Workforce Management Software Suite by Phi Division Ltd, Hungary
The challenge Workforce planning which fits for call curve Number of agents in half an hour call queues above the line, free agents under the line Call curve Workforce cost is 40-60% of total operational costs of a call center. Distribution of incoming calls is not constant. Not controlled, external factors affect call quantity. Strategic service expectations (service levels, call control theories, open hours, composition of workforce, division of labour) creates the framework of workforce planning. 20-30% average fluctuation should be calculated which influences availability of call center. 12 10 8 Number of calls 6 4 2 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 interval (min) Underseating – reduced free time Overseating – growing free time Underseating – negative customer experience Overseating - no significant added value Workforce planning which fits for call curve or optimization of existing workforce for call curve? 2
Management Summary Contact center outsourcers, financial services, travel, hospitality, telecommunication providers, retailers and e-commerce companies are some of the most frequent users of contact center WFM software. Enterprises using WFM solutions to report the following achivements: Reduced time to create agent schedules by 45 percent to 90 percent Increased service levels by 10 percent to 13 percent Decreased payroll costs by 10 percent to 13 percent Decreased call abandon rates to 3 percent. (Overall call abandonment rates consistently average around 7 percent; however, the best-performing 25 percent of desks average only 3 percent abandonment, according to Gartner Measurement.) MarketScope: Workforce Management Software for the Contact Center Wendy S. Close, Tom Berg Document Type: Strategic Analysis Report Note Number: R-21-0614
System implementation Case study: workforce mgmt automation for a leading telecom service provider Manual roster making, based on historical data Chaotic seating, decreasing workforce satisfaction Varying quality parameters Human resources out of proportion to service level achieved • Flexible roster that easily follows unpredictable events • Supporting infrastructure, time of preparing roster reduced radically • Steady service level Requirement analysis, determined by planning inputs Preparing system specification System implementation and support
Architecture of main modules Phinesse DB planner Phinesse overtime Phinesse outbound Csiribiri client Csiribiri client Csiribiri client Phinesse DB roster builder Phinesse DB Csiribiri admin Application & communication server XML loader Phinesse XML planner Imitátor client Imitátor client Imitátor client Event/ message server CMS loader XML files Phinesse XML roster builder ACD CTI CMS
Integration to AIC AIC Phi- Nesse Break/ACW/consulting transfer permission Break/ACW/con sulting transfer request On line traffic data by skill Agent state actual data AIC Phi- Nesse
Making roster without WFMS A roster can be made this way as well… WFMS is most useful at 35 agents or above Depends on complexity of roster (number of skills, cross-login etc.)
With PhiNesse This way: Input data processing Output data
Input data Time period Planned traffic by skill (historical data + trends) Labour regulations Service-Level expectations Requirements of agents (ad hoc / regular) Quantity of back office activities, outbound campaignes, breaks Planning of outsourced workforce Number of seats (by skill) Agent-Skill, cross-login Parameters of shifts (length, start time, etc…)
The ideal output Should fit annual Service Level requirements. No accumulation in Back Office activities. Efficiency of outbound campaigns should be satisfactory for the procurer as well No violation of labour laws, local regulations and collective agreements Agents work in a friendly atmosphere, where their personal requests are considered in the roster. Shifts should be swapped simply and without managerial involvment between agents with the same skills.
The process of automated roster Initial, static parameters Erlang table for workforce requirement based on expected service level Prediction of traffic curves based on historical data and trends Workforce requirement calculation based on Erlang values by skill Defining forecastable personal requirements, assigning skills Defining shift plan Defining dynamic parameters Building the roster automatically
Initial static parameters
Erlang table for workforce requirement calculation
Prediction of traffic curves based on historical data and trends
Workforce requirement calculation based on Erlang values by skill
Defining forecastable personal requirements, assigning skills
Defining shift plan: parameters of shifts for each skill can be defined (start time, length of shift, required number of people)
Defining dynamic parameters and building the roster
Different views of the complete roster Global coverage
Different views of the complete roster View by agent
Different views of the complete roster What the agent receives
Different views of the complete roster What the agent receives (Excel export)
Swapping of ready rosters Shift exchanger module Swapping of ready rosters No need of management control Popular (25% of shifts is swapped)
Windows for shift swapping Client side windows of shift exchanger
Bidder Agents can choose shifts within the given possibilities
Client side of the bidder application: bidding for the shifts of virtual roster
Supervisor can review the actual state of bid
Agent state managing module Scheduling of breaks is a critical factor in call centers
Planning of breaks: overseating
Break and back-office activity schedule by agent
Break planning and control: static vs. dynamic