Until a few years ago, contact centres were focused on handling inbound telephone calls from customers – predominantly order handling, customer service, support, appointment confirmation and of course switchboard services.
Contact centre agents acted as the intermediary between the company and its customers. In recent years, a lot has changed. Today, customers communicate with agents through a variety of communication channels, and many contact centres have been transformed into multi-channel contact centres. Customers have the choice of contacting companies by telephone, email, white mail, and increasingly, using web-chat or by pressing a ‘call me’ button from a web page.
From the company’s perspective, the move to new communication channels has benefits in terms of both differentiated customer service and cost reduction. Research has shown that, on average, the cost of handling calls is five times higher than that of handling emails. The telephone still represents the lion’s share of the communication mix, but it is also by far the most expensive channel. The business objective, therefore, should be to provide both service-oriented and cost-effective multi-channel communications.
The decision about which communications channels to make available to the customer should be predicated on the need to balance the cost of service and the benefits of high customer satisfaction.
The increasing number of communication channels, however, poses a challenge for the contact centre planner in providing the requisite number of appropriately-skilled employees at all times during the day for all channels. On the one hand, the planner should avoid over-staffing, where agents sit idle and are not being productive, while on
the other hand understaffing results in reduced service standards. Customers not only expect a high degree of availability on the telephone, but there are also high expectations concerning the response times to emails, faxes and other newer channels like SMS (text messaging) and social networks (e.g. Facebook, Twitter). Response times of several days are increasingly unacceptable for these alternative forms of communication. It is therefore desirable for companies to handle non-call, non-abandoning contacts such as email (which does not have the immediate response criteria of telephone queries) primarily during times with lower call volume, in order to increase productivity across the labour pool. Appropriate service level standards need to be defined for all contact channels.
This article explains how the workforce management (WFM) can enable planners to carry out forecasting and scheduling in a multi-channel environment and how customers may achieve the proper balance of service between channels by using service level simulations.
Definition of Multi-Channel
A multi-channel environment is typically a set of centrally managed devices for communicating with a company’s customers, including all communication technologies by which customers interact with the company: inbound telephone, document (email, fax, letter), contact forms on an internet page, call-back, chat, SMS (text), outbound telephone, IVR self-service, Internet self-service, etc.
With the introduction of multi-channel environments, the forecasting and scheduling of contact centre agents becomes more and more complex. This is due to the mix of channels and the difference between response time expectations for each of these channels. The problems are compounded when the contact centre operates across multiple sites and where the agents have multiple skills.
The resulting schedules should take advantage of the lower cost of the alternative channels, using the work to fill the expensive available time required to meet telephone service levels. However, in order to succeed in today’s multi-channel world, companies must consistently achieve their various service level goals. Today’s customers are increasingly demanding and discriminating; they expect an appropriate level of availability and swift responses regardless of the communication channel that they chose.
In a traditional contact centre, agents handle inbound calls from customers. The first crucial step in workforce planning is to predict the number of callers that will want to reach the contact centre per skill. Note that the number of callers is different from the number of calls. Traditional workforce management (WFM) systems use either calls handled or calls offered as an input to forecasting. If these metrics are taken at face value, without accounting for the customer behaviours driving the volumes, the accuracy of the forecast may be questionable. Factors such as repeat callers – abandoned calls where the same caller tries to get through multiple times – and the times of day that they are most likely to call (or abandon) must be taken into consideration, otherwise the basic forecast will be wrong, and inaccurate forecasts lead to inaccurate staffing models.
The ACD (Automatic Call Distributor) provides the workforce planner with information about the typical daily and weekly distribution of incoming calls; they show what day of the week and what time of each day customers are most likely to call (or least likely to call). Based on this data, the employee requirement is generated with respect to service level goals and shifts are assigned to available agents.
Generally, the algorithms used for the calculation of the employee requirement are based on a modified Erlang formula. Erlang-C is a mathematical method to plan for randomly incoming workload (e.g. calls) which is based on known information (e.g. average duration of a call, service level response times). It returns the number of employees required to handle the workload in order to meet the service level response times. However, the Erlang formula is considered by some to be outdated as it is based on some false assumptions, such as the assumption that callers will wait forever in the queue. If you don’t consider that a portion of your calls will abandon, then your staffing calculations will be based on an inflated workload, which would lead to overstaffing.
New versions of the Erlang formula are being developed to cope with the limitations of the traditional Erlang formula and to provide the most accurate forecasting tools available.
Inbound calls are very different from documents (i.e. email, fax, white mail) and outbound calls. Inbound calls (and other ‘abandoning’ media such as chats) have to be handled (answered) with near-immediacy, often within as little as thirty seconds. They cannot be postponed to another point of time. If they are not handled within the defined time period, e.g. 80% of the calls answered within 30 seconds, the service level will be missed. When telephone service levels are missed, the result is high abandonment; but those callers who hang up in queue due to extended wait times cannot be neglected. They will call back, which increases the overall call volume, and/or contact the company through alternative channels, such as email. What’s more, those emails are likely to be duplicates of issues that are open in the contact centre, leading to redundant work. If the work is not completed in a timely manner, customer churn is a likely outcome, so top-line revenue is affected.
