Customer Success Story

Salesforce Slashes Escalations and Gains Manager Productivity with Real-time Sentiment Data

Salesforce has deployed SupportLogic’s support experience (SX) management platform across its core businesses – including Sales Cloud, Marketing Cloud, Service Cloud, Mulesoft, and Tableau.

Leveraging signal extraction for sentiment analysis and escalation prediction, Salesforce cut their escalation rate by 56%, acted on product insights surfaced from customer sentiment signals, and returned an hour of productivity back to every support manager’s day.

Founded

1999

Company Size

79k+ employees

Website

salesforce.com

Industry

SaaS

Proven Results with SupportLogic

56%

REDUCTION IN ESCALATION RATES

<2%

ESCALATION RATE SINCE USING SUPPORTLOGIC

13%

PRODUCTIVITY GAIN FOR EVERY SWARM LEAD AND MANAGER

About Salesforce

Salesforce is a multibillion-dollar maker of cloud-based software designed to empower companies of every size and industry to connect with their customers through the power of AI + data + CRM + trust so they can find more prospects, close more deals, and wow customers with amazing service. More than 150,000 companies use Salesforce CRM to grow their businesses by strengthening customer relationships.

As Salesforce grew through innovation, the team turned its focus to making their customer success and support operations more efficient than ever. Coupled with changing economic conditions and the company’s desire to increase productivity, the team knew they had to work quickly to achieve success.

How Salesforce Transforms the Support Experience

Salesforce targeted an escalation target of 2%, but rates were fluctuating between 2.5 and 4%. In late 2021, the company experienced an influx of support cases which added to the pressure to reduce escalation rates. They focused on looking for a solution to predicting which cases were likely to escalate versus those that could be resolved quickly. Getting ahead of those likely escalations was critical to improving the support team’s productivity.

“We were really interested in the needle in the haystack problem. We knew we were going to have cases that are difficult, that escalate, that run forever,” recalled Jim Roth, President, Customer Success at Salesforce. “But the problem is you don’t know which ones. If you knew which ones earlier in the cycle, you could pay a lot more attention to those cases and prevent them from being so long-running.”

To start tackling these challenges, Salesforce needed better insights into customer support data, and needed those insights now. The key to this change was in leveraging customer sentiment scores. Sullivan recalled an aha moment she had when reviewing data that showed how sentiment could predict CSAT scores.

“We noticed a sharp uptick in negative sentiment, and sure enough, two weeks later, our weekly CSAT scores took a significant hit,” Sullivan recounted. “We had observed these signals in the sentiment analysis from the preceding weeks, but at that time, we were still acclimating to the potential of sentiment signals, and we underestimated the direct impact of immediate action on CSAT results. We had the foresight weeks in advance, but we failed to act on it. Moving forward, we aim to respond promptly when we detect these early warning signs. To do that effectively, I emphasize the importance of having reliable and actionable data at our fingertips.”

Knowing that Salesforce needed a comprehensive platform that provided real-time insights to improve its support experience, Sullivan also had to consider the company’s scalability, productivity, and ROI expectations.

Salesforce initially launched SupportLogic’s AI-powered support experience (SX) software as an easy-to-deploy solution that provided comprehensive sentiment analysis out of the box for its Marketing Cloud and Mulesoft solutions. The 40+ customer sentiment signals SupportLogic detects far surpassed other solutions Salesforce evaluated, and those signals would form the basis of its efforts to reduce escalations.

Salesforce started by fine-tuning the SupportLogic machine learning (ML) models to reflect the nuance and history of its customer support efforts and identify when customers were confused versus frustrated. But, the company quickly shifted into success mode as it saw clearly that escalation rates were correlated with SupportLogic sentiment signal data. These insights helped Salesforce get ahead of escalations such as by offering phone calls earlier in the support cycle and giving swarm leads more data to be proactive when pulling in help from SMEs.

“SupportLogic has powered our ability to take proactive actions throughout the journey of a case. We can detect real-time negative and positive sentiment of our cases as they are progressing through the lifecycle of resolution.” Sullivan said. “Our swarm leads are able to quickly engage to help engineers with an escalation alert, de-escalate the situation and provide real-time product training or coaching to the engineer. This motion not only improves the customer experience, but also improves the support engineer’s experience with live enablement.”

Customer sentiment signals from SupportLogic also provided a two-week leading indicator on CSAT, which expanded the benefit far beyond improving escalation rates to understanding Voice of the Customer (VoC) and boosting support productivity – and even taking a more data-driven approach when collaborating with other stakeholders.

“SupportLogic is revolutionizing my interactions with business partners. When I engage with the Engineering team, I bring data that not only highlights a negative sentiment with a specific product area but also identifies the source of frustration and cost escalation. This contrasts starkly with my previous approach of merely presenting case volumes. Today, I present actionable sentiment insights, enabling us to proactively steer our course and effect meaningful change.”

