3 Secrets To Automated Call Handling Success

1. Test, 2. Test, and 3. Test Again


Why Test?

• To complete more calls,
• To reduce the cost per call,
• For a positive public perception and reception,
• To optimize call center performance,
• To perfect your automated call handling system.

Stated simply, to achieve optimum results from an automated call handling system, “market testing” is a fundamental requirement.

What Does “Market Testing” Mean?

Any testing of your automated voice system must include ALL callers, or a statistically valid subset, regardless of whether they were served or disserved by the automated system. This “Market Testing” is similar, incidentally, to testing performed at many live call centers. Testing is repetitive and not a one shot deal. For whatever reasons, implementers of automated voice systems have avoided this regimen.

Yes, there are those that profess to routinely test their automated centers by having staff members call to test each of the branches of a program tree to check that the program and equipment work as designed. Those tests, which ConServIT also routinely performs, are NOT market tests.

Others test their automated systems to obtain data for internal reports. Information gathered typically includes caller anecdotal evidence and testimonials, surveys of callers that completed the call progress labyrinth (but omitting those that hung up or opted out to an operator), panel studies, and internal management opinions. These, as well, are NOT market tests.

Internal report testing is often perceived as a short cut for the more rigorous market test. A large, well known retailer used such a study to justify expansion of their trial automated order taking system to supplement a live call center. The report showed management that 25% of the callers preferred the automated system and 50% had no preference one way or another, so that, they summarized, 75% either preferred or had no preference regarding using the automated system. Sounds good! Right?

Wrong! The analysis was flawed because it excluded the callers that hung up on the automated system. If the study had included all callers, rather than ignoring those that hung up, the results would have shown over 60% found the organization’s automated system unacceptable, only 25% had no opinion, and less than 15% found it preferable. The results were skewed to support management’s objectives in preparing the report.

Let there be no doubt that the perception of the public as to how an organization’s calls are handled is a marketing issue, not addressed by either technological tests or tests to support management’s perspective.

How Is A Market Test Conducted?

Market testing is necessary to achieve a high “Yield” (Yield is the percentage of all legitimate callers that successfully complete their calls). ConServIT’s natural speech technology (NST) uniquely, first and foremost, analyzes calls where the caller was NOT satisfactorily served, such as those when callers hang up prematurely.

If a problem is found with the program (a “program” includes more than the script and the call progress path), we drill down through other tests to find out what is happening to callers at that point in the program. For example, if callers are hanging up during the opening prompt, we perform drill down tests to find the reason so the problem can be resolved.

The testing becomes more difficult as the “yield” increases. To increase the yield from 30%, for example, a sample of two hundred callers might be sufficient, because in that sample, the 140 callers that did not complete their calls would make clear the problem needing to be corrected. To increase yield above 90%, however, would require a sampling of at least 1,400 callers to analyze the same number of lost calls.

A script improvement intended to reduce the calls lost for a particular reason will likely be offset by callers lost for some other reason. Any program or script change is a trade-off – gain a few and lose a few - hopefully, but not always, resulting in a net increase in the yield. As the yield increases, the trade-off issues become more complex, requiring even larger sampling sizes.

To achieve its design yield of over 90%, ConServIT’s analysis focuses on the unsuccessful calls. Users of other automated systems are satisfied with low yields, rationalizing their use of an automated system on a unsubstantiated premise that all callers who wanted to use the system were able to use it or on the likewise unsubstantiated premise that, over time, users will become familiar with their system.

The Technological Hurdles

ConServIT’s testing regimen is more complex than it may seem at first glance. How do you find out why callers hang up before they have been identified? How do you find out why callers had problems or what the problems were when, within seconds after hanging up, callers have forgotten details about their call experiences, thereby rendering follow-up surveys ineffective. The capability to perform a market survey involves a technology and technique that ConServIT has developed over the years, necessary to attain the levels of caller satisfaction for which it is known.

If, even with these resources, ConServIT finds its drill down capabilities still too limited, a sampling of callers may be answered live, using the same script as the automated system. Because the NST script replicates how a call is handled live, ConServIT can find out what happened at the moment the caller event occurs and design the appropriate resolution. In contrast, with generic automated systems, it is impossible for an agent to say to a caller the verbiage of the automated script. The agent will inherently revert to his or her own natural speech, which, for technological reasons, the ASR or IVR system cannot replicate.

It is important, as well, to keep track of call disposition in the day-to-day processing of calls. The universe of callers changes for a variety of reasons that may require, on short notice, scripting changes or other corrections. Those shifts can be picked up immediately with ConServIT’s call processing and reporting procedures.

Interesting, But What Is The Bottom Line?

The calls completed can rise by 50% to over 300%.[1] This has now been repeatedly shown when callers programs are transferred from another automated system to ConServIT’s facilities, without any change in the call volume of the universe of callers.

