Key considerations in post-interaction surveying: which data collection channel is right for you?

It is a common question in marketing research: "Which data collection channel is right for my program? Telephone, Web or IVR?" If you are talking about a contact center post-interaction survey, the answer may not be a simple one-size-fits-all solution. Each data collection channel has its own pros and cons. There are three sequential steps in the design phase of a contact center survey program:

* Developing the survey based on goals and parameters such as budget;

* Comparing each channel according to its merits and shortcomings; and

* Putting the channel you choose into operation.

Developing The Survey

Your research goals will dictate your first step in selecting a data collection methodology. These goals range from providing a valid and reliable read of customer satisfaction at the enterprise level to driving results to the agent level for one-on-one coaching. In the case of gaining an enterprise or even regional view of the company's overall performance, telephone and Web each offer multiple benefits. Due to cost considerations, however, these channels may provide a relatively low number of completed surveys (e.g., 400). If your goal is to complement a contact center quality assurance program with agent-level "voice of the customer" feedback, IVR is a cost-effective option. However, whereas this type of program offers a large number of surveys (e.g., 40,000), IVR data may not be consistently representative due to low response rates and polarized customer ratings.

Some other considerations include:

Customer segment. Which customer type is the focus of your research? Different customer segments have different "channel preferences." Some people prefer to be contacted over the telephone while others prefer the Internet. If you are surveying respondents in an older age segment, a telephone approach is ideal. If you are targeting technically savvy respondents, Web or IVR both work well. Because different segments of the population have different reactions to technology, be cognizant of these variations when designing your approach. One quick solution is to survey respondents using the same channel through which they contacted the company.

Budget. Many companies prefer the "high-touch" option offered by phone-based research programs but are prevented from exclusively using this channel due to budgetary constraints. IVR and Web both offer the appeal of relatively low investment compared to the phone. Beware of hidden costs in IVR implementations that may not be included in your early budget, however; the technology required for telephony and CTI integration as well as labor resources can add significant incremental expense to your program. As an alternate low-cost option, one growing trend is to capture telephone-based survey data in a facility located offshore. This approach offers the benefits of telephone data collection at a significant savings, sometimes as much as 40 percent.

Performance incentives. Increasingly, companies are relying on customer feedback as a key indicator of contact center service quality. In an effort to drive results to the frontline agent, performance incentive programs are established whereby individuals receive positive recognition or, in some cases, monetary bonuses based in part or in whole on customer feedback. When compensation is at stake, it is especially important to choose a defensible and unbiased measurement system. The need to tie results to individual employees can make automated techniques like IVR ineffective, especially in a multicenter model, because most locations do not share identical technology infrastructures, thus creating an obstacle in "passing" agent or customer information out of the system.

Linkage analysis. Many times, customer feedback alone is insufficient for answering all of a company's information needs. An effective technique for providing more in-depth exploration is through linkage analysis, a technique where customer scores are statistically "linked" to other operational or quality data sources. To enable a linkage approach, key identifiers like the agent who handled the call and the reason for contact must be easily "passed" out of the data systems in which they are warehoused. As noted above, the complexities inherent in an IVR solution can prevent the ease of establishing these linkages; by comparison, these obstacles are minimized for phone and Web surveys.

[ILLUSTRATION OMITTED]

Survey length. Another critical consideration in designing a marketing research program is respondent fatigue, or the point in the survey at which respondents simply stop answering questions because they are bored or have become otherwise engaged. Market researchers must continually balance the need for in-depth insights against the reality that customers won't spend a lot of time completing a survey. To that end, each of the three data collection channels discussed above offers varying success in expanding the survey length without affecting fatigue levels. Because the drop-off rate is very high with IVR technology due to lack of familiarity and low comfort levels, IVR surveys must be much shorter in length, and as a result offer "quick-hit" insights rather than more in-depth results. In contrast, Web and telephone allow for longer surveys with more attributes, deeper insights and probing of comments.

Timing of survey invitation. IVR offers the appeal of invitation immediacy. The immediacy of the survey--which is conducted as soon as the contact concludes--is commonly viewed as being critical to ensuring accurate respondent recollection. For surveys via phone or Web, the invitation may not reach the respondent until hours or even days after the contact.

Selecting The Channel

With the design in hand, now you are in a better position to make informed decisions about the channel. Table 1 is an overview of what each channel offers.

Operationalizing The Channel You Choose

How will each data collection channel be operationalized in a contact center environment? Depending on which channel you select, your end customer will experience something different.

Interactive voice response (IVR). The key to using IVR technology in survey research is selecting the invitation approach. There are two primary invitation strategies in contact center environments that are commonly used for IVR surveying (see box). It's important to note that, regardless of invitation, all agents must disconnect the call before the customer does which requires retraining of the agent population.

Once the customer is placed in the IVR survey, he or she proceeds through it by providing responses using the keypad and/or speaking into the mouthpiece.

If technology is inconsistent across contact centers, it can be extremely difficult to consistently "pass" agent or customer identifiers--critical

for back-end analysis and one-on-one coaching--from the front-end IVR to the IVR survey. This is the most common point in the automated invitation methodology at which the process breaks down.

With an agent-initiated invite, the actual invitation process should be closely watched by operational managers to ensure agent adherence and compliance; otherwise, there is no guarantee of an unbiased sample. This is particularly important when survey data are used in performance incentives programs.

It is always critical to remember that any time the agent is directly involved in the survey invitation process, biases will enter the process.

