State Of The Speech Analytics Industry
Analytics, specifically speech analytics, has generated a great deal of excitement for over 10 years. The industry’s enthusiasm lies in the potential value that can be garnered from the countless customer conversations occurring each day. However, for so many, speech analytics hasn’t met expectations or never makes it past the “wish list” stage.
This article provides a state-of-the-industry update and shares some views on how to address the challenges that may be holding you back. As we enter 2015, solutions are more accessible and value more achievable. If you haven’t considered speech analytics in a while, or haven’t gotten value from what you bought, it’s probably time for a fresh look.
Challenges
Strategic Contact and Contact Center Pipeline recently conducted a survey on analytics use and vision (see “2014 Contact Center Analytics Survey Final Report,” see link below). In reviewing the survey results, Lori Bocklund states, “Of participants who have not yet implemented analytics, two-thirds plan on pursuing it. However, only about 15% of those have specific plans; the rest can only hope at this point. Perhaps this reflects some understanding that analytics can be valuable, but requires a large, multidimensional commitment to people, processes and technology, along with a thoughtful business case to back it up.”
Speech analytics is not simple. The solution can be expensive considering hardware, software, implementation services and staff to use it. You need buy-in from multiple departments to achieve the full value potential and buy-in from IT for implementation and support. You need a trained, dedicated analyst, not just an available reports person. And, you need time and focus to reap the benefits; it doesn’t analyze and draw conclusions on its own. That’s a tall order, even before you start pursuing the changes suggested by the analysis.
Value Potential Keeps Us Interested
To gain enterprisewide support, you need a value message that is achievable, high-level sponsorship, and collaboration among a variety of stakeholders, which can be difficult in many organizations. In a quick sampling of clients who were early adopters of speech analytics and others who are in hot pursuit, we found a common theme of struggle to achieve the outcomes they sought when they embarked on their analytics journey. However, none had “given up” and whether it be technology or resource hurdles, or the need to revise plans, they are tackling the challenges and forging ahead.
The good news is that, even if you just focus on the contact center, there is large potential for impact in compliance, customer satisfaction, customer retention, cost management and sales effectiveness. If you can get crossfunctional ownership/involvement (from marketing at a minimum), there is even greater potential beyond contact center processes. Small changes to marketing messages, scripts or processes can yield tangible results. Many departments can benefit from improvements in the upstream and downstream impacts to end-to-end business processes, customer experience, market intelligence, and risk management or compliance.
Solution Enhancements Add to Value
Most vendors are adding new capabilities that make speech analytics more enticing. Limitations in language capability have been a challenge in the past, but vendors are expanding their language libraries. Some can analyze close to 50 languages. Accuracy and analysis speed have always been a challenge with tradeoffs depending on whether your solution used Large Vocabulary Conversational Speech Recognition (LVCSR) or Phonetic indexing. And that speed and accuracy came at a high cost in processor demands (and therefore cost). Vendors are now combining the approaches with more sophisticated algorithms and adding direct phrase recognition to maximize both speed and accuracy, and they are able to do so more efficiently. Solution usability has also been enhanced with improved tools for identifying appropriate word or phrase searches, finding the pertinent calls to listen to and drawing correct conclusions from the calls. Sentiment or emotion detection has become a common component of many solutions. Vendors approach this functionality differently, focusing on tone (e.g., volume, pauses, cross talk, laughter, pitch) or searches for specific words or phrases that flag concerns or opportunities during the call.
Most solutions offer an improved ability to reveal insights that drive action. Metrics have been expanded and displayed in more userfriendly reporting. Key performance indicators (KPIs) can be defined to evaluate occurrences and trends. One of the most market-changing enhancements has been the emergence of realtime components. Real-time analysis enables a reaction during calls based on insights gained from root-cause analysis performed on past calls. For example, you can explore why customers get frustrated and leave your company. Once you determine the root cause, you can use the system to watch for the situation to occur and notify a supervisor, who may IM the agent. Alternatively, using business rules, the system can notify the agent during the call to offer a particular discount or promotion to defuse the situation and retain the customer.
Besides the improved functionality and reporting, there are new options for how speech analytics is delivered. Some vendors offer managed services, providing analysis through professional services. This offering reacts to the challenge many companies have in dedicating qualified and trained resources. Managed services can address short-term needs or focus on targeted improvements to a perceived problem (e.g., IVR, FCR or the customer experience). Managed services can be appealing to small-to- medium businesses as it is more common for large companies to have “big data teams” with analysts. In addition, vendors offer multiple pricing options (e.g., cost per interaction, cost per agent). And like most technology, speech analytics can reside in the “cloud,” enabling buyers to avoid capital expenses and reduce the need for internal IT involvement.
Perhaps most importantly, vendors offer options that can help reduce the complexity and time to generate valuable results. Many vendors offer expanded professional services that can support your implementation process through building libraries and queries. Figure 1 shows the options for technology delivery and resources to drive value from that technology.
Vertical solutions are becoming common with specific libraries and search algorithms already built for financial, healthcare, insurance, government, consumer goods and retail, transportation and logistics, manufacturing, and help desk, among others. Vertical solutions reduce the time and cost of implementation as well as the required skills.
Compliance deserves special mention and applies to several vertical markets. Compliance solutions can apply to collections environments with scripted messages, claims or other areas requiring “voice signatures,” and sales environments (especially cable/wireless/TV) where services are switched frequently and require validation. Speech analytics (including real time) can identify calls where the right compliance phrases are not heard and bring those calls to the attention of management for appropriate reaction.
Vendor Changes—Where to Look
Most analysts agree that, even though the solutions have evolved, speech analytics is not a fully mature market (see “Market Maturity” sidebar). Here’s what you need to know today.
Most vendors use value messages that speech analytics will:
- Reveal insights on reducing costs
- Improve products, processes and the customer experience
- Identify competitive threats and opportunities
- Reveal the root cause of customer dissatisfaction and call volumes
- Identify and validate “moments of truth” in the customer journey
- Help you understand drivers of customer loyalty
As a result, you are unlikely to differentiate vendors based on their value messages. You will have to dig deeper to narrow your search to those that feel right for you.
The original pioneers in speech analytics were Utopy, CallMiner and Nexidia. CallMiner and Nexidia are still operating as standalone companies focused on this niche. Utopy has been bought by Genesys, bringing speech analytics into a much broader contact center technology suite. Most, if not all, performance suite vendors offer a speech analytics solution which expands your sourcing options. You can buy their full performance optimization suite or just add speech analytics. Typically, the performance suite vendors include options for data, text, speech and desktop analytics.
Figuring Out the Fit
Speech analytics has come a long way, and its potential has only grown. Decide if the time is right for you to pursue speech analytics and define how you can achieve value. Focus on the specific uses that make sense for your type of business, and for the business goals you are pursuing (not just the canned list of market hype messages that all the vendors use).
Assess whether your solution will be contained within the contact center or if you will have enterprise involvement. Discuss the level of involvement you can expect—and can get—from IT. Based on their interest in supporting your solution, you may target solutions that can limit IT involvement, such as cloud offerings. Determine who is going to pay and get resource commitments.
Meet with a crossfunctional team to determine your organization’s willingness to invest and dedicate resources. Discuss how the results will be used to take action and make changes. Then you can determine the services you will require (e.g., implementation, application support, managed service) and the vendors that will be most appropriate for you.
– Reprinted with permission from Contact Center Pipeline, www.contactcenterpipeline.com
You must be logged in to post a comment.