The cost per completed call is significantly decreased because the equipment needed to process the calls, equipment capacity, and supporting services are all decreased when an application has been properly developed using market testing techniques.

The bottom line is more completed calls at a lower cost per call. The bottom line is a more successful automated call handling system. The bottom line is improved public perception and increased use of your call center facilities.

Please call at 847-265-4901 or email me at theis@conservit.com if you have questions or would like to discuss this topic further.

Sincerely,

Peter F. Theis
President

[1] From a yield of 30% to 90% is a 300% increase, and from 60% to 90% is a 50% increase.

What Would You Charge To Handle Our Calls?

How To Get a Meaningful Quotation


The relationship between a client and potential outsource call center frequently begins with the question,

“How much do you charge to answer a given number of calls per month?”

to which the call center would respond:

“So much per hour”.

The response is not only non-responsive to the client question, but meaningless as well. The client that accepts a quote based on a price for a time period is not much better served than not having any quote. Acceptance of the cost per time period is losing the negotiation at the opening bell. It is the equivalent of writing a blank check.

The cost per hour response wins the competitive bid for the call center without any real commitment. This figure is so non-descriptive and vague that it allows the provider wide latitude to adjust its costs and revenues to maintain profit margins. The quote based on the cost per hour ignores many significant factors that cause client fees to mount or diminish caller service levels, such as:

a. Callers being placed on hold while the billing clock continues to accumulate time;

b. Callers hanging up while on hold;

c. Callers hanging up while being served by touch tone or word recognition machines;

d. Callers calling multiple times;

e. Inefficient operators (including operators intentionally managed to be slow) and machines (such as those that repeat an account number and ask if it is correct ad nauseam) that take much longer than needed to serve a caller;

f. Operators that engage callers in non-relevant dialog (whether intentional or otherwise), such as the weather and “where you are”;

g. Inadequately prepared or inadequately managed operators that take unnecessarily long because of poor training, inadequate resources, minimal supervision, or slow computer systems – busy periods being an opportunity for a call center to increase revenue and cap costs;

h. Assigning to a client a call center owned phone number previously assigned to a high volume client whose callers were continuing to call the number, but for which the new client would be charged; and

i. The use of automated systems that redirect an inappropriately large portion of callers to more expensive live agents to intentionally increase call center revenue.

Eliminating the Provider Controlled Variables

Take a second look at the above factors. Notice that every one is under the immediate control of the outsourced call center. Bid pricing can be founded on offering a low price per hour to win the business, but either through increasing revenue or decreasing service levels, profit levels can be maintained within a wide range of per time period quotations.

The way to obtain a legitimate quote is to ask for the price for each completed transaction, a transaction generally being a record having sufficient information to have value to the client. If a call center won’t provide a cost per completed transaction, the call center either doesn’t know its business, or if they know their business, they want wiggle-room to bid low and bill high.

The client does need to provide the call center with the information it needs from the caller, the mix if there are multiple types of calls, and the anticipated call volumes for each type. The professional call center can do the rest to provide a cost per completed transaction.

In the case of the typical outsourced automated call center, quotes are almost universally based on cost per time period plus a second amount for “transcription”. This fee structure can be particularly abusive for clients by covertly passing to the client costs for the above referenced risks.

ConServIT, on the other hand, prefers to quote based on the cost per completed transaction (Because ConServIT does not “transcribe” caller information, there is no additional charge for transcription.). ConServIT’s fees are simply the cost per transaction multiplied by the number of transactions completed. Let’s look at a simplified example showing how this works:

Client: “How much does it cost?”

ConServIT: “The cost per transaction is ninety cents.

Automated Call Center: “The cost is twenty five cents per minute, plus forty cents per transcribed call.

For a client wanting to obtain the names, addresses and email addresses from 10,000 callers each month, which alternative is a better deal for the client?

ConServIT: Assuming ConServIT converts 90% of the 10,000 calls into completed transactions, ConServIT would charge $8,100/month[1] and provide its client with 9,000 completed transactions.

Generic Automated Speech Provider: Assume only 50% of the callers complete their calls (hangups, transfers, etc.) and that the average incomplete call is one minute in duration – for the cost of $1,250[2]. The remaining 5,000 transactions take three minutes each, for a time charge of $3,750[3] and a transcription charge of $2,000[4].

Summary of Results:

The cost for handling 10,000 calls using ConServIT, based on a quote per completed transaction, would be $8,100 per month for 9,000 transactions, or $0.90 per transaction.

The cost for the generic automated IVR/ASR service would be $7,000 per month for 5,000 transactions, or $1.40 per transaction.

Valuing Lost Transactions:

But this is still not a level playing field, since ConServIT captured almost twice as many transactions as the generic automated system.

Losing transactions costs money, figured either based upon the promotional cost to get the caller to pick up the phone and call, the customer service value of the transaction or the revenue generated. For this example, let us assume a value to the client of only $10.00 per call lost.

