ICMI is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


A Timely Conversation: Interval Scheduling in the Contact Center

Today it’s all about CX, #Custserv, NPS, and FCR.  Too many buzz words? Sometimes I think we couldn’t exist as an industry without acronyms.  The bottom line is we need happy customers and engaged agents. We have all heard the idiom, “You can’t manage what you can’t measure” but the reality is, unless you know what’s in a number, the number may not tell you what you need to know. If you want to deliver consistent and repeatable customer experience, you have to create an environment conducive to repeatability.  That’s where your friendly neighborhood WFM team comes in.

Prelude to a Catastrophe: "The Current Quiet Interval Will Not Last..."

I read a blog written by Dan Hunter.  I just knew it had to be about workforce management, but she is actually a science blogger and science fiction writer.  As I reflected on this, I thought “Yeah—and a lot of great folks I worked with through the years would say WFM is science fiction.” Last year, a front-line agent even put in an engagement survey “They say we have to do this because of forecast—but c’mon..no one can really predict how many calls we are going to get!”  The reality is, we have tools, historical data, and (I know this will raise eyebrows) intuition.  Forecasting is a science, but also an art. I have asked forecasters throughout the years, “Do you believe it?”  If I hear—"Well, that’s what the tool says” it’s time for writing sentences on the chalkboard.  If I hear, “Yes, based on current trends, normalizing noise and inputs shared by partners-it’s the best I can do.” then it’s time to rock and share the plan with partners.

Contact center WFM: interval schedules

What’s in an interval?

Looking at daily, weekly and monthly numbers is dangerous. If you look at your monthly service level and see you made it, you say “great.” In reality, you may have been slammed to the teeth and holding calls for 5 to 10 minutes the first week of the month, but sitting around the rest of the month. If you solve weekly, you might be so busy you can’t think on Monday and Tuesday, but the numbers may not tell the full story.  If you solve at the daily level, you might be sitting at 99% SVL at 4 PM, but hanging on for dear life by 11:00—but thinking “whew” we made our number!  The opposite can happen.  I’ve seen companies who are running smoothly, but send agents home and purposely let queues hold so they can save money. Again—solving for numbers, not customer and agent experience.  I guarantee the customer who waited 10 minutes to have his call answered does not think fondly of his experience. And agents who work the busiest shifts will catch on and either try to move to more optimum shifts or leave! Only when you solve at the interval level will you see true consistency and across the board satisfaction. 

Join Todd at ICMI Contact Center Demo! He'll be speaking in session 101. 

Before creating an interval forecast, you need to define the interval.  Let your business dictate what works best for you.  Standard intervals are usually 15 minutes or 30 minutes. If you have calls that average 20 to 30 minutes, 15 minutes may not make much sense.  Some agent tools only report at the hour (example—chat and social tools). You may have to run with that.  Good workforce management tools can be configured to the interval that works best for you.  Match your interval history, not just volume but AHT.  Look at realistic service level goals and related occupancy (it will be different in busy and non-busy intervals).  Plan shrinkage.  Put it all on the table and begin the adventure.

How much? Volume

Contact volume is what most folks think of when you mention “forecast” It’s usually in the first column or “box” on a report.  Often, it’s heard, “Well, volume was right on forecast. Why didn’t we make service level? “I’ve actually changed reports to show capacity and demand as the first item seen, to illustrate workload.  At the interval level, look at historical daily patterns.  Look at abandons (How many calls were re-dials in busy periods? If you are adequately staffed in the future, these won’t occur. If you are not, they will happen again. Adjust for fluctuations (more about anomalies later.)  Remember each day of the week can have its own unique pattern.  In some businesses which week of the month can change daily volume distribution (example—banking==pay days). Look at Holiday factors and marketing input. Compare these items against your WFM tools “curves” and adjust and lock.  Be prepared to explain the “whys.”  You know they will ask.

How long? AHT

Utilize a similar process for Average Handle Time (AHT).  In few circumstances do contacts have the same AHT at every time of day.  During working hours customers may be looking for quick answers, while in the evening they may be more conversational. Tenure of agents and who works when can lead to a time of day variation. Those with a graveyard team know that in most cases that’s a whole different world. Additional considerations include sales drops, web page updates, changes in tools, and changes in processes. Your marketing team may decide they want every call documented differently during a sales drop between 5 and 7 pm.  That is going to add 15 seconds to every call relating to the ad. If you don’t adjust your AHT (based on a percentage of sales), you will be short staffed and what could have been a great revenue driver will turn out to be a flop—all because interval AHT was not considered.  Make sure to build in a glide path, too, for when new hires will be on the phones.  Know the rate at which they “get up to speed" and use it. If they are in a “nesting” period, make sure you know what their schedules are and when they will take breaks.  Failure to plan is planning to fail.

