Beyond surveys: modernizing your     approach to CX

In last week’s blog, we explored use cases for AI that go beyond call centers and opportunities to create value by applying AI in marketing, sales, and service.

This week, we’ll take a look at ChatGPT’s answer to the next question in our series: How can AI be used to get beyond the limitations of survey-based approaches? We’ll outline our own thoughts, after providing our take on how well ChatGPT did.

Chat GPT’s Answer (repeated here, verbatim from ChatGPT): 

While surveys are a valuable tool for gathering feedback from customers, they do have some limitations, such as response bias, low response rates, and limited scope of questions. AI can be used to get beyond these limitations and provide more in-depth insights into customer behavior and preferences. Here are some ways AI can be used to augment or replace survey-based approaches:

1. Social media monitoring: AI can be used to analyze social media data to understand customer sentiment and feedback. This can provide a more natural and unobtrusive way of gathering feedback, and can offer a more comprehensive view of customer perceptions.

2. Voice and text analysis: AI can be used to analyze customer interactions with call centers, chatbots, and other customer service channels to gain insights into customer behavior and sentiment. This can provide more nuanced feedback than surveys and can help identify areas for improvement in customer support.

3. Behavioral tracking: AI can be used to analyze customer behavior on websites and apps to understand how they interact with the brand. This can provide insights into customer preferences, pain points, and opportunities for engagement, without the need for surveys.

4. Recommendation engines: AI-powered recommendation engines can be used to suggest products or services based on customer behavior and preferences. This can provide a more personalized experience for customers and can help businesses understand their customers' needs and preferences without relying on surveys.

5. Natural language processing: AI-powered natural language processing technology can be used to analyze customer reviews and feedback to identify key themes and areas for improvement. This can provide a more comprehensive understanding of customer perceptions and preferences.

Overall, AI can be used to augment or replace survey-based approaches, providing more in-depth insights into customer behavior and preferences. By leveraging AI in new and innovative ways, businesses can better understand their customers and deliver more personalized and effective experiences.

How Did ChatGPT Do?

This isn’t a bad answer by any means! ChatGPT identifies ways to use AI to augment or replace survey based approaches, as well as specific sources of data to focus on for your analytics. It also calls out bias and constraints on the number of questions you can include in your surveys, both of which an AI-driven approach to modernize your approach to CX listening helps overcome. ChatGPT’s point about response bias is one where further conversation with it would have provided additional insights into not only the impact on response rates, but the importance of gathering insights from a broader sample not skewed to those with especially positive or negative experiences. After all, it’s the 3 star reviews that tell you the most about what it takes to get to 4 or 5 stars.

Engaging the “Movable Middle”

Let’s dive deeper into the response bias issue further. How does this impact your ability to make the best decisions for where to invest resources to improve your CX focused on the right customers for the right moments that matter? How does this impact your ability to deliver personalized, real time experiences using a data-driven approach to journey orchestration?

Experience management investments have the maximum ROI when they are based on actionable insights into the propensity of customers to engage with your brand, and what is amplifying or hindering their engagement. Your CX investments will have the highest ROI when they are targeted to those with a propensity to engage with your brand that wouldn’t buy the product anyway. This doesn’t mean that fixing pain points for your most loyal customers isn’t important, but that there are trade-offs for whom to focus on and the right balance between acquisition and retention oriented investments. Retention focused investments typically have a higher ROI than acquisition focused ones, but making things marginally better for customers that already have a good experience will yield a lower ROI than alternative investments.

MMA Global, the association where I’m acting as a CX subject matter expert, has a powerful growth framework called the “Moveable Middle” that taps into this key principle. The ROAS for paid media campaigns focused on customers with some propensity to engage but that aren’t already promoters is an order of magnitude higher than the control for media campaigns. Optimizing the effectiveness of campaigns targeting the Moveable Middle - including not just paid media but also your owned and shared media (social media where you aim to encourage sharing) - requires first party data to understand customers’ propensity to engage with your brand, as well as other insights to be able to tailor the content and calls to action to these customers along their path to purchase. As we’ve explained in other blogs in this AI & CX series, you can use AI to better leverage both your X and O data signals to drive a continuous improvement cycle. Experience data (X data), is what people are thinking, feeling and saying during the customer journey, and includes not only surveys, but the much larger set of unstructured data signals from ratings and reviews, social media, calls to your contact center, chatbots, messaging, emails, etc. Operational (O data) are the digital footprints your customers leave along their journey that are captured in your IT systems. These X and O data signals can be used to generate extremely valuable insights that help you target customers with a propensity to engage with your brand and to optimize the experiences you design for them. Focusing efforts on your “Moveable Middle” will yield the highest ROI for these investments, including your paid, owned, and shared media campaigns, as well as human centered design investments for physical and digital experiences along their customer journey.

Note that Net Promoter Score (NPS) can be a part of understanding the propensity of your customers to engage with your brand, but it isn’t a golden metric, nor is it a scalable approach to target customers based on limits to sample size. You need to tap into the broader set of X data to build deeper insights and create a scalable growth flywheel to drive targeting and ongoing optimization for how you engage your “Moveable Middle.” Not only will these insights be less biased than if you relied on survey based approaches given the much broader set of X data, but you’ll also be able to generate insights for a wider set of topics given the constraints for how many questions you can fit into a survey. Moreover, you’ll need to enhance your analytics of your O data and link it more closely to your X data analytics to fully realize the potential to raise the ROI on your experience management investments.

Improve survey-based approaches with AI

Although AI is turning unstructured data into an extremely valuable source of insights, surveys will remain a useful part of your approach, though the way you use surveys will evolve. For starters, relying on surveys only after the end of the customer journey is not as useful as gathering feedback throughout their journey. Micro-surveys that are shorter (3-5 questions) and with a mix of closed and open-ended questions are a better way to gather feedback that is closer to specific moments that matter and also gives you time to address pain points during vs. after the customer journey. Moreover, micro-surveys have a higher response rate. Finally, including a mix of open-ended questions along with a more focused set of closed-ended questions allows you to tap into AI as a way of developing more actionable insights. In many cases you can also use the open-ended questions as well as other unstructured data sources to predict how the customer would have answered an additional survey question, or in the case of mining unstructured data, eliminate the need to ask a survey at all (e.g., post call surveys after someone has already spent time with an agent).

Up Next Week

In next week’s blog, I’ll be writing about how AI can be used to create a more customer-centric culture.

If you’d like to see the full set of blogs in this series so far, click here. If you’d like to see some of the video interviews that address related topics, click here to see the videos done with our strategic partners Experience Engineering, Farlinium, and GK5A.

We’re looking forward to continuing the conversation!