For ConServIT, the 1,000 lost transactions would have a value of an additional $10,000[5]. Adding that cost to the cost for handling the 9,000 transactions completed, the cost is $2.01 per transaction[6]. For the IVR/ASR, which bid only charging $0.25 per minute, the value of the lost transactions would add an additional $50,000[7]. Adding that to the cost for handling the 5,000 transactions completed, the cost is $11.40 per transaction[8].

Regardless of whether the value placed on lost calls is a fraction of a dollar, or in the three digits, the cost value of the calls lost or diverted becomes paramount.

As an aside, when ConServIT replaces an existing generic IVR/ASR system, the number of completed transactions routinely increases from 50% to 300% and, yes, sometimes even more. The increase is not from any increase in the number of client calls ConServIT’s receives. It results exclusively from the number of calls that are converted into complete transactions by ConServIT, but were lost by the generic voice system.

Being charged per transaction also is an incentive to the provider to maintain the level of the “caller experience”. If one vendor successfully completes 90% of all transactions, while a second vendor is in the 50% range, this methodology will reduce the revenue of the poorly performing vendor, and increase its costs. This transaction based billing is largely self policing, protecting the client from low bids and high bills.

Budget Controls

We have had many clients whose promotional campaigns were successful beyond their wildest imaginations – and, because of that success, their costs exceeded their budgets. Clients, where a budget is a concern, should confirm that the outsourcer has the ability and is willing to limit the number of callers in a way other than discontinuing the service. Budget overruns, regardless of the value of the success, is often an overriding client issue. The only client more upset than one that does not receive an anticipated call volume is the client whose program is so successful that, because of its success, it is invoiced for sums exceeding its budgeted amount.

If you have further questions, or want a quote from ConServIT, please call me at 1-877-503-7131 or email me at theis@conservit.com.

Yours truly,

ConServIT

Peter F. Theis
President

[1] 90% X 10,000 calls X $0.90 per completed transaction.
[2] 50% X 10,000 calls X $0.25 per minute.
[3] 50% X 10,000 calls X $0.25 X 3 minutes.
[4] 5,000 calls X $0.40 per call.
[5] 1,000 transactions lost, each having a value of $10.00.
[6] $8,100 for transaction costs plus $10,000 for the value of the lost calls, the total then divided by 9,000 transactions.
[7] 5,000 transactions, each having a value of $10.00.
[8] $7,000 for transaction costs plus $50,000 for the value of the lost calls, the total then divided by 5,000 transactions.

Is Caller Experience Irrelevant For Call Automation?

Does Anyone Really Care About the Caller?

Excellence - The Bigger Issue

An organization has to care about all its callers, not just the callers who happen to reach a live center. Even with an acknowledged Best of Class live center, overall performance becomes mediocre if half the callers hang up trying to get through the automated front-end system. If excellence is to be a call center objective, it is necessary to encompass the entire spectrum of callers, and not be limited to those callers selected as representative by management.

In this day and age when there are such incredible pressures to cut costs and so much hype about this or that new, better, or improved technology, it is often difficult to relate cost savings to its impact on overall excellence.

Technological advances for call centers tend to be implemented not to do a better job at a lower cost, as with other industries, but rather to do a lesser job to achieve lower costs. The caller experience has been pushed off the sponsor’s radar screen, while expense reduction and technology have moved front center. For overall excellence to even be a consideration, caller experience and expense have to both be on the screen, with the caller experience paramount and front center.

The Double Standard

Corporate acceptance of the disparity in caller experiences between calls handled live and calls handled automatically never ceases to confound me. This double standard seems to be accepted by the corporate world as inevitable for some unexplained reason. A caller is a customer, or prospective customer, regardless of how the call is handled. There is no rational reason for accepting different standards, on one hand when a call is handled live by the sponsor and on the other hand, when it is handled by interactive voice response (IVR) or automatic speech recognition (ASR) systems.

Callers seldom control the path of the dialog. The path is designed and controlled by the sponsor. Logically, the quality of experience from the callers’ perspective should be the same regardless of whether the sponsor leads the caller directly to a representative or to an automated system. Whether the difference is acceptable needs to be addressed.

A short while ago, the speaker at a presentation by one of the region’s largest financial organizations expounded at length upon the success of the efforts to maximize the caller experience of their live call center. The speaker, as a passing reference, boasted that 75% of their calls were handled automatically without the caller ever needing to speak to a representative. The speaker subsequently acknowledged not having any information about how well automatically handled callers had been served. The organization didn’t seem to know, and didn’t care about callers diverted to its automated systems.

