Lean Six Sigma for the Contact Center

Recently I conducted a webinar on Lean Six sigma for Contact Centers.  There were over 500 attendees, and some great interaction.  Below I have taken a few of the questions from the webinar with my answers.  Maybe you can glean a few morsels from the interactions.

 Question: Are there recommended Lean Six Sigma training programs specific to call centers that you could mention? It could be ones that others have found useful. -Thanks.

 Answer: LSS training strictly for the contact center has been a need for awhile.  I believe that is why hundreds of people show up on the webinar series.  CCNG and Contact Center Pipeline would be a good source to look toward.  Also, because of demand, OpenSpan is considering providing LSS for the CC courses, perhaps as a standalone or additional certification for Yellow, Green and Black Belts.

 Question: How different is Hospital Call Center than Manufacturing, Banking or other commercial centers?

 Answer:  My experience with hospital contact centers has shown that there are three areas of focus, in general.

The most common is for collections activity, which includes collecting from patients and collecting from insurers.  These centers operate substantially the same as credit card companies collecting bills.

The next most common is pure customer care.  Friends, family, press etc. trying to find patients or get patient updates.  Smaller hospitals may make this a reception function, and others aggregate these calls in a central center. These guys operate like a customer care center in almost any other industry.  They have regulation (HIPAA for hospitals) and policies, and multiple systems with which to engage the caller. 

The third area is most closely analogous to the hospitality industry. These “agents” are responsible for filling up operating rooms and beds.  They may make both inbound and outbound calls, and some of the agents operate like telemarketers, others like reservationists.

I usually refer to contact centers as a horizontal spanning all verticals.  Contact centers usually have more in common with CCs in other industries than departments in their own company.  There are three general CC disciplines:

1. Sales
2. Customer service
3. Technical Support (technical could mean technology, or could mean a highly paid worker with special knowledge or training. )

 Question: A great number of the professionals within the call center are great "people" skills driven and "training" driven; less skilled in the analytical/critical thinking skills. How do you deal with this to manage some of the activity covered today?

 Answer: Great question, and not unusual in contact centers.  There are several dynamics that impact their ability to think analytically.  Most often, I find it is lack of proper tools.  Centers are drowning in data, but have few tools to organize it across all platforms and across all agents.  So, to get a true picture, these individuals drowning in data have to be picked up out of it all and given higher level views. Some centers have employed business analysts to do that for the managers, but often these guys are the first ones to be cut in lean times.   

OpenSpan Events is the best tool I know of to get this cross-platform and cross-agent data.  Once you have this, you begin to see bottlenecks in technology and resolution rates for agents much more clearly.  No longer can IT and workers blame external sources for their lack of adherence.  You know, for 100% of the calls, what the impact of any activity really is.

Training can only address the human component in a relaxed environment.  The biggest reason soft-skills training often doesn’t work is because the agent is having to juggle to many applications and rules during the interaction in a live environment.  A good analogy is almost all of us can remember a phone number for 5 minutes, unless everyone around us is demanding attention and relating other numbers and competing concepts.  Simplify as much as you can in the agent environment, and monitor and measure everything, then use technology like OpenSpan and/or business analysts to gather and digest data for your people managers.

 Question: Is there a formal calculation for CSR/supervisor ratio?

 Answer: Not to my knowledge.  There are benchmarks, and you can find these at places like benchmarkportal.com.  I try to calculate the supervisor/agent ratio only on time a supervisor is available for escalations or intervention activities (like real-time remedial coaching).  Everything else is clerical or training, and time spent there should be calculated as such.  Otherwise, you could have a proper staffing level, then heap more and more non-contact activity on them and gradually starve your agents of the attention of their supervisors, impacting both customer satisfaction and agent satisfaction.

 Question: If we have to have minimal AHT, how will it impact our customers’ experience! Does it leave a good impression to wrap it up instead spending time and rapport building.

 Answer:  You can have calls that are too short, no doubt about it. And focusing your agents on AHT is a less than optimal success strategy.  This, too, has been proven over and over. But don’t get confused by what you tell your agents and how your CFO evaluates the operation.

Here is a way to get an indication of whether your call length is optimal.  I call it the MIO, or Minimum Improvement Opportunity.

Take the AHT for your center.  Graph it on a population curve, also known as a bell curve.  It should naturally look like a bell.  If it is wide and flat, you have lots of room for improvement.

Next, pick your top five to ten agents, based on things that matter to you, not just AHT.  This could be First Contact Resolution, Net Promoter Score, Scorecard scores, or revenue attainment.  Determine their average AHT.  The difference is the amount you should be able to improve without impacting customer satisfaction and loyalty.   The reason is that these top agents are better at manipulating the technology and anticipating activity.  If you listened to their calls, you would hear fewer pauses and less filler conversation, neither of which help build rapport with a customer.  They are getting down to business quicker with the customer, which is the best thing you can do.  The things behind the scenes that these top agents do to get to resolution quicker can almost always be facilitated and/or automated.  OpenSpan is superior technology for this type of improvement effort.

