JUSTIFICATION – WHY IVR?
IVR is one of the oldest and most widely adopted customer contact automation technologies. Why, in an economic downturn, is it more important than ever? The reasons are, at first glance, rather simple:
- IVR automates customer interactions, and therefore it cuts costs;
- IVR smoothes the peaks and troughs of call queues in call centers, thereby adding a degree of predictability to staffing and call levels;
- Speech recognition technologies are set to ratchet up the usefulness and customer-friendliness of IVR;
- If there is one area of CRM which will maintain its momentum in an increasingly tight economic climate, it is (customer, employee, supplier, distributor) self-service. Speech-enabled IVR systems can give to the voice channel which web self-service has given to the e-channels.
The assumption among the general public is that IVR is as unpopular with customers as it is popular with call centers and companies. Badly designed corporate voice self- service applications, with their endless choices that lead nowhere, are mainly to blame for this negative perception. Examples of badly designed IVR systems abound. One particularly poor one automated booking system does not let users know the dates for which tickets are available. It instead forces users to make guesses until they happen to find a suitable time. The damage such IVR systems cause can be great. This particular system has become so unpopular that customers have started to talk about it with their friends and family. The obvious consequence of this is that potential customers were warned against using this particular service and are likely to attempt to contact the business in ways that prove more expensive to operate.
Generally speaking, many end users feel that IVR systems do not provide an adequate level of interaction and ‘dictate’ rather than ‘interact’ with the customer. In fact, until very recently, the term IVR was itself something of a misnomer: no IVR was truly providing an interactive self-service experience. Despite the possibilities touch- tone offers, the type of applications vendors can develop are quite limited, due mainly to the restrictions of telephone keypads. The direct result of these limitations is that, in some cases, instead of relieving call centers and switchboards from some of the incoming calls, IVR does the exact opposite. Callers simply bypass the IVR and go straight to the operator.
As a result, IVR has not brought as much benefit as it should have. In a survey of over 300 call center managers in 2000, Datamonitor found that on average 35% of calls to a call center are front-ended by IVR, but that only 12% of calls are handled completely by automated systems. In other words, most call centers are using IVR as an added routing engine, sometimes with limited success.
Despite these problems, the IVR market is large. With the help of new technologies, IVR vendors and their customers can shake off the current image if IVR, and start to improve customer service at the same time as reducing costs. There are two main initiatives that are stimulating the IVR market:
- Using speech recognition technologies to design user-friendly, more sophisticated and more diversified applications
- Migrating to open standards to enable developers to write applications faster and more easily.
Outline of the brief
IVR (Interactive Voice Response) is a technology that enables callers to press buttons on a touch-tone phone to choose between options presented by an electronic voice.
In most cases, IVR serves one of two purposes:
- Automation of customer services: IVR systems allow customers to service themselves by accessing a company’s database through the use of a touch-tone keypad on a telephone;
- Routing of calls to the appropriate agent: IVR systems allow a call center to obtain information about the caller in order to route the call to the appropriate agent.
The vast majority of IVR systems use Dual-Tone Multi-Frequency (DTMF), the audio signal that is generated when buttons are pressed on a touch-tone telephone phone, to determine the user’s response to recorded prompts.
IVR is the most widely used call center technology worldwide after the ACD. The market for traditional IVR in Europe and the US is already mature and is growing slowly – with much of this growth coming from existing customers. In the rest of the world, technological and cultural barriers have prevented the market from growing as quickly as it may otherwise have done. However, despite relatively slow growth in comparison to other areas of CRM, the IVR market is not static. IVR is as essential as ever to most medium-sized and large enterprises because it allows substantial cost savings, and it is therefore a priority investment area for a significant number of companies even in difficult economic times. After all, cutting costs is the order of the day in many companies.
Moreover, IVR technology is also evolving speedily. Voice self-service is no longer exclusively seen as an automation tool or a cost saving tool – it can now also be considered a means of improving customer service. The technological advances that enable these improvements entail a radical change in the way IVR systems will be built and implemented. This brief will examine them.
In the course of the analysis for the future of the IVR, Datamonitor identified four key Action Points:
- IVR vendors must embrace speech recognition;
- IVR vendors must understand the telephony world is changing into a more
web-like open environment and adjust accordingly;
- IVR vendors should focus on core skills of application development and back-
- IVR vendors must target the right customers.