It is counter-productive to invest in improving the caller experience for callers served live while simultaneously downgrading the caller experience for those diverted to automated systems. To portray an image of call handling services by only measuring live services, while ignoring similar measurements about callers handled automatically, is misleading and hypocritical. The bottom line is that the call center should be viewed as a single entity regardless of how callers are handled. Such an analysis may show results contrary to a sponsor’s intended objectives of improving services and lowering costs, reflecting instead reduced services at higher costs!

A double standard is a non-issue for ConServIT’s natural speech technology (NST) because of the inherently similar experience for the caller served by a live center and ConServIT call center. It should be mentioned, however, that the standard using NST would generally exceed the standard for the live center. That is why ConServIT can guarantee that it will outperform any other alternative!

Does Anyone Care About The Caller Experience?

Peruse the industry articles and news releases. There is a never-ending litany of neat stuff going on with automated speech. Read about how pleased the sponsors are, how much was saved, how great the technological leap, and how well the companies marketing the devices and services are doing. But there is no one proclaiming about how the poor caller is fairing! Only at social events does the objectionable side of telephony automation have a public hearing (Just the mention of being a participant involved with automated call handling often necessitates a prompt flight for cover).

Bid specifications generally ignore caller experience requirements. Rather, the requirements relate to technical details, connections with databases, reliability and benefits for the sponsors. The entire perspective is on the technology and whether the proposed system will operate according to those requirements. “What operating system does it operate under?” “What programming language do you use?” “Whose platform?” Those are the questions asked. They certainly are not caller-centric.

Return-On-Investment (ROI) is the foundation upon which purchase justifications for voice techniques are based. The value of the caller experience is not even a line item on the cost justification spreadsheet. But if ROI is king, it doesn’t take a rocket scientist to conclude that the highest ROI can be attained by eliminating telephone services entirely. Use a cheap machine to tell callers “We no longer accept phone calls.” That is the ideal way to maximize savings and minimize investment to achieve the highest possible ROI - and it, too, totally ignores the caller experience. Who cares?

Occasionally, firms will survey callers whose automated calls were successfully concluded. Callers hanging up prematurely are excluded from the survey results, under the premise that the caller was not “dissatisfied” but made a voluntary election to bail out early. But, curiously, when callers bail out in the middle of a live call, it is considered a manifestation of extreme dissatisfaction – the opposite of the presumption for the same action for a machine-handled call. It just doesn’t make sense to make contradictory assumptions about a pool of callers sharing common characteristics.

A major retailer based a purchasing decision for automatic telephone order taking equipment on survey results showing that three quarters of the callers either liked the pilot automated systems or didn’t care whether their call had been handled live or by a machine. They rejected as irrelevant information that half of the callers hung up during or at the beginning of the call. If that information about callers bailing out had been included as manifesting dissatisfaction, as it did, the survey would have shown that only one in eight callers expressed a preference for the automation, only two out of eight of the callers did not care whether the call was handled live or automatically, and a whopping five of the eight callers expressed a distinct dislike. The service, incidentally, never got off the ground in spite of a major capital investment to acquire the equipment.

Perhaps the reason the caller experiences for IVR and ASR systems have been ignored may be that the results are not encouraging. Industry practices of referencing call volumes, anecdotal success stories, and analyzing only those calls that were satisfactorily completed is not reflective of the overall picture.

ConServIT’s NST Establishes Standards For Excellence

ConServIT’s NST establishes a consistent, objective standard of caller experience against which all other services can be measured. ConServIT’s standard is an objective standard, the yardstick of the caller experience against which the effects of internal cost saving changes can be measured. The standard is not dependent on the agent, the time of day, or any other variable, because all calls are modeled to handle callers just as an agent would. If a caller hangs up prematurely, those calls are categorized based on the reason. Were the callers improperly serviced, objecting to machines, dialing wrong numbers, placed on hold too long, playing games, kooks, repeating calls, etc.?

ConServIT has even encouraged live call centers to divert a small fraction of the calls to its facilities so that a set of statistically solid standards can be developed for the overall call center. That exercise forces any call center to define for its particular situation and objectives what excellence means and the elements of that standard of excellence, rather than relying on glittering generalites to obfuscate the lip service given to caller experience.

If your company is dissatisfied with the caller experience using IVR and ASR, and wants superior results, call ConServIT. As a general rule of thumb, ConServIT will improve the results experienced using IVR and ASR by 50% to 300% or more. We do better than having callers serviced by off shore providers as well. And now you can prove NST does better without making major commitments.

Call me at 1-877-503-7131 or email me at theis@conservit.com if you have questions or would like to discuss this further.

Sincerely,

ConServIT
a service of Conversational Voice Technologies Corporation

Peter F. Theis
President

Revealed: The Secret of Natural Speech Technology

Keep it Simple.

An observation frequently expressed about a call handled by ConServIT’s natural speech technology (NST) is “It was so easy. So simple. What’s the big deal?”

Simplicity is the Big Deal. An NST handled call is so remarkable because it is so un-remarkable. Amazing!

