Our KPIs used to be based on the dim light of what we could track. No longer.
6 minute read
Bruno Bertini (CMO at 8x8) and I recently discussed how CX capabilities have evolved, but the markers of success have stayed relatively the same. This blog post is an extension of that discussion and features choice excerpts from our conversation.
We’ve been talking about the valuable data hidden in conversations for a while now; there’s so much golden potential in what we finally have access to. In parallel to all this freshly available context, Bruno and I discussed whether or not the old ways of measuring success reflect our new abundance.
For decades, we’ve aimed to understand customers by recording everything around the conversation. Listening in on every single call wasn’t reasonable or even possible, so the contact center and the sales floor ran on the only data that we could reasonably track in bulk: call volume, handle time, and CRM fields a rep filled in after the fact.
The data that did exist wasn’t much better, because the vast difference between what the rep logged and what actually happened was where all the vital information was hiding.
Those numbers and scattered details turned into our benchmarks and KPIs. They defined the difference between success and failure on the job. I think we’ve always looked at the information we’ve been given as gospel truth.
We’ve rarely paused to question if they were the ones that tied most clearly to sales. Take car sales, for example. Is the clearest KPI whether the customer did a test drive, or is it more variable, like happiness with the overall experience, a positive interaction on the sales floor, or something else entirely?
But we built incentives and coaching on the numbers: calls started, calls ended, thumbs up, or thumbs down. We’re still optimizing for out-of-date metrics that might have little or nothing to do with what the customer wants or feels.
It’s not that everyone got it wrong; to grant some credit to the CX operators of yesteryear, data and KPIs were limited by what we could track at scale.
It was a dark art how somebody spoke to someone on the phone. You never knew, it wasn’t recorded, and you had no insight. Transcribing and reading a thousand phone calls a day wasn’t reasonable. We didn’t have the technology, so we counted what we could. What happened on a phone call was mostly a black box to everyone but the people experiencing it. We weren’t aware of the full potential that lay within, so we didn’t know what else we could measure.
Or, to use a more pleasant metaphor, it was like a wrapped-up present that we weren’t allowed to open yet. The call was always the richest record of why a customer did or didn’t buy; we knew there was something good inside, but all our guesses about what it was could only be based on size, relative shape, or the color of the wrapping paper.
The black box of the conversation was why voice was ignored as a valuable source of information. It was the channel, the method, but not much beyond that. But now, transcription is widely available, and AI has become useful enough to help scale the busy work when it comes to voice. We can finally rip off the wrapping paper and see what’s inside.
We’ve been looking for decades at how long it takes to sell a car. Where do you have to advertise it? What’s the demographic of the area? All this information became numbers on a piece of paper. A simple phone call has a hundred times as much information. It’s the golden egg within a phone call.
With access to all this fresh new information, it’s no surprise that 88% of organizations now use AI in some form. Moreover, according to McKinsey, redesigning workflows is the key to a successful AI implementation. The heaps of valuable data and the added AI layer have led us to a critical pivot: we now have to decide anew what is actually worth measuring and tracking.
The tricky part of all of this evolution is that the things we used to count were, well, easy to count. The things we now acknowledge matter more—steeped in subjective sentiment, tone, and emotion—are way harder to quantify, compare, and measure. That forces us to reckon with determining what actually drives a sale.
“We know now that there is a material benefit, there is a level of insight, there’s a level of changing your KPIs, improving your operations, building a culture of selling to the extent that makes your expertise better.”
Bruno Bertini
CMO, 8x8
That’s the work to be done. There’s still no definitive, standardized answer, but we at least know we can’t stick with the stats of yesteryear. We’d be willfully ignoring the valuable information that we finally have access to. And while we’re deciding, there’s no need to stay locked in place.
While we’re figuring out the KPI nuances, we can at least determine how the workflows will shift. We can already improve team behavior and training. Busywork is no longer the metric, and so people are now free to shine in the part of the conversation that has the most impact on the final outcome.
AI can handle the reading and the boring repetitive stuff: transcription, summaries, reporting, and documenting, 90% of the work by volume. Humans stay in charge of everything experiential.
If you can have your skilled sales executives just doing that 10%, the delight, the showroom experience, that’s the bit they should be doing. If you take away the mundane and ask them to do the stuff they really like doing, they’ll probably sell a lot more and make a lot more customers happy.
Customers benefit here, too. According to a Metrigy report, consumers increasingly feel comfortable with AI but still prefer human interaction when it counts. The goal of improving measurement is to determine where people matter most and keep them there.
The previous means of measurement shaped how people were trained and the incentives that motivated them. Old habits die hard, so adjusting them requires enabling people to use the new tools and actually experience the positive difference that they make.
New AI tools enable sentiment analysis while a call is still in progress. They can identify churn risk and even prompt the agent to turn things around before the call escalates further. This is a major leap forward, enabling CX operators to intervene live rather than only react when it’s too late to retain a customer. And even with AI analysis prompting, ensuring that agents practice their soft skills will only become more important.
Realizing the full benefits of this and other similar technologies demands updated training and new ways of working. Moreover, everyone should be encouraged to rip up the rule book and start asking, “What don’t we know about our business?”
“That's the change that's bottom up. You can have the best board strategy, the best manufacturer strategy. But if you don't have that level of embedding the technology, transforming around it; that's what makes the difference." - Bruno
According to a Gainsight report, the great KPI readjustment is a cross-industry pivot: 60% of companies interviewed said their KPIs are still “under construction.” Construction takes a while, however. The organizations moving fastest are the ones building a culture around this pivot. At LSH, a manager and a sales executive recently made a bet: lunch on the line for whoever moved their customer satisfaction score using AI coaching tools. No mandate from the board. No top-down initiative.
“We are going through a transformation — from structured data that was self-reported, to understanding that we can access the richest part of being a human, which is understanding each other. If we stop and evaluate the power that comes from voice communication, human communication — it's a game changer." - Bruno
The data will tell us where the conversation went wrong. What we do with that is still on us.
Chief Information Officer, LSH Auto
Chris Gensmantel is the Chief Information Officer at LSH Auto, a premium automotive retailer redefining how dealership operations run across the UK. He's spent his career in automotive IT, with senior technology leadership roles at JCT600 and Gilder Group before joining LSH Auto, and that experience shows in how he thinks about the problem: multi-site dealership networks are operationally complex, and the technology landscape reflects it. Chris is focused on building an integrated, repeatable model that works across the network, connecting the DMS, CRM, marketing, and communications tools each function depends on rather than replacing them.
Stay current on what matters in CX and IT. Subscribe to our LinkedIn newsletter for regular analysis on the decisions shaping customer experience, technology, and AI. Clarity you can act on.