Telephone. Telephone data collection offers a high-value, high-touch option for capturing customer feedback. It enables deep analysis into drivers of customer satisfaction. In a telephone-based survey environment, the process usually follows a common series of events: Within hours of the completion of a contact center experience, a customer receives a call from a research interviewer inviting the customer to complete a survey over the phone. If the customer agrees, the interviewer guides the customer through a survey that includes questions designed to solicit the customer's perceptions of his or her contact experience. The interviewer's role is to accurately capture the customer's answers and probe for additional insight where appropriate. (Note that in this scenario, a "research interviewer" is a professional trained in marketing research practices. It is not a good idea to use contact center agents to conduct marketing research surveys as this will invariably bias the data.)

Web. Many companies now capture e-mail information as part of broader marketing strategies, which allows an e-mail follow-up invitation to be considered in the measurement of contact center experiences. Web-based surveys are conducted by sending customers who have contacted a center an e-mail inviting them to provide feedback via an online survey about the service they received. Flexibility exists to establish time-sensitive rules as to when the survey invitation is sent as well as contact rules that control the number of times an individual can be invited within a given time period. E-mail invitations are customizable to differentiate them from spam. Within each e-mail invitation, a URL is embedded that customers click to hyperlink into the survey. Each customer is assigned a unique pass code and login, ensuring that only the invited customer completes the survey and preventing customers from completing multiple surveys. For non-responding customers, a reminder e-mail is sent after a certain period of time from the initial mailing.

There is little question that the use of marketing research surveying to assess the quality of a contact center interaction is invaluable. Within the complex contact center environment, however, several important factors must be considered before selecting the most appropriate data collection methodology for your program. The three most common data collection channels--telephone, IVR and Web--each offer benefits and tradeoffs depending on the goals of your program and the level of insight you require. In the event that you need a valid, representative assessment of performance at the enterprise level, then telephone or Web are ideal. If you are looking to complement an existing quality assurance program with "voice of the customer" feedback and are less concerned with data representativeness, then IVR is a great option.

Mike Cholak is a Convergys vice president in the company's Customer Care business, where he is responsible for Customer Intelligence Services. Convergys (www.convergys.com/research.html) helps clients optimize their customer management operations through a portfolio of professional services that enable companies to achieve their full business value potential.

If you are interested in purchasing reprints of this article (in either print or PDF format), please visit Reprint Management Services online at www.reprintbuyer.com or contact a representative via e-mail at [e-mail] or by phone at 800-290-5460.

By Mike Cholak

Convergys

RELATED ARTICLE: Two Options For Inviting Customers To An IVR Survey

Option 1: Automated

An automated invite is played as part of the front-end IVR menu system. A subset of customers is randomly offered the chance to participate in the survey. The IVR system automatically transfers a customer who agreed to participate in the IVR survey after that customer's service transaction is complete. Agents must disconnect the call before the customer does.

Option 2: Agent-Initiated

Agents offer a pre-scripted invitation to the customer at the conclusion of the transaction. If the customer agrees to participate, the agent transfers that customer into the IVR survey system. Another option is for the customer, at the completion of the call, to ask to be transferred to the survey.

          Telephone             IVR                 Web

Timing    Easily implemented.   More difficult to   Easily implemented.
          Typically 2-3 weeks   implement.          Typically 2-3 weeks
          to launch. Does not   Typically 3-4       to launch. Does not
          require additional    weeks to launch,    require additional
          resources for         can be longer if    resources for
          integration into      automated           integration into
          IVR or agent          invitation used.    IVR or agent
          desktop.              Can require         desktop.
                                significant
                                resources to
                                integrate into IVR
                                and/or agent
                                desktop.
Survey    Flexible design.      Less flexible       Flexible design.
          Longer survey (8-10   design. Short       Longer survey (10-
          mins.) provides       survey (2-5         12 mins.) provides
          richer insights.      mins.). Limited     richer insight.
          Readily incorporate   question types.     Readily incorporate
          call metrics (ASA,    Limited ability to  call metrics (ASA,
          AHT, agent, site)     incorporate call    AHT, agent, site)
          into analyses.        metrics (ASA, AHT,  into analyses.
          Verbatim comments     agent, site) into   Verbatim comments
          rich due to probing.  analyses.           rich due to
          No transcription      Verbatim comments   probing.
          requirements.         restricted in       No transcription
                                length.             requirements.
                                Transcription is
                                required.
Response  Highest (18-25%).     Lowest (1-5%). (1)  Higher (15-20%).
Rate
Results   Representative.       Highly polarized,   Representative.
                                skewed toward
                                positive.
Sampling  Proven                Limited             Limited
          representativeness.   representativeness  representativeness
          Conveys high value    due to low          due to lack of
          to customer. First    response rate.      e-mail addresses.
          attempt made within   Immediately         Invitation within
          hours of customer     following customer  24 hours of
          interaction. Can      interaction.        customer
          prevent respondents   Customer fatigue    interaction. Can
          who recently          (with longer        prevent respondents
          completed a survey    AHTs). Agent-       who recently
          from participating    initiated invite    completed a survey
          again. Easily         introduces bias.    from participating
          control all quota     Cannot prevent      again. Limited
          groups.               respondents who     ability to control
                                recently completed  monthly quotas by
                                a survey from       agent.
                                participating
                                again. Limited
                                ability to control
                                quotas, especially
                                agent-level.
Cost      Less cost-effective,  Moderately cost-    Most cost-effective.
          moderate with         effective.
          offshore option.

(1) IVR response rates are higher for agent-initiated invitations,
ranging up to 20 percent or more.

Table 1.