Outbound predictive dialling WFM
Predictive dialling inverts the inbound model: there is no queue, no arrival rate, and no Erlang C. The dialler controls agent occupancy through pacing. WFM controls staffing to give the dialler enough agents to pace efficiently without breaching drop rate compliance limits.
Note on legal jurisdiction
This guide describes UK GDPR and data protection obligations as they apply to contact centres operating in Great Britain. Data protection law varies by jurisdiction. Always verify the requirements applicable to your operation with your Data Protection Officer or legal counsel before changing data handling practices. This guide is for operational context, not legal advice.
Why inbound WFM principles do not transfer to outbound
| Dimension | Inbound WFM | Outbound predictive dialling WFM |
|---|---|---|
| Contact flow direction | Contacts arrive; agents wait and answer | Agents available; dialler places calls to fill agent time |
| Staffing model | Erlang C: forecast arrival rate → calculate agents needed for target SL | Campaign throughput: contacts to reach ÷ RPC rate ÷ hours → agent-hours needed |
| Primary SL metric | % contacts answered within X seconds | Drop rate: % live connects with no agent available (regulatory limit: 3%) |
| Occupancy driver | Arrival rate and agent count (queue dynamics) | Dialler pacing ratio (controlled by the dialler algorithm) |
| Intraday management focus | Queue depth, oldest call, SL per interval | Drop rate, agent count, RPC rate, contacts-per-hour throughput |
| Key intraday risk | Queue building — too few agents for arriving contacts | Drop rate breach — dialler over-pacing for available agents |
| Forecast input | Historical contact volume, seasonality, events | Campaign list size, RPC rate, call window hours, AHT |
Four outbound dialling concepts WFM must understand
Right-party connect rate (RPC)
Definition
The percentage of outbound call attempts that result in the intended person (the right party) answering the phone. Not all connects are right-party connects — a call may connect to an answering machine, a wrong number, a gatekeeper, or a different household member. The RPC rate is the key multiplier for calculating how many call attempts the dialler must place to generate a given number of productive contacts.
WFM implication
The staffing model depends critically on the RPC rate. If a campaign needs to reach 1,000 right-party contacts in a 6-hour window and the RPC rate is 25%, the dialler needs 4,000 call attempts to generate 1,000 live contacts. If the RPC rate drops to 15% mid-campaign (common at list end, as the easiest-to-reach numbers have been exhausted), the dialler needs to place 6,667 attempts for the same 1,000 contacts — requiring either more time, more agents, or a revised campaign target. WFM must monitor RPC rate intraday and adjust campaign pacing or staffing accordingly.
Pacing ratio
Definition
The ratio of call attempts per available agent. A pacing ratio of 2.0 means the dialler places 2 call attempts for every agent currently available. A higher pacing ratio generates more live contacts per agent-hour but also increases the drop rate (calls that connect to a live customer but find no available agent). A lower pacing ratio reduces the drop rate but reduces agent occupancy (agents waiting for live contacts).
WFM implication
The dialler's pacing algorithm manages pacing ratio in real time. WFM's role is to ensure the agent count is sufficient for the dialler to maintain a compliant pacing ratio without excessive drop rates. If agent count drops sharply (e.g. due to unplanned absence or extended breaks), the dialler must reduce its pacing ratio to maintain compliance — which also reduces contacts-per-hour throughput. WFM intraday must monitor agent count closely and ensure breaks are staggered to avoid sudden drops in available agents.
Drop rate (abandoned call rate)
Definition
The percentage of outbound calls that connect to a live customer but are dropped because no agent is available at the moment of connection. The dialler connects the call before an agent is confirmed available (to avoid agent idle time). If no agent becomes available within a short window (typically 2 seconds), the call is dropped. A dropped call may be played a recorded message or experience silence before disconnecting.
WFM implication
Drop rate is a regulated metric. In the UK, Ofcom requires that no more than 3% of live outbound calls are abandoned. In the EU and some US states, similar rules apply. WFM must ensure staffing is sufficient to keep the drop rate within compliance limits at the campaign's target pacing ratio. A drop rate breach is not just an SL failure — it is a regulatory compliance failure with potential fines. Campaigns exceeding the drop rate limit must immediately reduce pacing, increase staffing, or pause until more agents are available.
