Customer feedback is the best way to find out what your customers think about your product or service.
Whether it’s in the form of survey responses, social media mentions, product reviews, or chats with your customer support team, listening to the voice of your customers should be a central part of any business strategy.
Analyzing customer feedback provides valuable insights for product improvement, and leads to a better understanding of your customer’s needs and expectations. And when you monitor feedback on a regular basis, you can measure customer satisfaction over time to see if your business is going in the right direction.
In this guide, you’ll find everything you need to start gathering, analyzing, and visualizing customer feedback. We'll explain how you can use machine learning to simplify the tagging process, and how machine learning software like MonkeyLearn can help.
Read along, save it for later, or jump to the sections of your interest:
Let’s dive right in!
Gathering, analyzing, and acting on customer feedback is key for making data-driven decisions about your products and services, enhance customer experience, and improve KPIs. In this section, we'll cover the basics about working with customer feedback so you can set up an effective process that drives results to your business:
Customer feedback refers to the information and opinions provided by clients about a product or service. You can get customer feedback from different sources: surveys, social media, product reviews, chat interactions with your customer support team, among others.
For companies, understanding what their clients think is key for providing the best customer experience possible. By analyzing feedback, you can find out which aspects of your business are working well and which ones require improvement, and use those insights to make data-driven decisions that align your product or service to your customers’ needs.
Collecting and analyzing customer feedback has a number of benefits: it can help you measure customer satisfaction, improve your product or service, show your customers that you care about their opinions, and much more. Let’s go through some of them in more detail:
Reaching out for customer feedback consistently is the best way to measure customer satisfaction. We all know how it goes: the happier your clients are, the more likely they'll stay loyal to your brand and help you attract new clients through recommendations. In fact, 77% of customers say they would recommend a company to a friend after having a good experience.
Analyzing what your customers say about your business can help you understand what drives customer satisfaction. Often, customer service plays a very important role in this. According to a Microsoft report, 56% of customers say they have stopped doing business with a brand due to a poor customer service experience.
When it comes to measuring customer satisfaction, Net Promoter Score (NPS) is by far the most popular method. But we’ll dive deeper into this later.
Customer feedback comes directly from the people who are using your products and services. Therefore, it can provide many insights that help you improve particular features and the overall customer experience.
At the same time, customer insights can help you understand client behaviour by showing you how they interact with your product or service. What features do they find most useful? What are the main issues they have while interacting with your product? Are they suggesting new functionalities?
Making improvements based on customer feedback allows you to create products and services that really solve your clients’ needs.
Customers want to be heard. In fact, 77% of customers have a more favorable view of brands that ask for and accept customer feedback.
By asking your clients for feedback, you are letting them know that you care about their opinions and ideas. This helps build trust and strengthens their relationship with your brand.
However, you need to make sure that you take action on that feedback. If a customer complains about a flaw in your system, for instance, you should try to fix the issue as soon as possible, and send a follow-up message to let them know that you’ve listened and taken care of the issue.
For a company, it costs 5 times more to acquire new customers than it does to keep the existing ones. This proves that focusing on customer retention is an effective and profitable business strategy. Customer feedback is a great way to find out which aspects of your business are making your customers happy or unhappy.
Perhaps you are taking too long to respond to customer issues, or you’re not solving their problems effectively. By analyzing your customer feedback, you can prevent customer churn, as it enables you to take action on the aspects that your clients are most frustrated with.
By gathering and analyzing both quantitative and qualitative feedback, you can get actionable insights and use them to make smart data-driven decisions. For example, you can decide which areas require more investment, validate a business strategy (or not), or innovate based on customer requests (develop a new feature, for example).
Whether you are doing marketing research, following new trends within your industry, or trying to find new areas for growth, examining customer feedback can help.
There’s a lot of public feedback available: just think of product review websites, forums, or social media channels. You can easily take a look at what customers are saying about your competitors, identify new business opportunities, and find out more about your clients’ preferences and practices.
Listening to the voice of the customer (VoC) is becoming the cornerstone for long-term business growth. To take full advantage of its potential, it’s key to develop a customer feedback process that involves consistently collecting feedback, analyzing it, and transforming it into actionable insights.
Want to create a successful customer feedback loop? These are the main stages you need to consider:
1. Define your objectives: the first step before gathering feedback is to define what type of data you want to obtain. Do you want to get feedback related to your customer service? Or find out what customers think about your product and identify what feature needs to be improved?
2. Gather feedback: based on your goals, you will need to decide on the best method to gather feedback. You might want to send out an NPS survey, collect product reviews from review sites, or gather your social media posts.
3. Organize that feedback: once you’ve collected your feedback, it’s time to make sense of all that information. Whether you have quantitative or qualitative data, you will need to set up a system to categorize the data and gain meaningful insights.
4. Act on the feedback: finally, it’s time to make the most out of your customer feedback results. You can share the results with different teams within your company using reports or data visualizations, and collectively decide what actions to take based on the data. You can also follow-up with customers who shared feedback.
As previously mentioned, the first step to creating a systematic customer feedback process involves defining a series of concrete objectives that you want to accomplish. The method you choose for collecting feedback will largely depend on these goals, so make sure to devote some time to this task.
Here are some questions you may ask yourself when defining the goals for your customer feedback process:
What do you want to know from your customers? You may want to measure customer satisfaction, analyze customer service experience, detect product issues, or research new market trends, for example.
Are you interested in a specific part of the customer journey? Perhaps you want your customers’ opinions on your latest feature, or you’d like to find out the reason why most of them abandon your shopping cart before completing the purchase.