An NST handled call is no more noteworthy than a routine live call. No one extols the virtues of an operator handling a routine call routinely. It doesn’t happen! Instead, a routine call becomes remarkable when the caller has a negative experience.

The simplicity and ease with which callers are serviced using NST is a tremendous deal – requiring a leap in technology and common sense in scripting. For those into PCs, it is as much a big deal as going from a DOS based system to Microsoft’s Windows, with its icon based user interface.

Simplicity. That is THE secret of ConServIT’s NST. But simplicity doesn’t mean that an NST program can’t be incredibly complex and lengthy when looking at the call progress, branching and programming. Nor does it mean that thought provoking, open-ended questions can’t be asked. But what it does mean is that, to the caller, the call progression is simple, obvious, and intuitive, like falling off a log. A successful NST program has to be caller centric, first and foremost.

But that simplicity can often be difficult and complex to attain. Anyone can create a complex program that will boggle a rocket scientist’s mind. But to have a call that is simple for a universal audience is a significant accomplishment.

This apparent simplicity for the caller of ConServIT’s NST results in as high or higher a yield (the percentage of complete records to the number of calls answered), greater accuracy, and lower cost than any other alternative. Calls are much briefer in duration as well.

How does NST Make a Call Seem So Simple?

Explaining in these few pages why something that is inherently complex appears so simple and elegant is difficult and perhaps impossible. With that caveat, hopefully, my best efforts will at least provide some insight.

1. Program Objectives

A particularly difficult aspect of an NST program is for the sponsoring organization to decide its objectives and adhere to them. An NST program is objective oriented, and typically objectives are interdependent. They can be changed from time to time. However, if the objectives become moving targets due to frequent changes, simplicity and the broader objectives become victims.

2. Script Layout and Call Progress

Call progress, the sequence of prompts, for an NST handled call is distinctly different from the call progress using any other automated technology. NST call progress generally follows what the caller would anticipate, not what the sponsor wants or how the technology can be most efficiently implemented. An NST program script is not interchangeable with scripts using other automated technologies. It is sufficiently different that porting an NST script to an ASR/IVR system won’t work.

3. Phrases, Wording, and Ambiguity

An NST script is caller centric, whereas, for alternative technologies, it is client and, more particularly, technology centric. The scripts for those systems are much more constrained, having to work around the requirements and limitations of their embedded word recognition technology. NST uses a fundamentally different technological approach.

Because an NST program is designed to be successful with the totality of the callers, the words used in the script must be universally understood. The expression “What is your destination?” will not work because many people don’t know what “destination” means. Rather the question is worded “Where are you going?” Third grade English is best.

In any automated script, there can be no tolerance for scripting ambiguity. If he caller responds, “What did you mean” the call could be lost. Ambiguity includes the perceptions of the caller, outside the specific words used in the script, based on the caller’s experiences and prejudices. One of the reasons that a perfected NST script can be the model for a live script is because all ambiguities and glitches have been recognized and resolved.

4. The NST Technological Difference

In the real world, there is no absolute certainty about the speech characteristics of any caller, the specific utterances that the caller will use, or the quality of the telephone connection. NST is inherently fault tolerant, unlike other technologies that require digital precision in recognizing a particular word or phrase. An agent, as does NST, has the flexibility to adjust for these shortcomings. Other voice automation systems, on the other hand, are inherently less flexible and forgiving.

The alternative systems compensate for their inflexibility by telling the caller how to respond, and by repeating back the answer as understood by the automated technology. Often their scripted prompts become painfully convoluted to compensate for the mismatch between the caller’s thought process and the logic process of the machine. The effort to match the two processes is one of the reasons an ASR/IVR program can be so expensive to implement.

Comparative Illustrations

I have selected a couple of familiar examples of prompts to highlight significant differences between the NST approach and the approach of alternative voice recognition technologies (IVR and ASR).

a. Frequently, imbedded in the greeting for voice mail systems is “After you have left your message, please hang up or push 1 for further options”. Do callers really need to be instructed to hang up after leaving a message? This prompt is very patronizing.

“Push 1 for further options” is placed very much outside of a rational position. These options regard a message after it has been left, such as adding to a message, erasing a message, hearing it again, etc. Logically, before a caller has even started a message, there is little relevance and the tag line is a distraction. It is after the message that the option alternative should be offered. So why aren’t these options expressed after a message has been completed?

The Press 1 option is used by only a very small percent of the callers (reportedly less than 5%). Yet, the additional five seconds to repeat this tag line, seemingly a lifetime to the typical impatient caller, has to be listened to by every caller on every call. It is an unnecessary, unappreciated and annoying imposition.