With sentiment signals giving its swarm leads access to real-time data and drive support cases to the desired outcomes, Salesforce was able to cut its escalation rates by more than half for its Marketing Cloud product support in less than 3 months. Integrating SupportLogic with Slack also streamlined alerting and made the teams more effective.

“Our true escalation rate dipped from 3.9% in February to 1.7% in April, which correlated with a large increase in actions taken on escalation predictions from SupportLogic,” said Jennifer Weber, Director and Chief of Staff, Technical Support, at Salesforce. “That’s a reduction of 56% in one quarter, and we’ve maintained an escalation rate of 1.7% for 6 months now.”

The real-time customer sentiment data from SupportLogic also boosts support productivity at Salesforce by streamlining and accelerating how support teams work. With SupportLogic, teams can now manage 50 cases instead of 3 due to getting alerts through Slack and having the ability to easily search through support tickets.

“We’ve seen productivity gains of an hour per day across our 60 swarm leads and 25 managers,” Weber added. “That’s 85 hours per day we’re saving in total because they don’t have to dig through as many support cases to identify problems. And, it’s made our CSAT more predictable because we can estimate CSAT dips caused by suboptimal case prioritizations.”

Salesforce now has a new outlook on the customer support experience with real-time sentiment data it can use to improve outcomes, increase productivity, and drive more customer support improvements. It’s also expanding SupportLogic across additional Salesforce product lines and finding ways to use sentiment and support data to uncover adoption engagement opportunities.

“Now that we’ve harnessed the power of data, I’m armed with more actionable and real-time signals than ever before,” Sullivan shared. “We’ve already witnessed an impressive 56% reduction in escalations, which directly correlates with decreased negative sentiment and costs. Escalations tend to attract widespread attention and resources, which is not what customers desire either. Support is rapidly becoming the new, exciting frontier, as we play a pivotal role in driving a significant shift towards enhancing the overall customer experience.”

The Challenge: Reduce Escalations, Improve Productivity, and Get Better Insights

Salesforce targets cutting escalations by 56%.

Learn More

The Solution: Real-time Insights from a Centralized Platform

Salesforce uses customer sentiment signals to get proactive.

Learn More

The Result: Fewer Escalations and a Better Support Experience

Salesforce cuts escalations by 56% and sees productivity gains.

Learn More

In the Spring of 2023, Salesforce rolled out these capabilities to the entire customer-facing organization and expanded into SupportLogic’s quality assurance and automated coaching products. SupportLogic’s full suite of support experience products allowed Salesforce to consolidate vendor relationships and ensure a company-wide adoption process with less friction.

Katherine Sullivan added, “I eagerly anticipate taking our utilization of SupportLogic to the next level, as it promises to facilitate our scaling efforts and offer deeper insights for enhanced case management. We’re poised to dive into questions like ‘How are we presenting ourselves?’ and ‘What insights can we glean from our choice of words and communication channels?’ These present real opportunities for coaching and improvement. Our focus will be laser-sharp as we pinpoint areas ripe for enhancement, ensuring we become even more efficient and effective in our approach.”

“Whenever we witness a surge in negative sentiment, our team springs into coordinated action, and the outcomes we achieve are consistently on target.”

Katherine Sullivan, SVP, Customer Success, Salesforce

Why Salesforce Leadership Loves SupportLogic

Jim Roth

President, Customer Success, Salesforce

Katherine Sullivan

SVP, Customer Success, Salesforce

Jennifer Weber

Director and Chief of Staff, Technical Support, Salesforce

We were really interested in the needle in the haystack problem. We knew we were going to have cases that are difficult, that escalate, that run forever. But the problem is you don’t know which ones. If you knew which ones earlier in the cycle, you could pay a lot more attention to those cases and prevent them from being so long-running.

We can detect real-time negative and positive sentiment of our cases as they are progressing through the lifecycle of resolution. Our swarm leads are able to quickly engage to help engineers with an escalation alert, de-escalate the situation and provide real-time product training or coaching to the engineer. This motion not only improves the customer experience, but also improves the support engineer’s experience with live enablement.

We’ve seen productivity gains of an hour per day across our 60 swarm leads and 25 managers. That’s 85 hours per day we’re saving in total because they don’t have to dig through as many support cases to identify problems. And, it’s made our CSAT more predictable because we can estimate CSAT dips caused by suboptimal case prioritizations.

The Details Behind the Success

Hear from Jennifer Weber, Director and Chief of Staff, Technical Support, as she breaks down how Salesforce uses SupportLogic’s real-time sentiment signals to improve business outcomes.