What do you plan to do with the feedback you collect? The insights you get from customer feedback are only useful when you translate them into action. Is your company prepared to invest resources on improving some aspects of your business? Are the aspects that you want to analyze valuable and profitable for the business?
Once you’ve defined your goals, you will be ready for the next step: collecting customer feedback.
Customer feedback allows you to obtain relevant insights on what your customers think about your product and how they relate to your company.
This information empowers your business in a myriad of ways: you can measure customer satisfaction (and see how it evolves over time), find opportunities to improve your product or service, which in turn helps with customer retention, and make data-driven decisions to take your business to the next level.
By ‘listening’ to your customer’s feedback, you are showing customers that you value their opinions. To be able to listen to them effectively, however, you’ll need to define objectives. Once your goals are clear, you can move forward and implement a process that involves collecting, analyzing, and sharing feedback in a consistent way.
In the following sections, we’ll delve deeper into each of the stages of the customer feedback process, including gathering feedback, organizing it and putting it into action. So keep reading to learn more about the steps you need to follow.
So, you’ve defined a set of goals for your customer feedback process and you’ve got an idea of the type of data you’d like to obtain. Now, it’s time to decide on the most appropriate method to collect customer feedback.
There are four common reasons why businesses may want to ask for customer feedback:
This section will go through each of these scenarios and provide an overview of the most relevant methods you could use to collect feedback. Then, we’ll provide some useful tips and best practices to keep in mind when gathering feedback. Finally, we'll share a few examples of questions and surveys that you may find useful to collect feedback from customers.
When you measure customer satisfaction, you get an accurate image of how happy your customers are with your product or service. Periodically checking this metric allows you to understand (at least in part) how your business’ reputation evolves over time, and what are the most common issues mentioned by your customers.
There are three basic surveys to measure customer satisfaction. Let’s take a look at each of them:
This is one of the simplest and most popular approaches to measure customer satisfaction. Basically, NPS surveys consists of asking customers to score a product or service on a scale of 0 to 10 based on how likely they are to recommend it to a friend or colleague. This rating allows you to classify customers as Promoters (9-10), Passives (7-8) or Detractors (6 or less). The ‘Net Promoter Score’ is the result of subtracting the percentage of Detractors from the percentage of Promoters.
However, while the score may indicate how high or low customers rate your business, it doesn’t provide any extra information about the customer’s opinions and experiences with your product or service. That’s why a follow-up, open-ended question is often included, inquiring about the reasons for customer scores. These responses can deliver valuable insights on what drives customer loyalty, what causes customer churn, and various other pain points.
The advantages of NPS surveys are that they’re extremely simple to understand and require little effort from customers. That’s why they generally get high response rates. Quantifying results from the closed-ended question is also easy: you just need a spreadsheet, and you can immediately identify clusters of customers. Many businesses consider NPS a key metric, as it’s strongly related to customer loyalty and business growth.
However, scores can be ambiguous and a 6 or an 8 may not mean the same for every customer. An NPS score alone doesn’t offer enough information: it is necessary to dig deeper with a follow-up question. But categorizing the follow-up responses may be difficult unless you’re using machine learning tools.
This customer satisfaction survey asks customers to rank their level of satisfaction after a specific interaction with a company. You can find CSAT questions after you make a purchase (‘how would you rate the recent product experience?’) or at the end of an informative article (‘how helpful was this article?’), among other situations.
The Customer Satisfaction Score allows you to measure customer satisfaction at different points of the customer journey, and track how effective you were in making your customers happy.
CSAT surveys are simple to understand, short, and intuitive, therefore you’re likely to receive high response rates. You can personalize the type of score (stars, emojis, a numerical scale, etc) and the data is easy to analyze.
However, satisfaction is subjective and scores might translate differently depending on the country you’re in. It may also be difficult to act upon customer feedback because, although a CSAT score can tell you if a customer is dissatisfied, it won’t deliver insights on the specific reasons.
This Key Performance Indicator (KPI) measures the amount of effort required for a client to make a purchase, use a product or service, or solve an issue (‘Overall, how easy was it to solve your problem with [company] today?’). A high effort experience translates into unhappy customers, while a low effort experience means that customers are more likely to repurchase.
Though it’s quite new, CES is becoming one of the main metrics for customer success teams to gauge customer loyalty. Making things as easy as possible for customers is one of the main drivers for customer satisfaction, hence the importance of measuring the ease of a customer’s experience.
This metric allows you to detect flaws at different stages of the customer journey and focus your efforts where they are needed. However, you only get information related to a very specific part of your business. So, the best way to use this metric would be by combining it with others to get a broader view of what your customers think about your product or service.
Collecting customer feedback is one of the best ways of assessing the performance of your customer service team and the quality and success of the support your company is providing. There are different strategic situations where you can leverage customer feedback. Let’s take a look at some of the methods you can use:
This consists of sending your customers a short online survey immediately after they’ve interacted with your customer service team. By analyzing the results, you can evaluate your customer support representatives – their ability to listen, response times, and how friendly and helpful they were. At the same time, you can find out if customers were satisfied with how the issue was handled, and if they have any suggestions to make the experience better.
Online surveys are a simple and quick way of getting impressions from your customers right after an interaction with your customer service. You can customize the surveys in order to ask the questions that really matter to you. Be careful, though, not to spam your customers with online surveys every day. Survey fatigue may lead to low response rates. Also, if the survey is too long, customers may quit before finishing it. Finally, make sure that you formulate questions properly; otherwise you may not get insightful answers.
