appear as such on the other.

DNC sync between both systems

This is the piece most teams forget. When an agent on the old dialer marks a contact as DNC (do not call), that record needs to flow to VICIdial before VICIdial's partition could ever reach the same number. Export your internal DNC list from the old system daily - at minimum - and re-import it to VICIdial's system-wide DNC table. Most dialers can export a plain CSV of DNC phone numbers; VICIdial accepts the same format via the lead loader.

Traffic split ramp and measurement

flowchart LR
    A[Full list] --> B[Hard partition]
    B --> C[Old dialer 80pct]
    B --> D[VICIdial 20pct]
    C --> E[Daily DNC export]
    D --> E
    E --> F[Import to both systems]
    D --> G[Compare KPIs]
    G --> H{VICIdial meets threshold?}
    H -- Yes --> I[Increase VICIdial share]
    H -- No --> J[Tune and retest]
    I --> K[Full cutover when 100pct]

Start VICIdial at 20% of total volume. After a full shift, compare answer rate, contact rate, and average handle time between the two systems. Because the lead partitions are different, the comparison is not perfect - but gross problems show up clearly. If VICIdial is within 5-10% on key metrics, bump it to 50% the next day and repeat.

Ending the parallel run

The parallel run ends when VICIdial handles 100% of traffic for at least one full shift without incident. At that point, freeze the old dialer, do a final DNC sync in both directions, and export any remaining uncalled leads to load into VICIdial. Keep the old system accessible but idle for one week before decommissioning it.

For a complete picture of what to prepare before starting the parallel run, see our full VICIdial migration guide. For how to handle leads during the switch, see migrating leads to VICIdial.

VICIfast provisions a ready-to-dial managed VICIdial server in under 40 seconds - fast enough to spin up a parallel test environment the same afternoon you decide to try it. Check VICIfast plans for the option that fits your team size.