Amazon.com’s early Contact Optimization program slashed contacts per order shipped by 70% over a 3-year period, and continued to decline as the company offered Amazon Prime and other innovative solutions. Amazon’s FCR passed the 90% mark along the way. Few companies can match Amazon’s NPS.
A major wireless provider cut $250 million from their annual spend for customer service using Contact Optimization while clawing their way back to becoming #1 in the JD Power customer satisfaction surveys.
The leading online payments provider increased their revenues by a factor of 6 while keeping headcount the same across its contact centers by using Contact Optimization.
All of these companies, and many others, achieved these results before Big Data appeared on the scene, and thanks to the power of Big Data it’s easier than ever to cut half of your customer support costs AND improve customer experience.
Collect and act on NPS-powered customer feedback in real time to deliver amazing customer experiences at every brand touchpoint. By closing the customer feedback loop with NPS, you will grow revenue, retain more customers, and evolve your business in the process. Try it free.
Common Customer Support Patterns
Over the years I have worked with pre-revenue companies and global giants with more than 30,000 customer support agents, and it never ceases to amaze me when we keep finding the same patterns (box).
- On average 34% of customer support costs answer questions from customers that should be automated via self-service or prevented using proactive alerts.
- However in most cases they don’t know this figure, and there already exists self-service options or the ability to send alerts, but customers and agents are unaware of them, or they don’t work, or they are too hard to find.
- On average another 30% of customer support costs answer questions from customers that should be eliminated by applying root cause analysis to reduce errors and clear up confusion.
- However in most cases they don’t know this figure, and they don’t know where to start their analyses.
- On average companies are delivering only 65-75% FCR in their support centers, therefore they are “failing” 25-35% of the time.
- However in most cases they don’t have any clear way to measure FCR or figure out how to increase FCR or reduce the failure rates.
- Reducing the need for customer support via automation or elimination and higher FCR improved your customer satisfaction and customer experience, and also your agent sat and their experience, in turn leading to higher customer retention and spend, and higher agent retention.
In some of my earlier posts and both of my books1,2 (The Best Service is No Service and Your Customer Rules!) I have argued that companies can reduce customer support operating costs and improve customer experience. Now with the incredible power associated with Big Data, and two tools that you can readily apply (Skyline and Snowballs), it’s getting much easier to achieve this twin promise.
Enter Big Data
Because Big Data essentially allows you to “mash up” disparate data sources into a single model, then manipulate that model with a series of “what if?” and other scenarios, apply indicated “recommendations”, and then “learn” from actual customer behavior, you can now quickly assess the opportunity to reduce demand including reducing failure rates.
One of the key insights here is to measure and report customer demand for support on a CPX basis, meaning Contacts Per X where X could be the number of orders shipped or installed base or customers. By using Skyline and Snowballs companies can easily reduce CPX by 50%, meaning that they can basically double revenues but keep the same headcount and costs to support customers.
How is that even possible? Let’s go back to the two data points I shared earlier: If 35% of today’s costs could be automated, and another 30% could be eliminated, and FCR is only 70%, then by automating half of the 34% and eliminating a third of the 30% while increasing FCR by 10 percentage points to 80% = a 50% reduction. And when you factor in “downstream costs” that can also be removed when the upstream customer support costs, e.g. unnecessary re-work or in-home truck rolls, the cost reductions for the entire company become very compelling.
Skyline is a multi-level Pareto Optimization report that breaks down customer contact reason codes into one of four “actions” by asking “Is this issue valuable or irritating to our customers?” and “Is this issue valuable or irritating for our company?”. Here are the key steps to construct and manage Skyline as a Big Data solution:
1. Re-define your customer contact coding to become “customer reason codes”, in the customer’s language, describing why your customers had to contact you for support. I like to see under 50 customer reason codes, far fewer than the thousands of codes typically offered to agents.