Sending a follow-up email after a purchase is an excellent opportunity to gather feedback about the overall customer experience and how satisfied the client is with the product or service. By creating an additional touchpoint with the customer, you can improve customer loyalty and encourage future purchases. This can be particularly profitable for industries with a strong focus on service, like retail, automotive, real estate, or hospitality.
Besides sending the standard ‘order-confirmation email’, you can send a post purchase follow-up email to ask customers to rank their experience leave a product review, or ask what you can do better to improve the experience.
Sending follow-up emails is cost-effective, it can be automated, and the results are measurable (you can analyze open rates, click-through rates, etc). It’s an opportunity to thank your customers and nurture your relationship with them post-transaction, and a great way to get new product reviews. Also, by recommending other products and offering discounts, you can actively encourage your clients to repurchase. Again, you have to be careful not to irritate customers by bombarding them with too many emails, and you need to make sure that you’re targeting the right people. Annoyed or incorrectly targeted customers may ignore your emails, or even unsubscribe if they don’t like the content. Lastly, people often skim read emails, so make sure your content is concise and appealing in order to catch their attention.
Customer churn is one of the main concerns of any business, and customer service often plays an important role in this.
When a customer drops out your shopping cart at the final stage, cancels a paid subscription, or ends a subscription before the trial period ends, sending a short survey asking why they canceled can provide you with valuable insights.
For SaaS businesses, for example, you might realize that you need to offer a more accessible plan (or a free version of your product). Perhaps, allowing customers to pause their account or extend their trial period, as well as offering free access to educational resources or use cases, might also help retain and win new customers.
Customer cancellation surveys provide you with actionable feedback: are your customers leaving due to pricing? Do they require a feature that you are not currently offering? Or do they have an issue that your customer support team was unable to solve? Knowing the reasons why they want to cancel allow you to make improvements and reduce customer churn. The survey can also serve as an opportunity to win back clients by offering them other alternatives.
You’ll need to plan customer cancellation surveys carefully, to ensure high response rates and accurate answers, making sure that you include a wide spectrum of answers that reflect all the possible reasons behind a customer’s decision.
You can send an NPS survey to your customers after an interaction with a customer support rep, asking them how likely they are to recommend your product or service. The results can give you an overview of how satisfied your clients are with the customer service you are providing. Companies with low NPS scores are often associated with bad customer service.
The great thing about NPS surveys is that they’re simple, quick, and easy to understand and usually get a high response rate.
On the downside, the score by itself doesn’t provide information about why customers leave a high or low score, leaving you with few insights to improve your customer service. This is why analyzing the follow-up open-ended question on your NPS surveys is a great way of getting more in-depth insights.
Collecting and analyzing online reviews from different sources (from review websites like Capterra, G2 Crowd or Google App Store, to specialized forums or even your Facebook business page) can provide valuable insights on what your customers think about your customer service.
Online reviews are honest opinions that reflect customer experiences and offer detailed explanations to support their statements, allowing you to understand exactly what customers like or dislike about your service.
Qualitative feedback in online reviews allows you to get more actionable insights. You can understand the reasons behind a client’s opinion and the situation that motivated them to reach out to your customer service team. However, since reviews are unstructured, they are harder to analyze. Text data needs to be organized in a certain way before you can detect patterns and classify opinions, and sorting them manually can be extremely time-consuming and unproductive. In this case, you’d need to automate your text analysis process, something we’ll cover in later on in this guide.
Analyzing customer feedback allows you to understand how your customers use your product, what their favorite features are, and which aspects require improvement. By asking for feedback at different stages of the customer journey, you have a better chance of detecting if there’s something wrong with your product and can take action before it turns into a bigger problem.
These are some ways in which you can gather feedback to detect product issues:
As previously mentioned, NPS responses measure customer satisfaction by quantifying how likely customers are to recommend a product or service.
Customers that give high ratings are grouped as Promoters. Usually, high NPS scores are related to high quality, reliable, and easy-to-use products. At the opposite end of the scale, customers that give poor ratings are grouped as Detractors. Product flaws, bad UX design, or shipping issues can be some of the reasons behind a negative evaluation.
It’s key to analyze the response to the open-ended question to get more in-depth information on what the product issue is (if there is one), because the NPS score alone won’t tell you. Sometimes, you may need to reach out to each of the Detractors manually to find out the reasons for their score.
Keeping track of social media mentions (and analyzing them periodically) can shed some light on the challenges and pain points of your business. What are your customers’ most frequent complaints? What are the most urgent issues to fix? Listening to your customers on social media can alert you of product issues and allow you to handle them before they escalate.
Social media mentions are spontaneous, so provide you with instant and honest feedback without having to ask for it, but you will need the right tools to collect and analyze them in real-time.
If you have a product app, asking for feedback inside the app is a great way of getting to know what your users expect from your product. It’s a fact: understanding your customers is the key to creating outstanding products. You can ask them about their favorite features, inquire if the app is making things easier for them at any point, and encourage them to suggest features they would like to see.
In-app surveys can be as flexible as you want: you can tailor questions to your product or service (maybe you’ve just launched a new feature and you’d like to know how it’s working, or set priorities for your product roadmap). You can also decide which users get to see the survey (those who’ve recently made a specific action within the app, for example), and when the survey will be displayed.
Keep your survey short, otherwise customers will drop off. Again, this type of survey provide qualitative data, so you’ll need the right tools to process and analyze the results.