With NST, callers would only receive the prompt after messages had been left, making the call flow easier and simpler for everyone.

b. The unnecessarily complex prompt “If you want such and such, push or say 1” requires callers make two and possibly three independent decisions in response. One decision is whether to select the particular issue represented by the “1”. That, unto itself, is generally a difficult decision (The caller might be considering “Is that the best choice or should I wait to hear the next option”). The second decision is whether to respond using touch tone or word recognition. A third decision might well be “Why are they giving me this choice? Is there something wrong with one of these options that could lead me into automation hell?”

The prompt could be “If you want such and such, say one”. If the caller should push “1”, even though not specifically suggested as an option, the touch tone signal could still be recognized. Or the prompt could just say, “Are you calling about such and such?”

Telling callers to respond to confusing, complex, compound questions is definitionally the antithesis of simplicity. With NST, multiple decisions are seldom required from a single prompt, and when they are, it is only when the prompt avoids any ambiguity. The decision must be easy for the caller. An example of an NST multiple decision prompt with a single easy decision could be “What is your relationship to the person you are calling about?” The first decision in this example would be whether the caller is calling for himself/herself, and the second is, if not, what is the relationship.

NST - Packaged with a Bright Red Ribbon

The call progress and scripting are inextricably linked to the call handling software and technology of the equipment employed. The linkage may be more interrelated than even a computer program is with its operating system - a Mac OS program is only for the Mac and Windows XP is only for the PC.

ConServIT is packaged to offer a complete service – one stop shopping. It has the experience in designing and scripting programs to match its client’s objectives, using technology about which it is the unchallenged expert. That bright red ribbon is why ConServIT can guarantee superior results, an offer no one else makes.

Call me at 1-877-503-7131 or email me at theis@conservit.com if you have questions or would like to discuss this further.

Sincerely,

ConServIT
a service of Conversational Voice Technologies Corporation

Peter F. Theis
President

To Obtain Better Results, Don’t Annoy Callers

Use the Open-Ended Question

About Better Results

When scripting an automated call handling system, the ability to include open-ended questions is a powerful tool. “What did you think of our product” is an open-ended question. The caller’s open-ended response can be whatever comes to mind – in free form. In contrast, with a closed-ended question, caller’s closed-ended responses are limited by the question; “On a scale of one to five, how would you rate our product?” ConServIT’s natural speech technology (NST), with open-ended questions and free form open-ended responses, is a significant technological leap in automated call handling.

Test the importance of the concept by conversing with someone, using closed-ended questions. Then, to really make yourself odious, follow each response by asking, “Did you say … ? Answer ‘YES’ or ‘NO.” as IVR and word recognition machines are often programmed to do today. Absurd? Sure! People rebel at being treated as if on a witness stand. In ordinary conversation, the norm is the open-ended question and the open-ended answer.

ConServIT’s natural speech technology (NST) converses with a caller without presupposing the actual words in the caller’s response, and then responds appropriately, just as an agent in a live call center would do. This approach is more caller centric because the call is easier and moves along.

When clients migrate from other automated call handling systems to ConServIT, the number of successfully completed calls generally increases by 50% to 300% (Yes, those figures are correct.). This ability to handle open-ended responses is a principal reason. The quality of the answers received is also significantly better.

Branching Based On The Response

ConServIT’s NST has the ability to branch the program script based on the caller’s open-ended response. The NST greeting prompt might be “Are you calling to place and order, or how can I help you?” just as an agent would greet a caller. If the caller wanted to order, the system would branch to an order script. If calling for another reason, the caller would branch to another script, or be transferred to an agent, whatever our client directed us to do.

Branching within the call, rather than at the beginning, presents special challenges. A caller, for example, might be calling to request information about insurance and, well into the call, there would be a branch in the script dependant on the whether the caller’s interest regarded life or automobile insurance.

As each branch continued, the program script and the information provided by the caller in response could differ widely. The information after the branch needs to be integrated with any information the caller left before the branch. The data files generated under the entirety of each script would contain different fields. The life insurance and auto insurance response files might ultimately wind up in separate computer databases in different locations. Some program scripts incorporate multiple branches and those can be very complex.

The complexity of the metrics and the testing required with branching using open-ended responses becomes increasingly complex and technologically challenging.

Branching is more involved than simply changing the conversation path the caller follows – which is generally all our client sees. Most of the real work is in the back office.

Processing Open Ended Responses

Generic automated systems using closed ended prompts limit the choice of permissible responses, generally suggesting appropriate responses as an integral part of the prompt. However, with NST, the range of selections could be nearly infinite; too much to be included within the prompt (which is one reason NST prompts are more compact). For example, when asking the caller where he/she heard about the number being called, the responses suggested by other systems might be radio, TV, or the Internet. With ConServIT’s NST, there need be no suggestions. The caller responses could include answers the sponsor might not ever have thought of or that were too complex and numerous to include in the prompt, such as “I heard about it from a friend” or that the sponsor was not aware of, such as “I heard it on Good Morning America.”