IVR vendors must embrace speech recognition
Today, although most of the IVR systems developed by the leading vendors can be speech-enabled, only a fraction of all IVR ports are. This will change over the next decade, as touch-tone is gradually replaced by speech recognition. This is because in many cases speech recognition can enhance existing IVR by:
- Allowing the development of more sophisticated applications that can automate more complicated tasks. There are some interactions which are too complicated to be automated using touch-tone IVR. For example, asking for flight information using touch-tone is technically possible but would take a lot of time and patience. This is because the number of options would take too long to be prompted, and be too difficult to be selected from using a telephone keypad. Asking simple questions using voice, however, is quick and easy. Other examples include stock quotes, directory assistance and name and address collection. The range of interactions that can be automated will increase with time, largely as a result of improvements in natural language understanding technologies. And more automation means more cost savings.
- Allowing companies to offer value-added services that would otherwise be too costly (using agents), or too inconvenient (using touch-tone IVR). In addition to cost reduction, speech-enabled IVR can help improve customer service and generate revenues by allowing companies to provide value-added services cheaply. For example, in an effort to increase customer retention, some telcos offer their customers access to a personal address and telephone book through speech- enabled IVR. Customers using this service are less likely to switch operator since they are reluctant to re-enter the entire contents of their address book. Access to voice mail on mobile phones has also been made easier through the use of speech recognition by some network operators. Although these applications may not provide an ROI directly, they can help improve customer retention – and do it more affordable than would be possible with agents. This is crucial in sectors such as telecoms that are experiencing intense competition as a result of regulatory changes. These services can also help generate increased revenues, either from increased voice traffic, or from service charges.
- Providing a more human interface that can help to reduce the number of end- users that ‘zero out’ to an agent. In most instances people prefer talking on the phone to pressing buttons, since the interaction is more ‘natural’. This gives speech recognition an advantage over touch-tone IVR. Testament to this is the fact that, when asked, people who have used automated 411 systems in the US, do not realize that they have used speech recognition: the interaction was natural and did not stand out as one in which a new technology was used. Further, interactions can become more conversational and less menu driven. Since people prefer speech recognition to IVR they are less likely to ‘zero out’ to an agent and if a greater number of calls can be automated then greater cost savings can be made.
Obviously, the increased price of speech recognition over IVR means that there are trade offs, and there is a threshold in terms of call volume, and automation percentage, below which touch-tone IVR has a better business case. But on larger systems, the improved user interface that speech recognition allows can drive an ROI of its own due to this increased level of call automation. Further, in virtually all systems it can improve customers’ perception of service (although this is a much more difficult benefit to quantify).
IVR vendors must understand the telephony world is changing into a more open, web-like environment and adapt accordingly
The telephony world is changing rapidly. New technologies are allowing voice to be treated in similar ways data – transported on IP networks and manipulated using web-style standards. Forward-looking vendors are taking web paradigms and looking to abstract telephony applications from their enabling hardware and software.
Vendors must migrate to open standards
Today, the IVR market is broadly divided into two distinctive segments:
- Proprietary IVR hardware and software
- IVR software based on open standards and running on voice servers.
Proprietary IVR systems have been on the market for more than 15 years and account for the vast majority of the market. They benefit from a body of experienced developers, proven reliability, large channel number capability and existing integration work with back-end systems. However, these systems also lock customers in to one supplier. Their components are not to be easily interoperable with those provided by competitors or those provided by companies that are not close partners. Application development in these environments, while eased somewhat by graphical tools, is more complicated than the VoiceXML equivalent. Further, customers are increasingly looking for more flexible systems that can be easily modified and enhanced, perhaps even by in-house teams. This is driving a shift in interest, if not yet significantly in sales, to the more open approach. The heavy investments made in proprietary systems by the largest IVR customers, is restricting the pace of this change.
However, the future lies in interoperable components and this implies a greater focus on applications, for both customers and vendors. Traditional IVR vendors need to address this trend. While this may mean relinquishing some of their control of customers and shifting the primary source of their revenues, application development, integration and services are where future growth lies. The platforms that underpin these applications will become increasingly commoditized. This is not to say that these changes will happen overnight, and that there will no longer be money to be made from proprietary platforms, but IVR vendors must be prepared to shift emphasis.