Implementing a feature request board is a great idea for a customer-centric product strategy. You can ask your customers which features they would like to see, analyze the data and add some of the suggestions to your product roadmap.
It makes your customers feel that you value their opinions and hear their voice. By building a product collaboratively, you give your customers an active role, increasing trust and loyalty, and it also enables you to prioritize the requests of those users you care about most. The only issue with feature request boards is that they usually receive a low volume of responses, and often from heavy-users.
Analyzing online reviews can provide valuable insights for product improvement. You can learn what customers like or dislike about your product, and find out about more pressing issues.
They also help you understand how customers actually use your product, and the problems they may encounter along the way. Often, companies have large amounts of product reviews to analyze, and it’s time-consuming to go through them all manually. That’s why it’s important to have the right tools in place, before processing and analyzing this information.
Comment boxes, strategically placed at the bottom of a page can help you identify product issues. Simply asking your customers for feedback with questions like ‘how can we make this page better for you?’, can lead you to insightful data, especially when it comes to minor issues that wouldn’t be reported otherwise.
It’s a very simple way of allowing customers to share feedback when something on your page isn’t working right. Don’t worry, they won’t interfere with the customer experience, and you can change the location of the comment boxes on your page to see which placements generate higher response rates. Since comments are in the form of text data, you’ll need to equip teams with the right tools to handle this information.
There are plenty of public sources of data which you can monitor to gain insights about your competition, as well as follow new market trends.
Here are two ways of collecting this type of feedback:
Listening to global conversations on social media can be a powerful tool for market research. You can analyze what’s being said about your competitors, create a more detailed profile of your ideal customers (their interests, what they talk about, their opinions towards certain subjects) and follow the discussions around certain topics or themes related to your industry.
This is a cost-effective way of getting valuable insights from potential customers, by listening to what they talk about without any initial communication. You can uncover trends around certain topics and use that data to improve the way you advertise your products. You can find gaps in the market and understand what’s working (and what’s not working) for your competitors. While this is a cost-effective way to do research, you’ll need the right tools to gather this information and analyze it.
Online review websites such as Capterra or G2 Crowd provide access to millions of public reviews. By analyzing reviews from other companies within your industry, you can get valuable insights that can help you improve your product strategy.
For example, you may want to track reviews from your competitors to identify relevant keywords that will help you become more effective at marketing or targeting your product. You can also stay one step ahead by searching for the most common complaints in your industry, and fixing those aspects before they become a problem for you. Once again, you’ll need specific tools to analyze and classify large sets of reviews.
Once you’ve decided which is the most suitable method to collect customer feedback, it’s time to design the questions you are going to ask your clients.
Here’s some advice:
Now that you are familiar with the different methods you can use to collect customer feedback, let’s take a look at some examples of questions and surveys:
Basically, surveys can be divided into quantitative and qualitative. The former provides questions with options from which users choose a response, while the latter asks open-ended questions which encourage a more detailed answer.
NPS surveys are an example of quantitative survey questions. The standard question for NPS surveys is ‘How likely are you to recommend [product name] to a friend or colleague’. Customers need to give a score from 1 to 10, where 1 is ‘not at all likely’ and 10 is ‘extremely likely’.
Other examples of quantitative survey questions are:
For each of these questions, remember to provide a closed set of options.
Qualitative survey questions allow for a more in-depth response. Occasionally, quantitative survey questions are followed by an open-ended question inquiring about the reasons for a customer’s score: ‘What’s the primary reason for your score?’
Here are other examples of qualitative survey questions you could ask your customers:
Collecting customer feedback is the best way to get first-hand opinions about your product or service. NPS surveys, product reviews, social media monitoring, customer cancellation surveys: there are many different ways of listening to the Voice of Customer (VoC), and you’ll need to pick the one(s) that are most suitable for you and your goals.
In this section, we presented four different ways to gather customer feedback: by measuring customer satisfaction, analyzing customer service experience, detecting product issues, and researching new market trends.
We also listed some best practices that you should keep in mind when collecting customer feedback. It really isn’t that hard, especially when collecting quantitative feedback.
As we mentioned above, it’s harder to collect qualitative feedback if you don’t have the right tools. But, it’s worth investing in these tools because the insights you gain from open-ended responses are far more insightful than scaled responses.
In the next section, we’ll go into more detail about which tools can help you sort this type of data, as well as how to organize and categorize feedback effectively.
After collecting customer feedback, the next logical step is to organize and categorize the results. By turning unstructured data (which can be quantitative, qualitative or a mix of both) into meaningful and actionable information, you can understand what your customers are saying about your business and use those insights to make better business decisions.
Analyzing quantitative feedback is pretty simple (it’s all about the numbers!); the bigger challenge resides in finding the most effective way to analyze qualitative data, especially when you have to deal with large volumes of open-ended responses.
In this section, we’ll also explain how artificial intelligence can help you automate your feedback tagging process, and how you can analyze customer feedback using machine learning models to capture the subjective information present in each opinion.
Read along or jump to the sections of your interest:
Let’s get started!
Quantitative customer feedback can be gathered from surveys with closed-ended questions. These survey responses provide hard facts and show a broad overview of the customer’s opinions and motivations. Metrics like Net Promoter Score (NPS), Customer Effort Score (CES), or Customer Satisfaction (CSAT), are examples of quantitative feedback.
Quantitative data is expressed in a numerical way and reflects concrete and quantifiable aspects of a business, for example the data you collect might tell you the percentage of customers that would recommend your company, or how many clients gave low scores following their latest customer service interaction with you.