Can I breathe? Occupancy

Occupancy and service level are part of the same staffing equation and are not independent entities.  I remember when I started a new job a while back.  I had been there about a week when the VP called and asked “Why is SVL so high! We ran about 97% yesterday” (for the day by the way).   I asked the real-time team and received the answer, “We don’t solve for service level—we solve for availability.”  I asked what the goal was.  The answer—25% daily!  Service and occupancy (or availability) goals had been sent entirely independent of each other, in short, making it mathematically impossible to achieve both! They were holding extra agents at night and even asking for extra hours to get a higher daily availability!  On the flip side, I once had a new contact center leader who came from the accounting world ask, “How much could we save if we lowered service level to 65% in 5 minutes and ran at 100% occupancy?”  My answer was that we would have so much turnover and customer angst we would lose money.  I share these two examples as the occupancy equation really applies to intervals. During busy times of the day, you are going to run high occupancy to make service level.  Likewise, you must sit available to ensure capacity on off-peak hours. If you don’t look at interval parameters, you will not be able to create a uniform customer experience.   Agents will burn out during busy times of the day. AHT will increase due to high wrap (agents are just trying to catch a breath—they want to help customers but just can’t) and higher talk time will be inevitable (30 seconds added to every call caused by customers venting about waiting!). Occupancy at the interval level is imperative to success.

Investment in our lifeblood and work-life balance! Shrinkage

We need to look at shrinkage in a few different buckets if we want to create a successful plan. At the highest common denominator, let’s define shrinkage as everything that puts an agent in anything but a customer facing state.  Some of this is “all day shrinkage” Things like vacation, sick days, all day training or maybe a special project. There is also “daily event” shrinkage—breaks, lunches, one on ones, etc.  There is also “sometimes” shrinkage-- things like focus groups, maybe a monthly team meeting, one-off training, etc. You can see that all of these things don’t happen all the time, so why would you put in a blanket shrinkage number for all day every day? Account for what will be happening and adjust by time of day. Also—plan time off allocation in advance and communicate often.  In a seasonal business, agents know they are needed at the moment of truth.  They get it!

One word of caution.  If you really want to drive an efficient operation, schedule your business around your customers, not your customers around your business. What do I mean by this?  I once saw a weekly team meeting scheduled at 2:00 PM every Wednesday. This coincided with a weekly ad campaign that was consistent. I brought this up and was told: “that time works best for the leaders.”  In short, we were taking agents way from customers at the greatest hour of need.  Likewise, I’ve seen training requests come in for times of day agents weren’t even in the building!  The answer I received was “can’t you move their schedules, that time works best for the trainer.”  There are of course times it can’t be avoided, but by and large, through planning, we need to be there at times that cause the least impact to our agents and customers. After all, aren’t they (respectively) the face of our company and the reason we are even here?

It’s not always the forecast

There are times when historical data doesn’t really tell the story of “why” and shouldn’t indicate a need to change forecast. The data actually suggests a need to partner and “fix” a behavior or tech issue. Partnering with operations or IT can alleviate the need to tweak a pattern. Here are a few examples:

Late to phones/early off phones.  Sometimes folks are late getting on the phones at certain times of day.  This could be due to a logjam at the time clock, a process that causes them to need to stop by a leader’s desk, or even just one-offs that need a quick cup of coffee every day.  I have seen a process that “let people leave 5 minutes before their shift ended” to avoid being stuck on call. Finally, I have seen daily IT updates that caused everyone to need to reboot at a certain time of day.  In all of these cases, shrinkage could be calculated and utilized, but wouldn’t it be better to fix the problem?  I mean, what would senior execs think if they knew you budgeted time for missing in action team members.

AHT. Many years ago my WFM team was “chasing the tail” in afternoon intervals.  We continuously adjusted staffing, changed interval shrinkage and optimized breaks around a period we saw an AHT bump. Finally, one of the forecasters shared “You are sending me on a fool’s errand!  I am playing whack a mole.  Every time I make changes, it shifts to somewhere else!” She made sense.  We took a deeper look and used the "five why" exercise.  The cause was a few agents with an extremely high wrap/hold time in the late afternoon.  It was a time of day when there wasn’t a large number of staff, and just these few folks changed our numbers.  I went to the floor to discover why. What I found was we had a gap in floor management during this time of day.  For the next couple of days, I just hung out on the floor.  AHT came right back in line.  I passed it on and the evening leader changed their schedule a bit.  No one ever said a word to the agents.  It just “fixed itself.” The moral of the story, find out “why” before you change patterns.

Measure twice, deliver once

A chapter in a contact center book could realistically be written on every tenant mentioned above. It could be seriously argued you could create a series of tomes called “The Encyclopedia of Anomalies.” Of course, not everyone remembers what encyclopedias even were (maybe cave man Google?). The key is remembering and remembering in detail.  What do I mean? Well, we’ve all heard the famous quote “Those who cannot remember the past are condemned to repeat it.”  Do you know who said it?  Do you know why? I’ll give you a hint.  The man who said this wrote “A Life of Reason.”

We need to document deviations and normalize for them in the future to ensure the most accurate forecast is achieved.  Our WFM tools, cool as they are, are only as good as the data stored or input. If you don’t adjust for the deviations, you will forecast them (or your tool will). You will create a false staffing template, and wind up with unhappy agents and unhappy customers. Best practice? Partner with your operations and analytics teams. Most of all, listen to your agents and customers. Remember, when you peel an onion, you aren’t going to find a radish—just a truer form of onion.  That’s something you can count on.

Educate and Engage

One final tip that will lead to breaking down silo walls in the contact center—make sure you share and teach the importance of the interval. Have weekly interval review meetings with management.  Hold focus groups with agents that share what WFM is (and isn’t).  Share blogs. Let folks honestly know four things 1) Why we need to solve for intervals 2) We aren’t always going to be perfect 3) What’s in it for the customers and 4) What’s in it for them. I could, of course, share more, but my clock says in the next interval I have a meeting and I don’t want to be out of adherence. Customers and agents are depending on me.