VoiceXML: a first step
The lack of standardization within the industry remains a barrier to the disappearance of proprietary IVR hardware and software. However the first steps in this direction are already well underway. VoiceXML − a standard that describes voice interactions using XML as a base − is already changing the way IVR hardware and software are developed. Its advantages and drawbacks are summarized in Figure 1 and discussed below.
Figure 1: Advantages and drawbacks of VoiceXML
- Voice as a web channel: VoiceXML enables voice applications to be more easily integrated with other internet-based applications. Since it is a standard, it facilitates the running of voice applications on open servers instead of proprietary platforms. Its uptake means that it will be harder for IVR vendors to add significant value with new proprietary platforms.
- Ease the application development process: VoiceXML allows applications to be created in an environment familiar to anyone with any web development experience. It therefore opens up voice applications to a wider range of developers that do not necessarily have telephony experience. IVR vendors should maintain an advantage, however, since it is the human factors and dialog design which tend to determine the success of a voice application and it is in these areas that they have the most experience, and web developers have the least.
- A new breed of vendors is challenging the market leaders: The end of the proprietary era, as far as application development is concerned, means that new players can enter the market and compete with established IVR vendors. This is positive as the new generation of vendors help bring innovation, advance and competition to the market.
However, while VoiceXML is an exciting development, the hype surrounding it is not always justified. VoiceXML is not mature and has several important limitations.:
- Poor call control capabilities: VoiceXML’s poor call control capability is a serious barrier to its uptake. In the call center environment, call control is essential, particularly when used in combination with CTI applications such as screen pop and call line identification (CLI). However, the VoiceXML Forum is working to add this functionality to the standard – using SIP (session initiation protocol) as a framework.
- Functional limitations mean that it is often not used in a standard way: While VoiceXML is maturing (witness the recent release of version 2.0), it still has limitations and cannot move fast enough to accommodate all of the functionality that speech recognition engine vendors are developing. As a result, developers often design unique routines for more sophisticated applications, or use object tags to call routines specific to certain recognizers. This effectively means that most application developers are using a customized version of VoiceXML, making the end result incompatible with different voice platforms, recognition engines or applications. This means that while VoiceXML is a first step towards a truly open environment, and is much closer that proprietary IVR, it currently does not allow the desired level of interoperability.
- Requires replacing any existing IVR hardware and software: By far the majority of existing IVR systems have more or less proprietary platforms. While APIs are being developed to allow the use of VoiceXML for applications, many of the problems of closed platforms remain. Currently IVR vendors need to provide migration paths to VoiceXML, which can protect their customers’ existing investments while positioning them to be able to take advantage of VoiceXML as it matures. In the medium-term, IVR vendors need to be ready to truly open up their platforms and to focus on application development, back office integration and services for their growth.
Nevertheless, customer attitudes are changing with the release of each new version of VoiceXML. While the current lack of true standardization is an issue, the possibility of having interchangeable layers of software and hardware is likely to appeal to a large proportion of existing and new IVR customers. VoiceXML, despite its limitations, will have a long-term impact on the IVR market, but, the problems outlined are a serious impediment to its widespread uptake in the short term.
The ‘webification’ of voice allows voice applications to be delivered using the ASP model
Delivery of traditional hosted IVR services by telcos or IVR vendors themselves through a telephone network is no longer the only way to obtain hosted IVR services and to rent an IVR system. Voice applications can now be delivered using the ASP model through the Internet or through an IP VPN.
Theoretically this type of service should enable a wider range of companies to use IVR systems since it enables them to rent rather than buy. This is attractive because its:
- No risk. Companies with a tight IT budget and little experience in running new IT initiatives, or those who do not want to make a long term investment in voice solutions, like the flexibility of renting an IVR system rather than buying one..
- Cheap. Pooling infrastructure resources between companies provides economies of scale. It also means that customers do not necessarily need to pay for enough ports to cover peak load, only to see them sit unused at other times, since the burden can be shared. This can produce a cost savings that can be shared between host and customer.
- Flexible. Customers can be charged on a pay as you go basis. This flexibility enables providers of this type of service to target a wider range of companies who are reluctant to make capital investments.