When it comes to analyzing closed-ended questionnaires, assigning a numerical value to each response (a process known as ‘coding’) allows you to turn that data into indicators, metrics, data tables, and graphs.
Based on these metrics, you can identify different groups of customers. NPS surveys, for example, often enable you to classify customers as Promoters (those who scored 9-10), Passives (scores between 7 and 8), and Detractors (respondents who scored less than 7).
Also, you can cross-tabulate data to look for different patterns between two or more variables. For example, you might notice that most of your Detractors mention negative experiences with your customer service. Cross-tabulation can easily be done using an Excel spreadsheet, the results of which you can export to a data visualization tool to present in a more attractive way.
You can get qualitative customer feedback from open-ended questions, product reviews, comment boxes, chat threads, and social media posts, among other sources. Qualitative feedback is subjective and tied to emotions, and it’s usually referred to as the ‘why’, because it helps you understand the reasons behind customers’ opinions and behaviors.
But before you can start examining qualitative customer feedback, it’s necessary to organize all open-ended responses by assigning categories to each piece of text. Only after classifying each piece of customer feedback into different categories, will you be able to identify and quantify relevant topics, detect patterns and trends, and uncover relevant insights from your data.
Categorizing feedback provides a sort of structure to the data, and makes it a lot easier to understand the main topics your customers are referring to.
However, there’s no standard way to tag feedback: the right solution will depend on your goals. What kind of insights would you like to get from your customer feedback analysis? Are there any specific problems that you would like to address?
A useful way to organize customer feedback is by splitting it into three main categories:
Then, within each of these main categories, come up with subcategories by scanning your data. For example, subcategories for product feedback could be UX/UI, Bugs or Feature Requests, while for customer service you could use tags that reflect common issues raised by users such as Account, Billing, or Usability.
If you are a software company, a handy way to organize product feedback is with RUF; a simple categorization system created by Atlassian that categorizes all incoming product feedback into 3 categories: Reliability, Usability, and Functionality.
Reliability: feedback about performance issues, bugs, downtimes, incidents, or things that aren’t working as expected are tagged with this category. For example, “X isn’t working”.
Usability: feedback related to how easy (or difficult) to use is the product, how easy it is to navigate, etc. For example, “Not sure how to do X”.
Functionality: feedback about specific features within the product. “I Wish X did Y”
Drift has also come up with their own categorization framework for customer feedback, which can be particularly interesting for startups. Basically, they consider that the way customers ask questions entails a certain type of issue. Therefore, they classify their feedback into three categories:
User Experience Issues: customers asking how to get things done with your product. This includes questions like “How do I customize this feature?” or “I tried to customize this feature and now everything is broken”.
Product Marketing Issues: questions related to what things your product can do. “Can I use your product for this?” or “How do you compare with this other product when it comes to doing this?”.
Positioning Issues: customers asking if your product is for them. Questions starting with “I’m probably not your target customer…” or “I’m sure I’m wrong but I thought your product might be able to help me solve this problem…”.
As we’ve already seen, organizing customer feedback under pre-defined criteria (tags) is key to understand what your data is about.
But let’s face it: not all the feedback you receive is equally meaningful or relevant to your goals. Having a solid categorization strategy allows you to put aside the pieces of feedback which are not aligned with your hierarchy of tags. At the same time, by tagging feedback, you can uncover insightful information that you may have overlooked in the first place.
Tagging can be quite a subjective task. That’s why having a consistent, clear and controlled tag structure is key. A good recommendation is sharing with your team a list of your tags and subtags along with a description of how to use each of them, so you can make sure that everyone is on the same page. Being consistent when using tags also benefits the areas of your company that have to act upon that feedback, like product, marketing or sales
Whether you are tagging customer feedback manually or using machine learning (we’ll come back to this later), here are a few best practices that you should keep in mind:
In order to define precise tags, you first need to understand what your customers are talking about. Read at least 30 pieces of feedback, and narrow down the most relevant topics, themes, or issues.
Create guidelines with descriptions for each of your tags, so that everyone in your team knows when to use each tag. For example, there may be two or more tags that teammates want to apply to one piece of feedback, but after reading the descriptions alongside each tag they may choose just the one tag.
Try to group similar tags into one category: it’s better than having a lot of similar tags that overlap. Always favor broad tags over ones that are too specific or niche. Why? When tagging manually, teammates end up using the most frequent tags and forgetting about the rest. If you are using a machine learning model, on the other hand, you need to provide training examples for every tag in the list, so it’s important to define tags that are representative of the majority of the feedback.
Tag customer feedback that is related to frequent issues or topics. There’s no need to tag comments that are unique to one customer.
Quality is better than quantity. We advise using a maximum of 15 tags, so that your team can tag consistently. Going through a long list of tags is not only time-consuming, but also results in inaccurate data.
Whether you are tagging your feedback manually or using machine learning to automate the process, it’s always best to follow a hierarchy by defining main tags and subtags. Machine learning models can make more accurate predictions if they have a solid structure to follow.
Using tags to categorize customer feedback is the best way to make sense of qualitative data.
However, doing this task manually can be both time-consuming and expensive. Also, you should consider that humans make mistakes, due to fatigue, distractions, pressure, and so on, and they choose tags based on their unique criteria, influenced by social, religious, and political ideas.