To facilitate computer sorting and selection, NST converts the free-form responses into established categories according to preset parameters. A long verbal answer might wind up simply as a two-digit code. Generally, one of the set-up categories available would be “Other” for answers that fit no other category. Should a particular group of calls become a significant subset of the callers in the “Other” category, an additional new category can be established on the fly. This frequently happens where the sponsor may not be aware that its product or service was mentioned on a broadcast or linked to a popular website.

In another example, for people calling for information about one of a range of products being offered, such as mutual funds, the open-ended question might be, “Which fund are you calling about?” Various callers might reference a particular fund by a variety of names other than its given name. A precious metals fund might be labeled “gold fund”, “precious metals fund”, “metals fund”, and so forth. Call records for that single product would all be coded the same regardless of the specific words used by the caller to describe the product.

The range of permissible responses with NST is very large, and yet the accuracy cannot be diminished, a feature of ConServIT’s NST. The opportunity to recognize new situations not previously considered by the client is significant, considering that the audience and situations vary throughout the life of a program.

For survey work, ConServIT’s NST excels because of its open-ended questioning and categorization capabilities. This is a tremendous advancement over alternative voice or touch-tone systems, and Internet surveys where open-ended responses are discouraged or not accepted.

Consider again the question “How did you like our product”. The open-ended question gives the caller latitude to say what was good or bad, on the caller’s own terms rather than limiting the choice to the numbers one to five dictated by the machine, as the other automated alternatives would do. It is with that open-ended question that you can find out what is really on your customer’s mind.

I am regularly asked, “What happens if the caller says something the machine is not programmed to handle?” Because ConServIT’s NST is based on a statistical model, ConServIT’s systems can be programmed to handle just about anything. If the caller’s response is very unique, one out of thousands, it will be handled as it would be handled with a live center. What happens to that response, or the call, is a decision ConServIT’s client makes, not ConServIT. Incidentally, several of our clients use this capability to separate legitimate callers from people not calling for valid reasons, including pranksters and kids playing games.

The Benefits of the Open Ended Answer

Earlier, I commented that the ability to ask open-ended questions and receive open-ended answers was very important.

· The call is easier, and more comfortable for the caller. In short, a better experience.

· The yield (percentage of callers that successfully complete their calls) is dramatically higher.

· The cost per completed call is less.

What more could you ask for? Call me at 1-877-503-7131 or email me at theis@conservit.com if you have questions or would like to discuss this further.

Sincerely,

ConServIT
a service of Conversational Voice Technologies Corporation

Peter F. Theis
President

Making It Easier for the Caller = Better Results for You

Setting Up The Application

Setting up a natural speech technology (NST) application involves skills in human engineering, a unique computer technology and a specialized field of metrics. Most importantly, it is an art coupled with hands-on experience.

At the ConServIT call center we have over thirty years experience implementing natural speech applications - experience no one else in the world can match. This experience is crucial to implementing an application faster, more successfully, and at lower cost.

An NST program, once successfully developed, is so good it can be used as the pattern script for agents at a live facility. For those interested in thinking outside the box, performance metrics for NST systems employing that script are sufficiently consistent and similar to those of a live call center that they can even be used as a performance standard for the live center.

Scripting

The natural speech technology script is derived from what a call center agent says handling similar calls. Natural speech scripts, not being variants of scripts for other automated speech technologies, can neither be used by IVR and word recognition systems nor imported from them. Migrations between other automated speech alternatives and either NST or live are simply not doable, although many have tried.

Natural speech scripts emulate how the best agent would handle a call, after having handled fifty other calls for the same client on the same day. We know that an agent at a live call center, as the day goes by, will deviate from what is appearing on the agent prompt screen, using, instead, expressions the agent is more comfortable saying.

This is not simply the personal preference of the agent, as some would suggest. Rather, it is that the clear meaning of the words dictated by a written or screen script may not be what is being communicated to the other party.

For example, we all know what is intended by the words “Your call is important to us”. When people hear that message on the phone, the communication instead is a negative “Your call means little to us” and manifests insincerity. The communication is the opposite of the clear meaning of the words (which is why call center agents do not say it).

Each prompt, whether from a machine or live, “tells” callers something. “Communications”, in contrast, is what is communicated to the caller by the prompt. The two expressions, “tell” and “communicate”, are not synonymous. What is communicated subsumes the actual words spoken. This distinction is often very subtle. The agent’s spontaneous revisions of the script reflect a subconscious communication to correct the inadequacies of the scripted words.

ConServIT’s natural speech technology (NST) focuses only on what is communicated, whereas other voice technologies focus principally on what is told to callers. That difference is pervasive and overriding in the implementation of an NST script.

Failing to recognize this significant difference between NST and other automated voice systems, the client will often dictate all or a portion of the script, making changes in the verbiage under the fallacious assumption that callers share the client’s preferences or understandings of what they are being told.

