Speech analytics moved from the government sector to the commercial market in 2004 and quickly captured the attention of both contact center and enterprise leaders. The technology was compelling because it did something new and wasn’t viewed as just a replacement for existing systems or applications.
Instead, the technology was perceived as a high-value analytics solution that provided customer and operational insights not available from any other application. Speech analytics rapidly grabbed the interest of company leaders, who appreciated its ability to provide information about customer needs and wants—the Holy Grail for marketing and customer service organizations alike.
Although it found its initial home in contact centers, where the compounded annual growth rate of speech analytics implementations was 153 percent between 2004 and 2009, the technology has been deployed by many other departments, including marketing, sales, operations, product development, credit, and collections, where customer insights can yield a cost and/or strategic advantage.
What is Speech Analytics?
Speech analytics, also known as audio mining, is software that converts unstructured conversations, such as phone calls, into structured output. The conversations are structured using a variety of techniques and turned into metadata. Then, the output files can be analyzed and used by the enterprise. Most speech analytics solutions use recorded conversations for their input, but progress has been made recently in real-time analysis. Real-time speech or interaction analytics applications that address both verbal and written interactions (e.g., Web chat, email, short message service, and tweets) are expected to play an increasingly important role in contact centers.
The primary uses of speech analytics today are:
1. Root cause analysis: understanding the reasons that people call.
2. Trend analysis: identifying volume trends of both anticipated and unexpected reasons for customer calls.
3. Emotion detection: understanding callers’ emotional state (some solutions can detect emotional changes and heightened states).
4. Talk analysis: understanding caller and agent talk periods, including measuring silences, holds, transfers, and talking over.
5. Script adherence: monitoring how well agents follow their scripts, communicate required information, or utter inappropriate phrases or words.
6. Quality assurance: identifying conversations that require management’s attention.
Benefits of Speech Analytics
When used properly, the technology delivers rapid and quantifiable benefits throughout companies, categorized as follows:
• Operational: improvements in the operating performance of a department.
• Productivity: ways to make an organization more productive, which would reduce operating costs.
• Efficiency: enhancing employees’ job performance.
• Quality: improving the products and services delivered by an organization.
• Customer satisfaction: improving how products and services are perceived by customers as well as other constituents.
• Cost savings: reducing costs by finding better ways to do things.
• Cost avoidance: finding ways to avoid additional expenses.
• Reduced customer attrition: identifying and reducing the number of at-risk customers.
• Increased revenue: using speech analytics to increase revenue and profit margins.
• Branding: improving how a company is perceived.
• Risk and liability avoidance: reducing an organization’s exposure to risks.
The average payback from speech analytics applications is three to nine months, but results vary greatly among organizations. Companies that have put in place processes that enable them to apply the findings from speech analytics on a timely basis are realizing great returns. Unfortunately, a large percentage of users are struggling to apply the findings and insights uncovered by speech analytics and, consequently, are disappointed with their solutions.
Making Speech Analytics Actionable
The number-one complaint among users of speech analytics solutions is that they find it difficult to apply the identified insights and cannot make the solutions actionable enough to pay for themselves. There is a clear disconnect between the findings reflected in reports and dashboards and using them to effect the necessary changes to fix the underlying issues.
In other words, the current generation of speech analytics solutions does not have a built-in change management process that facilitates the practical application of findings in contact centers and across the enterprise.
Speech analytics has hit a critical juncture. Few technologies break into the market without hiccups, and speech analytics is no exception. The overarching question is: What will competitors do to help their customers overcome these serious challenges and to realize benefits? This is not a simple issue. All that vendors really want is to sell more software, because margins on software are much greater than those for professional services, workshops, or training. On the other hand, vendors need to help develop and share speech analytics best practices in the user community. If they don’t improve the success rates of implementations, market interest and the adoption rate will surely wane.
The speech analytics vendors hear this demand from their customers daily and are working hard to meet the challenge. They are delivering more packaged offerings that come with results-focused libraries, reports, and alerts. They are providing tighter integration with other workforce optimization modules, such as quality assurance and coaching, as well as ways to feed other enterprise systems beyond the contact center. Finally, they are developing and delivering new workshops, training, and professional services to help users derive the full benefit of their speech analytics solutions.
Even with the current challenges, the speech analytics market should grow rapidly over the next four years. DMG Consulting projects growth of 40 percent, 42 percent, 32 percent, and 25 percent a year between 2010 and 2013, respectively. Those rates are expected to slow in later years, because DMG Consulting believes the market will have grown quite large by then. In early years, when adoption is small, it’s relatively easy to grow by 50 percent to 100 percent, but as the installed base becomes larger, it’s much more difficult to sustain that rate, even if the number of seats sold continues to rise sharply.
Entering its second technology generation, speech analytics is the most mature of the analytical applications used in contact centers. It is helping enterprise leaders appreciate the value that contact centers bring to the entire organization. Better harnessing the information that arrives continually in contact centers, both internally and for the broader enterprise, remains a challenge for all but a few highly progressive companies. These companies have successfully prioritized enterprise revenue and cost goals over inertia and organizational politics.
Social media is accelerating the need for multichannel analytics to address all forms of spoken and written communication. Calls, emails, text messages, blogs, Twitter, Facebook, and other forms of social media are increasing the visibility of speech analytics beyond the boundaries of contact centers. To take advantage of those opportunities, speech analytics vendors must build multichannel solutions and prove to the market that their applications can be used to drive the necessary changes to improve the enterprise’s bottom line.
Donna Fluss (email@example.com) is founder and president of DMG Consulting, a leading provider of contact center and analytics research, market analysis, and consulting.