Let’s imagine you need to analyze over 1,000 survey responses to open-ended questions: categorizing each response manually would be hard because, one, by the time you’ve gone through each response it may be too late to retain customers who expressed negativity towards your product or service. Two, data might be time-sensitive and, after two or three weeks, irrelevant. And three, teammates will get tired, bored and frustrated, and end up tagging inconsistently just to get through responses.
Eventually, as your data grows, it’s necessary to automate your tagging process, which is exactly what Artificial Intelligence (AI) is equipped to do.
But how? Through machine learning, a subfield of AI that creates algorithms capable of understanding human language. Machine learning models can be trained to perform different tasks, such as classifying texts or extracting relevant data.
To automate the process of tagging customer feedback, you’ll need to create a text analysis model that’s able to automatically categorize your data. You’ll need to define which tags you want to use, using the best practices we outlined earlier on, and train your model with relevant examples. Once it has been fed enough data samples, your machine learning model will be able to start making its own predictions.
Keep reading about the advantages of using machine learning to automate the task of tagging customer feedback, and learn how you can get started by using an AI platform like MonkeyLearn.
Every minute, social media users post 474,000 tweets and 510,000 Facebook comments. With such an overwhelming amount of digital data, it comes as no surprise that companies need to find fast, accurate, and cost-effective ways to deal with information on a large scale.
Some of the advantages of using machine learning to automatically tag feedback like survey responses, chat threads, social media posts, and product reviews are:
Machine learning allows you to analyze large sets of data in a very short time. Imagine a task that would normally require weeks of manual processing and analysis, to be completed in just a few minutes.
One of the best things about machine learning is that your model can function 24/7! You can monitor what your customers are saying on social media in real-time, or analyze the latest product reviews as soon as they get published, allowing you to handle dissatisfied customers and take care of urgent issues right away.
Based on experiences, emotions, or beliefs, humans will apply different criteria to the same situations. The same applies when manually tagging customer feedback. One teammate might believe one set of product feedback is talking about UX, while another might think it’s focused on Delivery. Machines, on the other hand, always apply the same criteria, making results more reliable.
If you craft a well-structured tagging system you can get a wide variety of insights from your customers. Thanks to machine learning, you can instantly identify the main topics, sentiments, and even intents of each piece of feedback, providing you with a deeper understanding of how your customers feel about and interact with your brand.
When analyzing customer feedback, the first thing you need to find out is what your customers are talking about: which aspects of your business are they referring to? Ease of Use, Support, Features?
However, this information doesn’t tell you how your customers actually feel about the different aspects of your business.
Aspect-based sentiment analysis is a machine learning technique that identifies the main topics within a set of data, and detects the sentiment related to that aspect (sentiments can be positive, negative or neutral).
Let’s say a high number of your product reviews refer to Customer Service. They could be either good or bad. Thanks to aspect-based sentiment analysis you can automatically find out which of the opinions about your customer service are positive and which are negative.
MonkeyLearn can be used to analyze all sorts of customer feedback, from social media to online reviews. For the purpose of this tutorial, however, we'll focus on NPS surveys since they are one of the most effective ways to measure your company's successes, areas for improvement, and get critical feedback from your revenue sources.
The process of analyzing NPS surveys can be split into two different stages. The first part is quite easy: it all comes down to calculating the average Net Promoter Score by subtracting the percentage of Detractors from the percentage of Promoters.
The second part, however, presents more challenges as you’ll need to go through the customers’, open-ended responses. By running aspect-based sentiment analysis, you can easily be able to:
a. identify relevant topics (learn what your customers are talking about) and, b. extract sentiment out of your data (understand how customers feel about each of those topics).
Here are the main steps you should follow to create an aspect-based sentiment analysis model with MonkeyLearn and use it to analyze open-ended NPS responses:
Depending on which tool you use for gathering NPS responses, you’ll find options to export your data or download a CSV or an Excel file.
Reading several of your customer’s responses will give you an idea of the main topics they are mentioning. This will help you define the tags you will use to build your aspect classifier.
For example, this is the taxonomy that Retently created to tag their NPS feedback:
Customer feedback often contains more than one opinion related to different aspects of a business. Take this example:
“The user interface is nice and clean, but it has very poor (almost non-existent) reporting tools”
The customer expresses a positive opinion about the product’s interface (“nice and clean”), and a negative one concerning reporting tools (“very poor, almost non-existent”).
Dividing each piece of customer feedback into opinion units makes it more manageable for machines to assign one sentiment to each sentence. Also, it allows you to get more granular insights, as you’ll cover all the topics your customers are mentioning and how they feel about each one.
So, the first step to analyze customer feedback with machine learning is splitting your text into opinion units. What does this mean? Basically, dividing each piece of feedback ―in this case, open-ended NPS responses― into smaller fragments of text.
How can you do that automatically? You can try MonkeyLearn’s pre-built model: opinion unit extractor.
Once you’ve extracted all the opinion units from your NPS responses, it’s time to build your aspect-based sentiment analysis model. Actually, you’ll need to create two different models: a sentiment analysis classifier and an aspect classifier (also known as topic analysis). Then, you’ll combine both to analyze your feedback.
A sentiment analysis model can automatically identify and extract the opinions from a text. By using it to analyze customer feedback, you can understand how customers feel about a certain aspect of your business.
1. Choose a model type
2. Choose a type of classifier
Next, you'll need to select the type of classifier you want to build, click on ‘sentiment analysis’:
3. Upload your data
You can upload an Excel or CSV file with NPS responses, or upload data from online sources like Twitter, Gmail, Zendesk, Promoter, etc. This data will be used to train your machine learning model:
4. Start training your model
By tagging different examples manually, you’re training your machine learning model so that it can understand which opinions are considered positive, negative, or neutral:
The more data you tag, the smarter your machine learning model becomes. After tagging several examples, the model will start making its own predictions. You can correct the model if you notice that it’s not tagging samples correctly.