However, these advantages rely on a critical mass of subscribers. Smaller ASPs will not be able to offer prices as good as their larger competitors – and while there are other reasons for taking the ASP route, transparent, predictable and most importantly low price points are the most striking advantage. This exerts an irresistible pressure pushing voice ASPs towards consolidation. Datamonitor expects that in the long term voice ASPs will consolidate around two or three leaders within each country.
So far the most prominent voice ASPs are the reformed standalone voice portals and the VoIP network providers that are looking to offer value-added voice services.
However, it is not clear that either of these kinds of companies will be the dominant voice ASPs in five years. They face competition from a number of sources, and perhaps the key will be to act as enablers of the offerings of some of these larger competitors. The obvious place to look for the future giants of voice hosting is the telcos. These companies have established networks, established billing systems, established brands and, most importantly, established customer bases.
Network-based solutions delivered by a carrier have a future with smaller companies as long as the providers of such services can demonstrate they can integrate data and voice and ensure that the system is secure. In the long-term, the roll-out of IP telephony networks, be they public or corporate, will make the delivery of such services easier.
IVR vendors should focus on core skills of application development and back-office integration
IVR vendors must shift their focus away from telephony platforms. Instead they should look to improve their application development skills and to find ways of packaging their existing work. They should also look to monetize their existing back office integration.
While the market as a whole is growing, one of the most notable trends over the next few years will be a shift in value from enabling hardware and software to applications and services. Enabling technologies, including both telephony platforms and voice servers, are becoming increasingly commoditized.
As a result, established vendors that rely on the sale of these proprietary platforms have to alter their strategy. Traditional IVR vendors will increasingly need to focus on the application development arm of their businesses. However, competition in this area could be intense, as developers from the web world start exploring the possibilities of voice and gain expertise in this area. Professional services, including ASP, consulting and systems integration, are also a high growth area that will attract IVR vendors. There is a lot of value in the work that vendors have done to integrate their IVR with disparate back office systems. At the moment this is locked up in proprietary code. Vendors should look at ways in which this value can be packaged and monetized.
The challenges new vendors are facing are different. In many cases, these vendors are already focusing on application development. However, these new vendors can have trouble competing with established vendors on back office integration. For these new vendors it is therefore preferable to partner to be able to offer comparable integration and service delivery options.
IVR vendors must target the right customers
Speech recognition is not suitable for everyone. Vendors should be aware of the following barriers to the adoption of speech recognition:
- Shrinking IT budgets;
- Unsuitability of speech recognition for some applications;
- Cheap labor costs in certain parts of the world.
Shrinking IT budgets.
Voice applications are expensive, since most require a high level of customization. Further, speech recognition engine licenses remain costly. In an economic downturn this means that even though speech recognition can have a good ROI, many potential customers are likely to postpone projects that require significant capital investment. There are three possible responses to this:
- In situations where speech recognition provides an ROI over and above touch-tone IVR, the message is clear: The sooner you invest, the sooner you will save money. By not investing companies are losing money unnecessarily. And the good thing is that ROIs for speech recognition can be proved, and there are numerous successful case studies already. ROIs for other technologies are frequently less clear-cut, if they exist at all.
- Outsourced IVR enables customers to account for their expenditure in different ways, and reduces the upfront costs of implementation. At the same time, it enables them to reduce some of their existing call center costs, potentially take advantage of better equipment (that economies of scale enable the suppliers to provide) and also to have the option to pull out, if necessary, without having to write off sunken costs.
- Creative financing can bring many of the pricing benefits of the outsourced and ASP models to customer premises equipment, effectively offering customers some of the benefits of both models. By leasing equipment, or by charging customers on a per port per minute, or a per port per month basis, vendors can offer their customers ROIs, while at the same time reducing the risk (for their customers, although not necessarily for themselves). There are also some adventurous companies that are generating revenues from their voice applications by taking a share of the cost savings that their customers receive – a very compelling pricing model to potential customers, and one that brings recurring revenues. However, such ‘shared risk’ revenue models have proven thorny in the past. Those going the shared-risk route will need to agree on a set of complex metrics beforehand – while there is a link between ‘saved costs’ and value for the end-user, there is not necessarily a link between end-user ‘saved costs’ and the real revenue needs of the equipment or applications vendor.