 Question: What if your products are so variable that high variability is not avoidable?  Our contact center can take calls from customers about hundreds and hundreds of different areas of the product we support.  How do you reduce variability in this environment?

 Answer: This sounds like a tech support function, and it can be a bit more difficult to drive results, mainly because the subject matter experts must be involved to drive processes.  The upside is usually that training materials and troubleshooting processes are often well documented, as this serves as reference material for the reps intra-call. Here are some process path scenarios.

Scenario 1 – The center is staffed by people with hands on experience with the product because there is very little documentation and very few job aids and process flows. You need to discover what the better agents do, then document and teach others.  Breaking your call into sections of a flow, and then evaluating the best performers at each stage of the call flow will help you organize and identify the best processes. 

Scenario 2 – The center is staffed with very skilled individuals, but the support so many products they cannot get enough experience with each one to be an expert.   Several techniques can be employed here.

 First, using skills routing in the ACD and having each product set up as a skill will have immediate impact.  Note that this is not the same thing as setting up specialized groups.  The calls are not routed by an IVR or other system to a group, they are still handled and routed to a general population.  It’s just that the caller is more likely to get an agent with some knowledge of their product. If this moves the dial only 3-5%, it is still significant for your center and customer satisfaction.

Second, regardless of the products, teach a common call flow, and then facilitate with technology each stage of the flow. Take this flow as an example:

Contact start – Single Sign On customer 360, if CLI can be used.

Verify – if there are multiple ways to verify a customer is allowed to get help, aggregate these methods in one screen, and allow the agent to record the appropriate verification automatically, rather than manually.

Research – this is the stage that can be most lengthy.  Navigating to multiple provisioning and support systems and researching the issue can be cumbersome and time consuming.  Unified notes screens, which pull caller notes from multiple systems for history can get an agent oriented quickly. Unified search fields allow an agent to make a single search query but all possible systems are searched concurrently.  Contextual help, and guided (automated) activities will reduce variability in the agents’ responses.

Resolve – Once again, guided activity will drive to resolution with fewer tangents and pauses.

Record – often, to resolve issues,   agents are either doing things in systems or reading information.  Actions can be automatically recorded, and information read can be automatically noted by a clicking on the info as it is being read.  These two techniques often reduce wrap time to almost nothing, or save significant time during the call if the agents make notes while the caller is on the line.  

Fulfill – if information or product needs to be shipped to complete the activity, agents don’t generally have much to do.  If your center requires agents to initiate other actions, usually there is already enough information gathered during a call to facilitate this.  Stock levels and shipping times should be made available to the agent during the call, then the fulfillment activity should be triggered (automated) based on rules, not based on the agent remembering to do it.

Once you extrapolate all contacts to a single flow, your work then is to make sure the technology is aligned to support the flow. OpenSpan is superior at enabling this activity, but you can often do the same thing with internal analysts and IT staff. The center I ran in Tallahassee did not have the benefit of OpenSpan, as it had not been created yet, however we were able to answer the phones with a small set of agents for 1,500 different customers, all with different products.  We had 40 programmers keeping things up to date, and it is my observation that a couple of people armed with OpenSpan could do what my 40 developers did in keeping things integrated and optimal.

Question: I have a transportation company and we created a call center to inform our customers where their loads are. We have different customers and everyone has different requests. How can we standardize our calls?

Answer: You can never standardize calls.   What you standardize is the way the call is engaged by your agents.  80% of the information callers want can usually be pulled and aggregated in a common customer view.  At OpenSpan, we often pull information from multiple systems, and even external sites like FedEx, DHL and UPS, simply based on identifying the caller or the order number.

You have to think of the calls in a unified call flow fashion.  Refer to the previous question for an example of a higher level (more universal) view of a call flow.

Question: How do you apply these concepts to the at home worker?

Answer: At home workers may have their own technology, but usually they are using company-supplied computers and phone connections.  The OpenSpan technology easily resides on the desktops of at home workers.  If you supply them a virtualized environment  (e.g. Citrix), the OpenSpan technology can reside within that Citrix environment and provide the monitoring and automations within those applications.  Likewise, the OpenSpan technology can automate non-virtualized desktops.  If you mix the two (e.g. some apps run in Citrix and others don’t), OpenSpan can handle that as well, although the implementation is a bit more complex than a standard. 

Your at home agents need the same improvements and need to be monitored as if they were in house, and the concepts discussed, supported by the OpenSpan technologies, facilitate their processes as if they were in house.  Geography should be immaterial to the ability to utilize the LSS concepts, especially through the use of OpenSpan to bridge the gaps in your current environment.

Which comes first - automation or monitoring?