5. Test your classifier
By clicking on the ‘run’ tab, you can paste a piece of customer feedback and see how your sentiment analysis model classifies it:
You can also click on ‘build’/‘stats’ to check metrics of your classifier’s performance. By checking stats like precision and recall, you can understand how well your model is working, or if it still requires some more training. You can also take a look at the most frequent keywords for each tag. Check out our blog on understanding classifier stats.
6. Put the model to work
Now it’s time to check if your machine learning model is ready to analyze unseen data. There are three ways of analyzing new data automatically with your model:
Batch Processing: you can upload a CSV or an Excel file with new customer feedback. The classifier will analyze the data and give you back another file with an extra column indicating the sentiment for each piece of feedback.
Integrations: you can use some of the integrations available at MonkeyLearn to connect a data source (like Zapier, Google Sheets, Zendesk or Rapidminer) to your sentiment analysis model:
A topic analysis model (in this case, an aspect classifier) automatically identifies the recurrent themes or topics in a text. You can train an aspect classifier to analyze customer feedback and assign pre-defined tags or categories based on the topic of each piece of feedback.
Before creating an aspect classifier, you need to define the most appropriate tags that you will use. Here are the steps you need to follow to create a custom aspect classifier with MonkeyLearn. The process will be similar to creating a sentiment analysis classifier, but will have some subtle differences.
1. Choose a model type
2. Choose a type of classifier
Now its time to define the type of classifier you want to train. This time click on ‘topic classification’:
3. Upload your data
You can upload an Excel or CSV file with your NPS responses, or import them from an external source. This data will be used to start training your model:
4. Define your tags
As you are now creating a topics classifier, now you'll need to define the tags that your model will use. For this step, think of the main categories you’d like to use to classify your NPS responses. For example, you can use tags like Ease of Use, Pricing, Customer Support, or Product UX.
The example below shows how easy is to create tags, in this case to analyze a set of software reviews:
5. Start training your model
You need to tag different examples manually to show your aspect classifier how to understand each tag.
Make sure to have enough examples of each of the categories you’d like to use. Manually tagging as much data as you can will make your machine learning model more accurate.
6. Test your classifier
Click on the ‘run’ tab and type a review into the text box field to test your aspect classifier. If it needs more training, you can go back to ‘build’ and tag a few more examples.
7. Put the model to work
Using tags is the best way to categorize customer feedback, whether it’s quantitative or qualitative data. While you could tag each piece of data manually, it’s time-consuming, inconsistent, and expensive.
By using machine learning, you can automate the process of tagging customer feedback, and perform real-time and large-scale analysis in just seconds.
There are many advantages of using machine learning to automate tagging feedback, which we’ve outlined above. However, we can’t stress how important it is to create tags that align with your business goals, so that you can gain the most valuable insights from your data.
Once you’ve determined your tags, you can get started with aspect-based sentiment analysis right away. It will make analyzing customer feedback a breeze – in fact, just as easy as building your own machine learning models using integrations, or use MonkeyLearn!
In the following section, we’ll focus on how to act on the results of your customer feedback analysis. Let’s move forward!
Now that you have analyzed and categorized customer feedback, it's time to put the results into action. In this section, we'll cover:
Let's move forward!
Whether you analyze customer feedback manually or with machine learning, you’ll probably receive an Excel spreadsheet or a CSV file with rows and rows of results that are neither engaging nor easy to understand. Think of something like this:
Daunting, right? Luckily, data visualization tools can help you unleash all the power of your customer feedback data and take your presentations to the next level, as you can see in this image:
To help you find the best tool, we’ve listed a few of the best software you can use to turn your data into attractive and user-friendly reports:
Google Data Studio is a free data visualization tool that allows you to connect with different data sources, and turn your customer insights into custom interactive reports that you can easily share with your shareholders or colleagues.
With Looker, you can create data visualizations that update in real-time (directly from the data source). By using the platform’s pre-built pieces of code, called Looker Blocks, you can benefit from easy-to-use integrations and customized solutions. There are also several tutorials available to help you get started.
Tableau is an analytics and data visualization platform that enables you to analyze customer feedback in-depth, by using features such as filters, data breakdowns, and interactive charts. You can extract data from many different sources, and analyze large volumes of information without compromising the tool’s performance.
Customer feedback is only useful when it can be translated into concrete insights. So, once you’ve turned your results into easy-to-understand and eye-catching visualizations and reports, it’s key to:
a. Take time to analyze the results in order to discover key insights, and b. Share these insights with your teams.
Usually, the areas that benefit most from customer feedback are customer service, marketing and sales, and product.
Once your teams have had time to look at the results, they’ll need to define a clear strategy on how to act on that feedback. The most critical issues (for example, UX flaws and bugs that may affect your customer retention) should be addressed right away.
Then, you should come up with a list of priorities, by calculating the ROI of improvement (the impact it will have on your business), and the costs of implementing a given solution (considering your budget and resources), among other factors.
There are different ways in which customer feedback can drive business decisions. Listening to your customers can shed some light on existing features that require improvement (and new features that you can add to your roadmap), for example. Also, you can leverage customer feedback to redesign your marketing strategy. You might want to target more customers, implement strategies to retain customers at risk of churn or invest in more training for your customer support representatives.