2. Cost out each of these customer reason codes by multiplying the number of contacts (on an omni-channel basis, not just calls or email) times AHT times labor costs. For greater measure calculate all “downstream costs” that can also be removed.
3. Debate the value/irritant questions to associate each customer reason code with one of four actions per this Value-Irritant matrix:
4. Assign an “owner” for each customer reason code so that s/he will be able to take responsibility for the action, especially those to be Automated or Eliminated.
5. Set targets for each customer reason code over the next 6, 12, and 18 months.
6. Create the Pareto chart by mashing up the CRM and cost data, showing it on a CPX and on a number of contacts basis. Include historical results (e.g., same week last year or month over month).
7. Review progress for each customer reason code against its action and targets, noting the cause and effect of self-service enhancements on the Automate codes or Lean/Six Sigma root cause initiatives on the Eliminate codes.
I have used the expression Snowballs to represent repeat contacts, since snowballs rolling downhill will gain momentum and cause major problems when your customers do not experience FCR. Snowballs is an input-output model that connects agents who are starting snowballs (not resolving the issue) and agents who are melting snowballs (resolving repeat contacts, but also resolving the first time). Here are the key steps to construct and manage Snowballs as a Big Data solution:
1. Determine overall baseline FCR in the customer support center. (The same concept works in other channels, notably web self-service and IVR systems where the term is “containment rate”.) This might be “easier said than done” but using Big Data mash up you can factor in these inputs to estimate FCR today:
- Customer comments post-interaction to “Did we resolve your issue today?”
- Agent notations to “Did you resolve this issue?”
- Noting if the same customer contacted you again for the same issue within, say, 4 days
- Text analytics extracting VOC, e.g. “This is the third time that I’ve had to ask you about this!”
2. Split out FCR across the new customer reason code since not all issues or codes produce repeat contacts. If the average is 70%, some issues are the 20s while many might be in the 90s. See also my earlier post “Get Rid of Average Thinking, Make Every Experience Count”. [Bob, please link that column December 2016]
3. Track FCR and Snowballs at the agent level to identify which agents are causing vs. melting the Snowballs.
4. Roll up to the team level, to managers of supervisors, to centers, and to enterprises if you also use 3rd-party outsourcers. Here you discover that some agents or teams and some centers or enterprises produce FCR in the 90s (fewer than 10% Snowballs), while others are sinking your customer experience with FCR in the 40s (more than 50% Snowballs).
5. Dig into the reasons for Snowballs and analyze those agents, teams, centers, and enterprises that somehow achieve well above average FCR. This could include pressure to handle too many contacts per hour, outdated knowledge articles, inadequate training (always the first cost to cut, right?), and weak coaching.
6. Apply recommended changes to overcome the presumed reasons, and via machine learning discover their relative success.
7. Celebrate that success! I’ve seen teams and enterprises challenge each other to produce higher FCR and/or melt more Snowballs, an excellent bonding process.
Stop “coping with demand” for your customer support operations, and challenge that demand by setting aggressive but feasible goals like 50% reduction in CPX over the next 2 years and a 10-15 percentage point improvement in FCR.
Engage all stakeholders in this quest, not just your customer support teams and 3rd-party partners but also Marketing, IT, Logistics, and Finance since they will also benefit from lower “downstream costs” and higher customer engagement and loyalty.
Use the power of Big Data’s mash up, modeling, recommendations, and learning to gain the benefits of Contact Optimization using Skyline’s Pareto reporting solution and Snowballs’ input-output reporting solution.
Celebrate success along the way!!
1 The Best Service is No Service: Liberating Your Customers From Customer Service, Keep Them Happy, and Control Costs (Wiley 2008). There are 7 Principles of Best Service starting with “Challenge demand for service” rather than adding more capacity.
2 Your Customer Rules! Delivering the Me2B Experiences That Today’s Customers Demand (Wiley 2015). There are 7 Customer Needs, 39 Sub-Needs, and 4 Foundations that lead to a winning “Me2B” culture.