Like the best ‘steps to success’ plans, creating a Voice of Customer (VoC) program sounds so simple. All you do is collect feedback or run a survey and you're on your way to better CX... right?

It's true, no VoC program can exist without a collection of feedback, and if you don't have feedback, start collecting it now!

However, you will very quickly need a plan in place to create a positive feedback loop that leads to measurable CX improvements. A feedback strategy that simply collects will leave customers frustrated and dissatisfied with a lack of action. The responses you've worked hard to collect not only become stale and more irrelevant with time, but worse, you create a false expectation with customers that you're listening and will respond with action. This creates a feedback black box that can lead to a negative retention trend and customers that feel rejected.

Customer feedback is a trove of insight that tells you exactly what you need to do to earn a customer's loyalty. It's too valuable to resist. However, to create a positive feedback loop, you need to build a VoC program, and build one that is effective.

What makes a VoC program effective? An integrated, collaborative process for handling and responding to feedback with consistent, unbiased data.

OK; how do you enable this?

To simplify the path to "Effective VoC", we've broken down the process into a four-step sequence of listen, learn, act, and improve. So without further ado, let's step through the... steps. 🙄

The Four Steps


Collecting feedback from only one source can lead to a bias to build for specific users. This is why it's vital to get a holistic feedback picture by collecting from your direct and indirect sources of feedback.

Direct sources include forms and surveys where customers are engaging with an intent to give feedback: CSAT and NPS surveys, onboarding feedback, reviews, product surveys, cancellation surveys, and in-person interviews.

Indirect sources include support tickets, customer calls, social media and community chatter, and outbound replies (to name a few).

Depending on the stage of a customer's lifecycle, that feedback will exist within multiple channels, each containing uniquely useful insight into different areas of your CX.


Feedback is unstructured, meaning it has no measurable attributes. You'll need to transform your feedback into a schema so you can get measurable data. Why? Because measurable data is evidence and evidence inspires action faster than anecdotes. Knowing that something is highly requested by customers and will lead to a positive outcome helps your teams make decisions more quickly and confidently.

To structure your feedback, develop a tagging scheme to categorize feedback that works across all of your sources. This is often referred to as a "taxonomy".

We recommend starting with a simple taxonomy and building it up over-time. You can start by tagging feedback by Type, Product, and Feature Name.


What type of feedback is this?

Examples of Type tags include: Feature Request, Product Complaint, Customer Service Complaint, Education Issue, Bug, and Billing Issue.


What area of the product is this feedback for? If you have multiple product teams, tagging helps ensure feedback gets to the owner who can take action.

Examples of Product tags include: Mobile app, Integration, Admin, and Analytics.

Feature Name

Get specific about the feature that is being requested or bug being reported.

Examples of Feature Name tags include: Slack Integration, Notifications, Bulk Delete, Login, and Checkout.

PRO TIP: Use AI to automatically detect and tag your taxonomy. It'll save you hundreds--maybe thousands of hours in the long run and will free you to focus more time on the next step, Act.


Create a summary of your customer feedback by collating the resulting tags and associated pieces of feedback into a report that you can share with various teams. If you have a small collection of feedback, ~50 pieces, a spreadsheet tool like excel, Google Sheets, or Airtable can suffice. In it, calculate the popularity of each tag.

For larger collections, manual tagging and reporting becomes untenable as feedback floods in and inevitably, your taxonomy will expand. You should consider automating the tagging process so your taxonomy can remain fluid and grow. Also, integrate with a true analytics tool for real-time reports.

With reports in hand, schedule bi-weekly VoC meetings to review your reports with team decision makers. Ensure VoC reviews become a critical part of any quarterly executive meetings. By doing this, you make your customer a stakeholder and can use their feedback to develop strategies that will impact growth, retention, and expansion.


Measure the ROI of your actions. Look retroactively at your feedback trends so you can ensure your actions are having a positive impact on customer sentiment over-time. View your newly acquired VoC insights side-by-side with acquisition, activation, retention, revenue, and referral metrics. Drawing correlations between the qualitative and quantitative helps your organization find the root cause behind growth roadblocks.

Manual vs Programmatic VoC

Many companies already pace out the four “listen, learn, act, improve” steps in some manner. A true VoC program formalizes the process, with clear systems and objectives.

The starting point for smaller companies will naturally be manual—simply talking to customers on a regular basis. But this will not scale, with multiple channels and thousands or potentially millions of customer transactions to monitor. Sooner or later, decreed by fate, you will lose contact with your customers and maybe develop products that no longer resonate, because you no longer understand their wants.

Fact is, insights are hard to find, particularly during times of rapid product iteration and company growth. Complex and unstructured source data, with multiple topics running through each thread in a huge range of emotional tones, will cloud the issue many times over.

As the world becomes even more digitally connected, the challenges are increasing. Online feedback forms and surveys have grown into community chatter, video calls, messaging, online forums and more. Consumers and corporations use a wild mix of Facebook, WhatsApp, Slack, Zoom, Zendesk and many more for customer communication. Someone, somewhere, still uses Second Life to promote their brand.

Manual processes struggle, to say the least, to handle this digital complexity, variety and volume. Embedding an effective VoC program based on “listen, learn, act, improve” requires efficient and scalable processes – which leads inexorably to Programmatic VoC.

A programmatic system can detect and surface actionable Voice of Customer insights from thousands of data streams, without bias (and without sleep!) using AI. AI captures and understands customer sentiment in real time, and analyzes the trends that could influence product strategy and development—and ultimately, growth.

Caravel offers an all-in-one hub for customer feedback powered by AI technology to bring your Voice of Customer program to life. By excluding personal subjectivity, Programmatic VoC powered by Caravel generates consistent results with a transparent process. As your data volumes rises, you can rely on Caravel’s AI to accurately analyze your responses and point you in the right direction.

For companies setting a product-led growth strategy, Programmatic Voice of Customer is vital to success, and AI is leading the way.

If you're looking to learn more, stay in the Loop because in the coming weeks, we're diving deeper into how exactly to execute each of these steps! If you want free consultation from our VoC experts, reach out here! We'd be delighted to work with you to build your VoC program.