Closing the loop of your customer feedback process involves following up with the customers that shared feedback.
Finalizing interactions with your customers, not only to thank them for completing a survey or writing a review but also to let them know about changes or improvements you made or that you are currently working on based on their feedback, shows them that you care about their opinion.
Addressing your customers’ feedback is also a great way of building trust and loyalty. There are several ways you can finalize interactions, from sending personalized emails to your customers, to publishing a report explaining how you implemented their feedback.
And, you should always follow-up with customers that have raised complaints or had an issue they needed to solve, as well as dissatisfied clients at risk of churn. Of course, you should respond to these clients as quickly as possible, and proactively reach out to disgruntled customers.
Listening to the voice of customer (VoC) is the best way to understand what your clients’ think and expect from your business. In this section, we’ll present some use cases of companies that are successfully leveraging customer feedback and using insights to make data-driven decisions.
Getting started with customer feedback can be very simple, especially if you take advantage of the many resources available. To help you out, we’ve listed some of the best tools and software you can use for gathering feedback, centralizing feedback from different sources, analyzing customer feedback, and visualizing the results.
Let’s dive right into it!
Customer feedback can help you understand what your customers need, validate ideas, find opportunities for improvement, test new features, and open a communication channel for your clients to make suggestions, ask questions, etc. Depending on your business goals, you can collect and analyze feedback for different purposes.
As the founder and CEO of Superhuman explains, asking their customers for feedback was the key for the company to achieve product/market fit. This is one of the main goals for a startup and basically means building a product that really creates value and solves customers' needs.
The team behind the email app simply asked their users the question, “how would you feel if you could no longer use Superhuman?”, and provided three options: “very disappointed”, “somewhat disappointed”, and “not disappointed”. They also inquired about the main benefits customers receive and how they could improve customer experience with the product. Then, they segmented users based on their answers and focused their efforts on understanding the types of people that were benefiting most from Superhuman.
For ticketing platform Eventbrite, creating an effective feedback loop with their early adopters and taking immediate action on their feedback was essential for the company’s success. Actively responding to customer feedback allowed them to understand their user’s needs and paved the way towards their definitive niche: the entrepreneurial community. According to the company CEO, observing how their customers use their product has become second nature to the company, and is essential when it comes to developing new features and keeping an eye on emerging trends.
There are several online tools that can help you at different stages of the customer feedback process. These are some of the most relevant:
SurveyMonkey: a free online survey tool with different templates and predetermined survey questions.
Typeform: a very easy to use survey tool to create customized online surveys, forms, polls and questionnaires.
Drift: use their live chat feature to have conversations with your site visitors and interact with them at the right moment of their customer experience. You can also create email marketing campaigns.
GetFeedback: design user-friendly mobile surveys with templates to measure customer satisfaction, customer effort score, and product experience, among other metrics.
Delighted: this survey software focuses on Net Promoter Score (NPS). You can customize surveys, add follow-up questions and send them via email, web, or SMS.
Promoter.io: another great tool for gathering actionable customer feedback using the NPS system. It provides an end-to-end solution for creating effective NPS surveys with higher response rates.
Product Board: a single ‘inbox’ where product managers can centralize customer feedback from different sources, like chat conversations, survey responses, feature requests, and support tickets.
Roadmap: a collaborative tool to unify customer feedback in one place. It has an ‘idea backlog’ to organize and prioritize ideas based on feedback, and share them with your internal stakeholders.
Artificial intelligence makes it possible to analyze large sets of customer feedback in just seconds. Even though you can build an AI solution from scratch using open-source libraries, it takes time and resources. SaaS APIs, on the other hand, are a great choice if you need to get started right away, with little (or no) coding, and just a few steps to get you up and running.
You can create AI-powered solutions to identify relevant topics in online conversation or determine sentiments in product reviews or social media posts, among other tasks. These are some of the smartest tools on the AI block:
As we’ve seen earlier, these are the most popular data visualization tools you can use to create killer reports and presentations out of your customer feedback results:
Having a consistent customer feedback strategy is key for companies to learn what their customers think about their product, but also to understand how their needs and expectations evolve over time. Regularly asking customers for feedback is also the best way of coming up with fresh ideas, detecting opportunities for improvement, and discovering new market trends.
The first step to set up a successful customer feedback loop is to define your objectives. Knowing on which aspects of your business you’d like to focus can help you figure out which type of data you’ll need to collect and decide on the best methods to gather feedback. Depending on your goals, you can send an NPS survey, check out online product reviews, or implement a feature request board, among many other alternatives.
Then you’ll need to turn raw data into actionable insights that you can share with the different areas of your business, so that they can make improvements based on that feedback. While analyzing quantitative feedback doesn’t entail much complexity, processing qualitative feedback (like open-ended survey responses) requires an extra effort: you need to categorize each piece of feedback, providing a sort of structure that enables you to find patterns, identify frequent topics, etc.
Analyzing customer feedback often implies dealing with large volumes of data from different sources, and spending hours reading and tagging each piece of data. However, thanks to AI and machine learning, it is possible to automate your customer feedback analysis, allowing you to obtain results in a very short time. Getting insights from customer surveys, reviews, social media posts, and chats has never been so easy!
Even though the concept of machine learning can seem a little hard to understand, getting started is actually very simple. With MonkeyLearn, you can easily create AI models to automatically analyze your customer feedback and get revealing insights about your product or service.
Want to give it a try? Just contact our team and request a personalized demo!
Automate business processes and save hours of manual data processing.