Contact Center Convictions

[#1]

Hi.  I’m Drew Judkins, VP of Market Strategy for inContact, Inc.  I have a fun job, I help figure out what we are going to do.  This however is frustrating to my wife who is occasionally asked what her husband does for a living.  She feels dumb saying something like ‘he is a figure-outer.’  It would be easier for her to say a teacher, lawyer or fireman… but figuring out what to do is what I do, and it is fun. 

One thing I’m trying to figure out lately is how to make speech analytics technology available to inContact customers.  Speech analytics is a fairly new technology to the contact center; it analyzes recorded conversations and generates reports about why customers are calling and how agents responded.  Of course the technology for speech-to-text translation has been around for a while. We have all seen movies about governments or spy agencies spotting words while reviewing transcripts of our calls. But speech analytics for the contact center is much more than that.  Sure you can pull up a transcript of the call, but the analytics and reporting around what was said, in what tone of voice, with how much silence, etc… is pretty interesting and makes speech analytics very compelling to me.

Speech analytics vendors provide intriguing reports on why customers are calling, such as calls by category or a breakdown of call disposition.  Apparently the analytics engine can often disposition more accurately than agents, a significant savings to average handle time.  The software engine can also report on first contact resolution ratio (FCR), product performance, and even the number of times and context of when your competitors name was mentioned on the call.  Usually you find yourself drilling all the way down to specific calls and clicking to listen to the portion of interest.

Your QA folks generally only have time to monitor three to ten calls per agent per month (less than one percent of the calls that agent handles.)  Speech analytics gives you full coverage and a rich set of search criteria for your QA people to leverage.  For example, they may search for all calls that were really long and then see the context of those calls, or list the calls with long silence, or list each call that mentioned a competitor or certain product, or that had an angry caller, etc.  They can find the “important” calls and review those.  Management can ask ‘how many calls this morning did people call in upset and mention late fees.’  Or even better, they don’t need to ask, they will simply see a new trend that points out that a lot of people are calling today (or this week/month) and saying they are unhappy about the late fees.  I think you get the idea.  I wish I had this technology back when I was running a call center. 

For those of you using our post-call (or email or chat) quality survey tools (Echo) imagine using a speech analytics engine to crunch through the open ended question comments you receive.  Wow!  Again, the transcript is nice to have (and could be posted back for interested parties to review) but the search and reporting ability… very cool.

However, less than 3% of contact centers have adopted speech analytics.  Why?  Well, other than being new, it is expensive to purchase and then complex to implement.  Speech analytics vendors have a convincing ROI argument, but it is hard to overcome, expensive and complex. 

So… how can inContact deliver speech analytics to our customers?  We have the ability to record conversations and we also know details about the contact (skill, agent, queue time, caller-ID, etc.)  We may even have information about the caller that was queried from the CRM in the IVR (e.g. account number/type/balance/status, etc. handy stuff to have available to filter in the reporting and searching criteria.)  My thought is that inContact could quickly be configured to do the following:  First record each call.  Second, at the end of the call, send the recordings and the associated data to the speech analytics engine.  Third, let the engine crunch through them.  Fourth, give you access to the reporting and search interfaces.  Bing, problem solved, right?  Well, although this is a big step forward, we still have the issue of expense and the complexities of customization to solve.

Regarding expense, what if there was a reasonable setup fee, with the software delivered via SaaS and functional for call centers with 20 to 800 seats?  This would cover most inContact customers, which is why I titled this rambling essay  Speech Analytics for the rest of us.  More on this subject next time, and I still need to solve the complexity issue.  But at this point I would ask for your feedback.  Should we figure this out?  Should this be a priority?  Could the price issue and complexity issue be solved?  Love to hear from you online here and you can even contact me directly.

By the way, if you would like to learn more about the technology, DMG Consulting publishes an excellent annual in depth report on Speech Analytics (http://www.dmgconsult.com/services/speech/abstract.asp).

These vendor websites also have a lot of information about the speech analytics technology: 

http://CallMiner.com

http://Nexidia.com

http://UTOPY.com

If you are interested in speech analytics in your contact center please contact me first and we can discuss.