For example, ConServIT might propose, as part of a draft script:

And the ZIP Code?

The client might demand that the clause be changed as shown below, believing the script revision would be more polite, friendlier for the caller and clearer.

May I have the ZIP code please?

Or worse,

Please give me (or Please state) your ZIP code.

Because of the negative communication to the caller associated with the client permutation, the likelihood of hang-ups, errors and caller frustration increases. This differential can be cumulative over several prompts. Differentiating what is communicated as a concept distinctly separate from what is told is a critical element for a successful NST program.

When a client gives us no choice but to accept their script revisions, we have lost the ability to maximize the caller service level as a primary objective, and minimize expense as a secondary goal. The client focus has changed from reaching a high level measurable objective, to a non-descript subjective standard that undermines both objectives. The application implementation has become client centric, rather than caller centric.

The seriousness of this dichotomy is aggravated by the high “yield” expected from an NST program (the “yield” is the percentage of the calls received that are successfully completed). When the yield is already very high, it is difficult to increase it further, and easy to cause the yield to plummet. A good NST program at ConServIT’s call center will experience a 90+% yield (higher than a live center). What may seem to a client as a token script change might increase the yield a couple percent for the reason given by the client (such as clarity), but then reduce the yield a much larger amount for another reason (such as being patronizing), for a significant net loss. There is almost always a trade off.

The Voice Recording

ConServIT’s voice programs generally use a generic voice. At live call centers, the caller could be served by the next-door neighbor, your child’s college roommate, or someone in India about whom the caller knows nothing. ConServIT’s voices could be the telephone receptionist in the office next door. ConServIT’s selection would be based on the voice that produces the best results. Those are the people the real world expects to answer their calls.

Not suitable for NST applications, where results are paramount, are the “persona” voices, the “professional” voices used for the client’s other IVR applications, or voices that the client’s advertising agency promotes. The objectives of the NST voice and the IVR voice are divergent, the former being caller centric and results oriented, and the latter being client centric and image oriented.

Metrics, Metrics and Metrics

After a new program has been implemented, ConServIT measures how callers respond. Changes and corrections are made to the program based on test metrics. Testing is an arduous, time-consuming high-level chore. All the confidence and the best of technology are worthless if the whiz-bang does not work with real world, bona fide callers - the target audience. That audience for NST is the real world caller, not the client staff.

This simple formula is far from being obvious. It is the opposite of the generic IVR/word recognition approach, which generally employs small test panels to opine on improvements to be made. But the problem with this approach is that everyone, the panelists, the people that are designing the program and the technicians and client staff are outside the statistical body of people that will be using the system. The two sets are mutually exclusive!

A real world caller to a call center has no preparation, has no preconditioning, and only wants a service or problem resolved. That caller is not calling to find out how progressive the company is, to hear about expanded services (the caller’s ears are turned off), has minimal preparation for the call, if any, and is not calling to please anyone or be on the next panel. This real world is what ConServIT’s natural speech technology is all about and what it deals with on a daily basis in its call center.

Perhaps the characteristic that particularly distinguishes ConServIT’s natural speech technology from generic voice systems (and live call centers as well, incidentally) is that NST programs are designed to meet measurable objectives. To accomplish these objectives, NST metrics focus on the unsuccessful calls, not the successful ones, to make rational program adjustments to measurably improve the yield.

To know details about the callers that successfully completed their calls is largely irrelevant for NST. NST needs to know about the calls that were NOT successfully completed.

Obtaining those metrics is a science in itself. The metrics cannot be obtained through panel research. They cannot be developed by calling a sampling of callers whose calls were incomplete or transferring to a live agent a caller that has just hung up for some unknown reason.

When putting an NST program on line, ConServIT asks its clients to identify each of their “test” calls to our systems so they can be distinguished in our statistics from real world calls. Although a few unidentified test calls are not problematic if they are joined by thousands of real world calls. They are a significant statistical hurdle if there are only a couple dozen calls coming from the real world, as often happens when a program is just beginning to roll out, and when initial program changes are being made.

One client, just before a program rolled out, made several unidentified test calls (many of which we recognized and isolated) and then even accused us of trying to manipulate their results by our insistence that they identify their “test” calls. In fact, the client was unintentionally manipulating our statistics by not identifying their test calls.

Preparation of the Call Information for the Client

To reduce the expense and investment for our client, ConServIT formats all the caller information, in virtually any format and field order, and transmits it as ASCII digital text ready to be merged into its client’s database upon receipt. It is that easy and inexpensive for the client.

What This All Means for ConServIT’s Clients

It’s Incredible. It’s proven. It works!

I appreciate your comments and thoughts, and encourage you to call or email me at theis@conservit.com or 1-877-503-7131. Best wishes for the coming year.

Sincerely,

Peter F. Theis
President