Foundational Routing Practices: Clarifying Skill-Based Routing

Predictive techniques are enhancing routing in the contact center — when to reach out and what’s the optimum match between customers and resources. However, prediction doesn’t deal with training, contractual requirements or even the simple treatment models around hours of operation. Basic routing practices provide the foundation on which advanced routing rides.

I love to garden, flowers not vegetables; but when I first got into it, I was overwhelmed with the possibilities. And whatever I did seemed random. Then I came across an article that gave me some simple starting steps.

  1. Define a pattern so you have something blooming throughout the garden all the time.
  2. Within that pattern, start with your favorite plants considering color, shape and water.
  3. Repeat plants that are calming to look at and to create.

As you look at your contact center, it shouldn’t seem random, it should automatically handle spikes, it should be easy to evolve.  As I work with my clients, I find that basic practices in routing are required as a foundation to address pains today — and to enable success in the future. Just like my garden, you must define patterns to cover different situations, that reflect unique goals and that stay simple. And then repeat that.

In this three-part series, I’ll share my most common pieces of advice on routing. We’ll talk about common routing patterns, when to use them and how to get started in defining them for your business. The three most common routing models that companies need are shown below.  In Part 1 of this series I’ll outline skill-based routing in particular.

Skill-Based Routing Borrow/Lend Rules Routing Relationship Routing
Need: First contact resolution (or other key results) Need: Protect small teams Need: Personal relationship
Definition: Route to the appropriately skilled agent and then gradually expand to acceptable (but still capable) agents. Definition: Do skill-based routing with target expansion but protect small teams so that they don’t get overrun. Definition: Route to the personal agent for the customer and offer alternatives when they aren’t available.

Skill-Based Routing
This is the foundational model because it’s critical that you distribute work only to capable people. Let’s start with definitions that I use.

Skill: The characteristic of an agent that doesn’t change unless agent changes. Examples include training, job function, language, licenses.Rule of thumb is that new agents have about two to three skills. Good agents have about five or six skills.

Proficiency: Star agents have 10 to 12 skills. Attribute of a skill. There are two uses:

  1. Indicate how adept an agent is at a skill —  expert, experienced, new, for example.
  2. Indicate skill preference for different types of work — primary, secondary, tertiary, for example.

Rule of thumb is to only have three to five different proficiency levels.

Work characteristics: Attributes of the work. Example attributes: intent, customer segment, product, media channel, recent activity.

Skill-Based Routing Flow

Let’s dig a little deeper into each block.

1. Queue for preferred agents. Using work characteristics, define the skill expression that represents the optimum match for this piece of work. The skill expression is a collection of skills and proficiencies. Here is an example:

We need to follow up on a billing correction inquiry for a Chinese-speaking customer via email.

    • Work characteristics = billing, language = Chinese, media = email
    • Preferred skills/proficiencies = billing + Chinese + email/primary

2. Timeout passed. Waiting is a critical part of the routing logic. You need to let your staffing plan have a chance to execute by waiting for the planned (preferred) agent. Staffing plans almost always assume a service level that is not 100% of work handled immediately. So, it’s OK to wait. Here are some recommended practices around managing the timeout step.

    • If wait time is a key metric for your business, the total timeout values until you are fully expanded should be less than your target wait time. This gives you a chance to meet the wait time goal. For example, if your wait time target is 2 minutes, you probably want to be fully expanded by 1 minute 40 seconds.
    • Different types of work have different patience tolerances. Therefore, the timeout value might be different for different work. Allow variation in your logic.
    • If your staffing strategy assumes use of both preferred and “next best” agents to handle your expected volume, put a wait step in for the preferred staff. However, wait less time than you might otherwise since you know you’ll need the next set of people.

3. Expand queue/route to next best agents. The shorthand term for this step is “target expansion,” where the idea is to loosen the skill expression to have a larger pool of agents who could be routed work to handle either spikes in traffic or because your staffing plan requires more than the preferred pool of agents. If we continue our example above, the target expansion skill expressions might be:

We need to follow up on a billing correction inquiry for a Chinese-speaking customer via email.

    • (Orig) Work characteristics: billing, Chinese, email
    • (Orig) Preferred skills/proficiencies: billing + Chinese + email/primary
    • Target Expansion 1: Change email proficiency to secondary or primary, so that agents who handle email in a backup capacity can receive the work.
    • When fully expanded: Anyone in the billing organization who can type in Chinese and can handle email can be targeted (billing + Chinese + email).

Considerations With Skill-Based Routing

  • For many traditional contact centers, the “expand” step is done manually by reassigning, or “reskilling” agents. If you live in this world, you’ll understand the phrase “chain of pain” because you have to chase work and then reset the assignments to keep from creating a new fire after putting one out. I strongly suggest you automate this step with technology. Remember: Only skilled agents should perform work; a warm body isn’t a skill. You’ll waste everyone’s time if you distribute work just to meet a service-level goal and then transfer it or do it incorrectly.
  • The typical target expansion model is from highest skilled to skilled-enough — trying to use your best people first but settling for people who can do the job if the best are busy. However, optimal matches are the real goal, for example, medium-value to medium-proficiency, may be your starting point. Or you may want to save your top people for consulting or transfers, so you initially route to your good-enough staff and only expand to your best when it’s busy. Your expansion model needs to match your customer experience strategy to make it a “best for you” practice.
  • Don’t go crazy. To keep your routing from becoming unmanageable, expand fewer than five times. Try to make each round as big of a group as possible.
  • The timeout step means that you’ll have the dreaded calls-in-queue/agents available situation. You need to train your staff and your leadership to accept this reality. The thing to unlearn is the false assumption that all work must be handled immediately. The real goal is to handle the work optimally — with the right people and in the right order.

The 80/20 Rule

In most contact centers, skill-based routing is the most repeatable pattern, it can handle a significant portion (60% to 80%) of your work volume. The only difference is mapping — which work characteristics match to which skill expression. Done well, this model ensures that you provide the optimum service experience with the right people handling the right work — supported by a staffing plan to achieve it.

We just talked about the “80” in the 80/20 rule, but it’s the “20” that causes heartburn and, in some cases, represents the high value or high-risk work. In Part 2 of this series, we’ll explore the lower volume, but usually critical models.

To learn more about delivering a seamless omnichannel customer experience across multiple channels, read this ebook.