There are several desktop monitoring products out there, and various ways to automate contact center and back office processes. Rarely do you have the ability to get both for the price of one.  Every contact center I have visited recently has faced the question - "should we monitor for awhile then define our automations or should we automate some things we know might help, then monitor for other improvements?"  Given normal process improvement efforts, it is understandable the deliberation prior to a months long development effort. 

If you are facing this decision, you may be using a sub-optimal tool set or process to achieve your goals.  It reminds me of an effort I managed in my AT&T days.  For a couple of years, the unit I worked in retained several Foxbase/Foxpro developers for a bespoke application and database.  It was time for an update and the quote came back for multiple developers and multiple weeks. What was not recognized was that recent releases of FoxPro had incorporated "wizards" that dramatically reduced the need for the new functionality.  We re-examined our needs, and were able to meet the business objective with two days of work from a business analyst.  The contractors were not happy, but the business was thrilled, and was able to get on with the new effort months sooner.

So, what long held assumptions are you operating under?  You probably can roll out much needed automations and full desktop monitoring (invaluable for the next round of process improvement) in less time than just one of those efforts took in the past. Challenge the status quo!

In related thoughts, this is an interesting blog post from Francis Carden of OpenSpan. http://franciscarden.blogspot.com/2010/11/why-we-like-competition-at-openspan.html

Specialization is for Insects*

Several times recently, I have reviewed the operations of large employee groups.  Each of these groups faced rising complexity in the constituents they served and the systems they had navigate.  So the reaction of their managers was to make certain employees "specialists" in order to to keep up and serve their customers. 

Great short term move (days or weeks) to give management and IT time to address the real issue.

HORRIBLE LONG TERM STRATEGY!  In no case did I observe such a radically different business process that the only choice was set up special groups.  Occupancy and shrinkage get out of control the more you specialize. I witnessed groups which simultaneously had high "calls in queue" or case backlog, while also experiencing high idle time. The managers acknowledged that the condition they were addressing was NOT improving.  But they can be excused.  They were addressing the situation the only way they knew and could control, because they did not have the IT resources to do the right thing.

What is the right thing?  Look at the work flow at a level of abstraction above the tactical.  Understand that workflows and call flows almost always follow common steps, regardless of system differences per step. Then make sure upper management (executive suite) understands that throwing 15 to 20 CSR or knowledge workers at the problem indefinitely is not as cost effective as devoting 1 to 2 technical resources for a couple of months.  Time and again, this resource realignment has had dramatic positive impact.  So next time someone suggests setting up a "special queue", remind them that may work in the short term, and insist on immediately committing resources to the long term solution.  The money, resource and aggravation you will save will be your own!

*"A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects."

-Robert A. Heinlein

A Simple Formula

Since the 1970s, few departments and major corporations have had more technology thrown at them than the contact center. All of this technology was for the purpose of better managing a company's interactions with its customers. Today in many centers, there is a sense that any further technology purchases or process improvements probably will not increase productivity enough to make it worthwhile to perform.  I call this the "One More Thing" syndrome.  Unless the "one more thing" subtracts a couple of other things that have lesser reliability (throughput), keep it away.  The formula is simple and can be tracked with switch reports, ACD stats, or OpenSpan Events.  For each system, or for each process (e.g. identify caller, access account, record wrap notes) multiply the reliability of each and every part.  

A simple example: in a call that has three parts, each with a 99% completion rate, the formula predicts that the FCR would be 97%.  That is .99 x .99 x .99 = .970299, or rounded to 97%.  If only every call had only three parts or three technology platforms to deal with. 

Help for the human worker bee!

Remember those Sprint/Nextel commercials a few years back that showed humans taking flight in a big hive of activity? I think anyone who has worked in a large company or a contact center has felt like those worker bees. I joke with associates all the time about my "checkered" past and the different worker bee roles I have held. Having succeeded and failed at many endeavors and with many companies in the past, I always try to take something positive away from every experience.

As an active process consultant and project manager, I use my Six Sigma background and my experience at companies good and not so good to identify and prescribe best practices for companies far and wide. I am an unabashed promoter of the OpenSpan technologies because they are the best way I have seen to monitor, analyze and automate the activities of enterprise workers and contact center agents. If you have ever had to answer a phone and access a dozen computer applications while engaging the caller in a friendly manner and driving to a problem resolution in a few short minutes, you know what I mean when I say that us worker bees are the missing glue for the IT world. But it doesn't have to be that way. The customer or caller should be our primary focus, and the computer programs should serve us, not the other way around. In this space, I will share my observations and my recommendations with three areas in particular for this hive metaphor:

(1) I will discuss how to know what is going on in all areas of the hive and the worker bee population;

(2) I will help you analyze the activity and observations so you know what to focus your process improvement initiatives on; and

(3) I will suggest ways and means for acheiving remarkable improvements in a short period of time.

I welcome feedback and shared experiences, and will gladly link your blogs and sites so that we can build a connected worker bee community. I hope my observations spur some thought and activity on removing the stress and friction in dealing with your day to day activity.