Campaign window
Definition
The permitted hours within which outbound calls may be placed to a given list. In the UK, the TPS (Telephone Preference Service) rules and ICO guidance restrict outbound calling to certain hours (typically 08:00–21:00, with sector-specific restrictions). Some campaigns have narrower windows based on product type or customer segment. The campaign window directly constrains the available agent-hours for the campaign.
WFM implication
The campaign window is a hard constraint that the staffing model must respect. If a campaign window is 08:00–20:00 (12 hours), the maximum agent-hours available are 12 × (number of agents on shift). WFM must schedule enough agents within the window to reach the campaign contact target — taking into account the RPC rate, AHT, and drop rate compliance limit. Evening campaigns require afternoon/evening shift coverage, which has separate scheduling and contract implications.
Outbound agent requirement calculation
Formula
Agent-hours required = (Contacts to reach ÷ RPC rate) × AHT ÷ 3600 ÷ (1 − shrinkage)
Where: Contacts to reach = campaign target (right-party contacts); RPC rate = fraction of call attempts that reach the right party (e.g. 0.25 for 25%); AHT = average handle time in seconds (talk + wrap); shrinkage = fraction of time agents are unavailable (breaks, meetings, absence).
Worked example
Note: this calculation gives the agent count needed to process the live contacts the dialler generates at the 20% RPC rate. If the RPC rate drops to 15% mid-campaign, the same 13 agents will produce only 375 right-party contacts in 6 hours — the campaign target will be missed. WFM should monitor RPC rate intraday and alert the campaign manager when the trajectory deviates from the required rate.
Four intraday metrics to monitor during outbound campaigns
Drop rate / abandon rate
Live connects with no agent available ÷ total live connects × 100
Target range
Under 3% (UK Ofcom limit). Best practice under 2%.
WFM action if out of range
If drop rate >3%: immediately instruct dialler to reduce pacing ratio; call agents back from breaks; escalate to supervisor to request additional agents. Investigate cause — sudden absence, break bunching, or dialler misconfiguration.
Right-party connect rate (RPC)
Live contacts reaching the right person ÷ total call attempts × 100
Target range
Varies by list type. Warm leads: 30–50%. Cold list: 10–20%. End-of-list: below 10%.
WFM action if out of range
If RPC falls sharply mid-campaign, the list is near exhaustion. Reforecast contacts-per-hour and update the campaign completion estimate. Communicate revised estimate to the campaign owner — the contact target may need to be revised or a new list sourced.
Agent occupancy
Time agents spend handling live contacts ÷ total logged-in time × 100
Target range
Target 80–88%. Below 75% = dialler under-pacing (agents waiting excessively). Above 90% = drop rate risk.
WFM action if out of range
If occupancy below 75%: check dialler pacing configuration and RPC rate. If occupancy above 90%: dialler is over-pacing; risk of drop rate breach. Reduce pacing ratio or add agents.
Contacts per agent-hour
Total right-party contacts handled ÷ total agent-hours logged in
Target range
Derived from AHT. If AHT is 6 minutes: 60 ÷ 6 × occupancy rate = contacts/hr. At 85% occupancy: 8.5 contacts/hr.
WFM action if out of range
If contacts/hr falls below model: check AHT (may have increased), occupancy (may have fallen), and RPC (may have dropped). Each cause has a different intervention. Do not assume the staffing model is wrong before ruling out campaign-side factors.
Outbound dialling WFM questions
Why does Erlang C not apply to outbound predictive dialling staffing?
Erlang C models a random inbound arrival process: contacts arrive at a Poisson rate, agents answer them, and the model determines staffing for a target service level. Outbound predictive dialling has no random arrivals — the dialler controls when calls are placed and at what rate. The relevant model for outbound is throughput: how many agent-hours are needed to handle the expected live contacts from the campaign, given the right-party connect rate and AHT? The dialler manages occupancy through pacing; WFM manages staffing to give the dialler enough agents to pace without breaching the drop rate compliance limit.
Related guides
Outbound CC staffing
Broader outbound operation staffing
Customer callback WFM
Callback-specific outbound planning
Occupancy management
Managing agent occupancy levels
Compliance guide
Regulatory requirements in contact centres
Forecast error impact
Consequences of forecast inaccuracy
Staffing models guide
Voice, email, chat, and outbound models
Occupancy calculator
Monitor agent utilisation in predictive dialling
Productive capacity calculator
Contacts per agent per day for outbound planning