Drew

[#2]

After going through demos for all major speech analytics players in the market I am truly convinced they can bring immediate savings and build efficiencies that could not be obtained through any other tool or technology. I think the key barrier right now in widespread adoption of speech analytics technology is the startup costs and large initial investment. SAAS is the way to go with an MRC model, especially in light of current economic struggles and weariness to spend money within companies. If a true "try before you buy" model and technology was in place that would certainly help with adoption. It would allow IT/Technology folks to put it in place for their recorded calls and demonstrate over the period of a month what issues they could find, how they could find them faster than conventional methods and how they could save the company money. This way you have a true ROI to show based on fact and experience vs. a theoretical ROI that sounds great on paper but is harder to prove out. Speech Analytics should be adopted by all call centers now more than ever to help quickly find efficiencies in a more automated and less labor intensive way and help save the company money through improved customer retention and increased satisfaction (those things lead to real dollars, even though most execs have a hard time with that concept).

[#3]

As everything, a little research on what the majority of use for the technology will help. Call centers are intended for a better customer experience ( or at least that is what I think!) To be able to find angry customers and have a second call to make things right may bring an immediate ROI. Can we look at the words being used or the tome of the voice to detect or in fact which is less expensive to be accomplished would be the place I would start with.

The number of seats and what is the purpose of the call center of the company will determine if inContact needs speech analytics for the current customers. If the cost is an issue, training the agents to tag the calls right may be a preferable option.

[#4]

Speech Analytics is intended to be an enabling tehcnology for contact centers, just like speech recognition-driven self-service and others.  If you operate a contact center I highly recommend you understand your needs deeply before you become enthralled with the automation hype vendors spin.  Pick one area to focus on and test out your options on a small scale.  Before the project begins have a list of "must-have's" and "wanna have's" from the technology.  Also, be sure you understand what it will take to support an implementation.  Keep in mind, your application will need to change just as your call center business changes- sometimes (more often than not) unexpectedly.  How will you be able to deal with this? 

We could probably learn quite a bit from the history of another "enabling" technology- speech recognition.  Let's face it, there are more lousy speech IVR apps in place today than good ones- which is why www.gethuman.com and other sources of how to get around them- even exist.  There's a list of reasons, and excuses, why Speech Recognition, has such a black eye but at the end of the day it's up to us to know our business and know the best place to start with an enabling technology.  Take your technology vendor's recommendations and review them against the context of YOUR business needs. 

It's still relatively early on in the adoption of Speech Analytics.  The technology and business case for implementing is compelling but it's now up to the users to determine the next step and really test out the hype. 

[#5]

Drew ... As VP of Marketing for the leader in enterprise speech analytics, I feel like I owe you a promotional fee ... sorry, due to budget cuts, that line item has been eliminated!  All kidding aside, you articulate a clear understanding of the benefits of speech analytics.

If after reading your analysis of speech analytics, your readers still require addltional information on how speech analytics works and what benefits they can expect, I would encourage them to watch our newly released video available at http://www.callminer.com/speech-analytics-business-value.htm.

There is NO doubt that CAPEX budgets are challenged today; some of our mid-sized enterprises are beginning to inquire about CallMiner Eureka SaaS.  I would be interested in your take and your bloggers take on what that means ... what kind of monthly fees (hosted or/and nonhosted) will these enterprises be able to expense .... blog me back (BMB) with your response.

Speech analytics does automate call disposition through automatic call categoriz(s)ation (a nod to our international customers!). Not only is it more accurate and more complete (there are hundreds of data points for each call -- call reason, behavio(u)r, product or competitor mentions, outcomes, procedure, acoustics and call characteristics (aka meta data) -- but organizations can save $000's or the British pound equivalent with by using speech analytics for call dispositioning.  

CallMiner plans to release a white paper this Thursday, Feb. 25, on how speech analytics can significantly reduce Wrap-Up Time therefore significantly decreasing contact center costs.  Go to http://www.callminer.com/white-papers.htm for a copy of the step-by-step guide.