Tested Positive: Post-Sales Sentiment Analysis

“We love working with you, but it’s time to move on….”

I’ve seen dozens of customer success managers and professionals going through the traumatic experience of opening similar emails, churn they often don’t see coming. The biggest driver of this action is often negative sentiment.

It’s no secret that Customer Relationship Management (CRM) is extremely popular and being widely used on a global scale. With AI now being applied to all enterprise software aspects, it’s only natural that it’s also boosting post-sales services. 

Our Machine Learning (ML) team here at Staircase AI has recently developed a new AI sentiment model that’s changing the way we understand customer sentiment, engagement trends, and track stakeholder relationships. Post-sales teams can now use Customer Sentiment analysis to fine tune their playbooks and boost their book of business. It’s really that amazing! 

But before I dive into the details of the technology, let’s first go over the challenges and why sentiment scoring is actually needed today. 

Adopting a Customer-Centric Approach: The Challenges

While I hate conveying bad news, it must be said that more and more companies are not able to understand where their customers stand in comparison to what the product actually offers. There are many customer blind spots. The amount of variables is growing by the day – product feature engagement, service response times, and stakeholder relationship dynamics being just a few of them.

Collecting customer feedback, or understanding the meaning of lack thereof, has always been a daunting task. I would like to shed some light on this by highlighting three important and crucial aspects of customer sentiment analysis.

• NPS Scores – Collecting Net Promoter Scores (NPS) via surveys (in-app, directly, or via email) is in all CS playbooks. But it’s getting outdated due to its inherited shortcomings. Surveys are not real-time indicators, making them a less proactive sentiment analysis option. They also paint an incomplete picture because participation rates are low (around 5%).

• Accumulated Data – I’m also seeing CS teams getting overwhelmed by the amount of customer data from the various communication channels. For starters, it’s time-consuming and frustrating to manually process this siloed information. But this is just the tip of the iceberg. It’s almost impossible to gain historical perspective and make unbiased predictions manually. 

• Unsaid Sentiment – How about an OMG stat? As per a Salesforce report, 91% of unhappy customers will leave a brand without prior notice, a CS nightmare. For example, emails can be extremely positive and friendly, not allowing the detection of at-risk accounts. For example: “Your product is amazing and we appreciate your support, but 24/7 Slack support would have been great”.

Anyways, the good news is that modern AI-technology can help address the three challenges I have mentioned here.

AI-Powered Sentiment Scoring is a True Game Changer

AI-technology is essentially enabling a shift from the traditional NPS, which requires periodic intervention, to continuous NPS. Now, all direct and indirect customer responses are recorded, analyzed, and broken down into actionable insights on an ongoing basis. Also, sentiment scoring can be done automatically, just by giving the machine the opportunity to listen to cross-channel communications. 

Customer insights, in my opinion, support a strategy that’s genuinely customer-centric. Customer Scoring helps capture and track the level of satisfaction experienced by your customer towards your company, products, or service. But that’s not all. This automated approach also reduces manual labor, while also providing an unbiased outlook on what the customers are saying. 

Need more reasons to embrace this new metric? The numbers don’t lie.

• US-based companies alone lose over $40 billion annually due to just one poor customer experience, with over 50% of the people churning altogether
• HubSpot is echoing the same sentiment (pun intended). 93% of customers are likely to stay engaged or even upgrade when companies offer good service.
• As per an American Express report, 7 out of 10 Americans are willing to spend more to engage with businesses that provide better customer support.

I told my team that understanding customer sentiment was going to be our primary focus when we started working on Staircase AI a couple of years ago. We made it a priority to harvest and use all customer signals to create truly unbiased insights.

Our ML team worked hard on the algorithm and the model training process was unique – we debated positive and negative scenarios and discarded all disputed factors, while including only rules that were agreed upon by everyone. The results surprised us all – Staircase AI can now listen to customer communications on an ongoing basis and create accurate human signals for actionable insights. 

Staircase AI is becoming an industry disruptor because of its unique ability to read between the lines. Most leading sentiment analysis models are not flexible enough to address the dynamic nature of customer sentiment and relationship dynamics. Staircase AI can detect business context and extract business sentiment combined with topics and customer journey stages to reveal risks and opportunities. 

Disrupting the Post-Sales Space

Staircase AI models currently have over 90% accuracy, while most auto-ML models are hovering around the 55% mark (for business communication). We are prioritizing this innovative human-data-driven approach to eliminate blind spots and boost account visibility. Service teams can now leverage this technology to detect at-risk accounts and identify new opportunities, all in real-time.

Staircase AI was developed to cover all bases and provide comprehensive customer sentiment analysis. The customer sentiment score you get is based on human insights that have been derived from analysis of your customer interactions – emails, Slack activity, chats, video calls, and other communication channels. This is essentially the missing piece in the modern post-sales analysis toolkit.

I do hope that you’ll try Staircase AI and let us know what you think about its capabilities. We are always working hard to make it better and help CS teams achieve better results. Your direct feedback is what makes this possible.