Jul 1, 2020

Raising Your Company’s NPS: Are You Being Played?

By P.V. Kannan

Without giving anything away, the film “Now You See Me” is a magic-themed heist movie with a lot of deception, sleight of hand, twists and turns. I enjoyed watching some of the characters getting taken for a ride, especially when they deserved it, however, I imagine I would feel quite differently if I were the one being tricked and deceived.


Recently, Fortune wrote about the “small piece of data that reigns supreme” known as the Net Promoter Score, better known as NPS, and how it can be manipulated. Invented by a Bain consultant 17 years ago, its beauty is in its simplicity. It asks consumers “on a scale of zero to 10, how likely is it that you would recommend XX to a friend or colleague.” To develop the score, companies take the percentage of people who answered 9 or 10 and subtract the percentage who answered 6 or lower. The result is a number that ranges from -100 to 100.


Today, getting the highest score is something every company strives for. In fact, many executives are compensated for raising their scores, which is why some technology vendors promise that if you use their products, your scores will rise. In fact, a lot of vendors promise that, but how do they really deliver what they promise?


Improving Company NPS “Magic” Debunked


If a vendor tells you they can “guarantee an increase in NPS,” you should be very wary. The improvement in NPS happens over time as companies learn how to operationalize chat and improve the services they offer. In one scenario that I witnessed, there is a technology loophole that saw one company’s NPS suddenly jump by 20 points in the same channel. When you replace “Platform A” which is doing chat, with “Platform B” doing messaging, that should not result in a jump of 20 points. No company’s score suddenly jumps by 20 points. Ever. The sudden jump is because of a trick – a software flaw that can be easily exploited.


In recent years, contact centers have largely moved towards digital interactions with consumers, and as they do, they’ve been adopting chat and messaging platforms to enable their agents to interact with those consumers. Historically, chat platform providers (providers of software that agents use to chat with consumers) have systematically collected the data on how consumers felt about the interactions afterwards. Some even provided a URL that would go to the customer by email.


The problem with that is that there is a loophole that enables someone to use the same link multiple times. What that means is that a rogue agent could change their email address and use the link to rate their own interactions themselves to raise scores. When contact center executives see the rising scores, they seldom question it, and very few examine what is happening in detail. They think they’re providing great customer service, when in fact they’re being played.


To understand how this can happen, it’s important to understand the difference between synchronous and asynchronous chat. With live chat, or synchronous chat, an agent has a conversation with a consumer, and as soon as the conversation ends, the chat is closed and a survey goes out. Some systems however, offer asynchronous messaging, meaning that the conversation stays open until the issue is resolved. The benefit of asynchronous chat is that it allows a consumer to walk away from the conversation (e.g. to have lunch) and then come back to where he/she started.


The downside asynchronous chat is that an agent can leave the conversation open even after it is resolved, and with some systems there is a flaw that will eventually close the system without a survey going out. If the agent closes the conversation, the survey goes out to the customer right away, but if they don’t close it, the conversation closes after 24 hours and no survey goes out. That way if the agent has had a really bad conversation, they don’t close it. Some vendors have deliberately designed their systems so that they don’t send a survey. When a vendor says that your NPS will jump 20 points, that’s how they do it.


A secondary effect is that the volume of conversations an agent can handle goes down. A major telecommunications provider that used to handle six conversations per hour has now dropped down to two. Sixty five percent of the conversations the agents handle are now not closed. Now it’s not only inflating NPS, but it’s actually costing the company money.


Customer experience leaders who’ve seen their scores go up are now under tremendous pressure to keep up the lie. In some cases, CX leaders have put so much time, money and effort into getting a new system up and running, that they rely on NPS to justify their purchasing decisions. Even if the system is flawed, it’s hard to change course now. However, as leaders, you now have an opportunity to make sure the agent doesn’t manipulate the survey.


Here Are My Recommendations to Improve NPS:


  • Examine Your First Month of Data – Examine the first month of data on your messaging systems. Most conversations are closed within 20 minutes. If that is not the case, it could indicate a problem, either with the software or the operationalizing of chat.
  • Ensure that Conversations are Closed – If your messaging platform does not send a survey every time it closes, get the vendor to fix it. Ask them to patch the system to ensure this functionality.
  • Look at Unclosed Conversations – Look at a sampling of system-closed conversations. Read a random selection of those conversations and see what the sentiment is. If the conversations didn’t go well, you will know there’s a bigger problem.
  • Look for Second Generation Platforms – The first generation of chat platforms were not built for asynchronous messaging. Look for a platform that was built to handle asynchronous messaging from the beginning, and get accurate NPS ratings from the vendor.