Outbound campaigns are characterised by systematically calling existing or prospective customers in order to actively sell products, collect overdue payments, gather statistical data, perform customer satisfaction surveys, verify address lists, etc. Outbound campaigns are usually performed on a given contact list. Calls are made within
a specific time frame, and they require a unique agent skill set. Nowadays, predictive diallers enable efficient management of outbound calls. These are contact-centre specific software and/or hardware modules which automatically dial pre-defined lists of telephone numbers. Usually the call is immediately put through to an agent as soon as connection has been established. WFM can seamlessly integrate into most diallers and import the data required to create efficient schedules.
When calculating outbound staffing needs, the WFM solution also takes into account the parameters listed below:
- Gross Connect Rate – All contacts made by the dialler
- Net Connect Rate – % of the calls that are put through to an agent
- Right Party Contacts (RPC) – Subset of net connect rate, contacts where the person the call aimed at has been reached
- Wrong Party Contact (WPC) – Subset of net connect rate, contact where the person to whom the call is addressed is not reached (answering machine if not detected by a dialler, wrong number/wrong person answers)
- Number of call attempts – How often will the dialler try to call a contact before giving up Average Handling Time (AHT) Talk time plus wrap-up time plus probably a break between two calls; this can be split up into its components
- Talk Time – Time actually spent talking to the contact, different for RPC and WPC, and may be different during day times
- Wrap-up Time – Time needed after the call to note down the call details, can be different, or the same as talk time
Documents (Emails, Faxes)
Many customers nowadays choose to use email to contact companies and thus they are independent of the contact centres’ business hours. The advantage for the company in receiving contacts in written form is that the emails do not have to be handled immediately. However, there are customer expectations concerning response times and frequently service level targets or agreements that have to be met. Failure to meet these service level agreements can result in financial penalties, e.g. those imposed by government regulators or contractual penalties in the case of contact centre outsourcers. This service level is calculated not in seconds as for calls, but in minutes, hour or even days.
If not all documents have to be handled the day they come in, a backlog results and must be properly managed. This means that, when calculating the employee requirement for non-abandoning work such as correspondence, both the correspondence forecast still to arrive and the existing correspondence backlog have to be considered. It is important that the first in/first out principle is used, which means that the processing of the documents needs to be performed according to the order of arrival. The processing times (average handling times) and the required service level target determine the staffing needs. If employee productivity is less than what has been planned, not all documents will be processed and a backlog will be created that may lead to more required staff the following day in order to achieve service level targets. Proper planning allows employee utilisation to be kept on a constantly high level to ensure the achievement of cost goals.
Web Site Functionality
The web call-back function provides the possibility for a visitor to a web page to submit a request to be called back at a defined time, perhaps about a specific topic. These call times are determined by the customer and are hardly predictable for the planner. At short notice, e.g. if the customer wants to be called back the same day, they can only partially be taken into account by intra-day management. Over time, however, a pattern of web call-back requests will form, and this can be forecast in exactly the same way as inbound calls.
By using chat applications, customers can communicate with a contact centre directly from the company’s web page if they want immediate online assistance. Depending on the issue, the chat queries are forwarded to a suitably skilled agent, who can respond immediately. It is usually possible for an agent to handle chat threads with several
customers at the same time, which increases agent productivity and drives down the number of agents needed to service the workload. Depending on the nature of the contact centre’s business, we think that one agent can handle between 3-5 chat sessions simultaneously. Internet functions such as these also provide real business opportunities.
Research has shown that high chat availability can help increase online sales, helping to grow top-line revenues. High chat availability redirects a portion of the more expensive incoming calls to the more cost-effective channel, which lowers the overall cost per contact. Like call-backs, chat queries are also unpredictable for the planner,
who has to rely on his or her experience when scheduling employees for this purpose. Again, a pattern of chat volume will form over time, and this can be forecast in exactly the same way as inbound calls.
Co-ordinated Scheduling of Channels
The challenge for the planner is to co-ordinate all the channels within the scheduling process. Usually, there are not sufficient employee resources at any given day and time to meet the full staffing requirements of all channels. Therefore, resource planning has to be fine-tuned every day or even several times a day, in order to adapt to the current conditions as they arise. The staffing of the different activities is subject to a state of permanent re-optimisation. WFM is a tool that gives planners access to the information they need to make the right decisions on balancing these channels to ensure that each is staffed according to corporate policy.
Each communication channel plays a role in a company’s marketing strategy, but the relative importance of each channel may vary. Therefore, the planner needs guidelines from management about which communication channels should receive high priority and which should receive lower priority.
By simulating possible schedules using different staffing models, prioritisation of tasks, and also modified service level targets, the planner can measure the impact of the parameters on the schedule. The planner can also see how the service level goals on some tasks can be achieved by relaxing the goals on other tasks. If there are not enough employees available to staff all tasks according to the requirement, the planner can weight and co-ordinate the contact channels by simulation. By doing so, the most important (service level) goals of the company can be achieved.