Feedback analysis involves identifying the needs and frustrations of customers, so that businesses can improve customer satisfaction and reduce churn. It's often done automatically, enabling companies to sort huge amounts of data from various channels in a timely and accurate way.
Imagine the following situation. You receive a new batch of NPS survey scores (like in the image below), along with open-ended answers that explain these customer scores.
The problem with open-ended answers, however, is that they’re hard to measure. Raw data, such as customer feedback, needs to be structured before it can be ‘measured’ or analyzed. One way to structure this data is by reading every open-ended response and tagging them.
But what if you have thousands of NPS scores to analyze and new batches every week?
In this article, we’ll present you with an answer to your customer feedback analysis problems. What is customer feedback analysis? Why is it important? How can you gain actionable insights? And most importantly, how can artificial intelligence help? Read from start to finish, or skip to one of the sections below:
Let’s hop to it!
We all know what customer feedback analysis is, right? We just flip the words around and it’s the process of analyzing customer feedback… If only it were that simple.
Customer feedback analysis happens on a micro and macro scale, or in two stages if you like. First, teams need to analyze individual pieces of incoming customer feedback and determine their category or tag. Then, this ‘goldmine’ of grouped data is analyzed again to discover insights about a company’s strengths and weaknesses.
Let’s explain in more detail.
Stage one; customer feedback analysis doesn’t just involve tagging a survey response or review with the first appropriate tag in the list. There might be more than one tag you can use. You might even need more information before you can lump your golden nugget of feedback into a category. You might have all the information you need to tag your feedback, but you’re unsure if you’re applying the correct tag, so you ask your colleague next to you what they think, and they either confirm your feedback tagging suspicions, or they disagree leaving you second-guessing. It’s a convoluted process, but a necessary one to ensure that your customer feedback arrives in the best possible state for stage two.
Now that you have all your customer feedback correctly tagged in one tidy Excel sheet, you can begin the second stage of analysis. An overall view of what all this freshly tagged customer feedback is trying to tell you and your business, otherwise known as insights.
We’ll explain how to analyze large quantities of customer feedback using data analysis tools, later on in the article. First, let’s take a look at why customer feedback analysis is important.
Let’s begin with all that feedback you’ve collected over the years. 80% of 18-34 year olds have written online reviews – compared to just 41% of consumers over 55, which means you’re probably receiving more data than ever before. Let’s say you have a reasonable amount of customer feedback to analyze but, after going through all your feedback, realize you don’t have any takeaways to share with your team. 40% of marketers say that the insights delivered by customer intelligence teams are not actionable. Where are businesses going wrong?
Considering the vast amounts of feedback organizations receive every day, they often find themselves overwhelmed with information and struggling to transform it into actionable insights because they don’t have the tools or the right people to turn their data into actionable insights. In other words, they’re ‘data rich but insights poor’. While 74% of companies say they want to be data-driven, only 29% are good at connecting analytics to action.
Imagine, all those years running customer surveys and collecting data and not one single actionable insight. Perhaps product, customer experience, and sales teams have been receiving poor data reports that don’t provide value or, worse, lead them to make the wrong decisions. Maybe you’re focusing on the wrong area, or your tagging system is a mess. Well, you might have a Herculean task ahead of you – sorting through all that collected customer feedback and tagging it, in an attempt to make sense of it all.
But why go to the trouble of sorting through customer feedback and tagging it? The simple answer is, to understand your customer. If they’re talking about you and your competitors constantly, you’ll need to stay on top of these conversations because their opinions can help ensure that your end product or service will actually meet their expectations, solve their problems and fulfill their needs. Tagging feedback also helps every team, from customer support to product development, make more informed decisions. So instead of second-guessing what might frustrate or satisfy customers, you can support or dismiss theories based on cold hard facts.
Take Netflix for example. It uses its massive user base to gather data, make better decisions, and create accurate algorithms. As part of the onboarding process, Netflix asks new users to rate their interest in movie genres and rate any movies they’ve already seen, so that they can recommend new movies and TV shows to users. Their data-based recommendation algorithm has resulted in an impressive 75% of viewer activity based on these suggestions.
So, it’s clear that by listening to your customers and using the information that’s right in front, you can gain valuable data insights that will help your business grow.
When it comes to customer feedback, not all of it will be useful to your organization. In fact, we can split customer feedback into two types of data: insightful and non-insightful.
Non-insightful data is information you already knew was an issue, whereas insightful data is information that contradicts your knowledge, confirms your suspicions, or quantifies the importance.
For example, if the feedback analysis for a SaaS reveals that 90% of customers are unhappy with the usability of their app, this is insightful data that they can take action on, in other words, actionable data or insights.
There are three types of actionable insights you can get from customer feedback analysis:
So now that you know the difference between insightful and non-insightful data, let’s look at how to get it!
Before starting any kind of feedback analysis, you’ll need to find quality customer feedback. You’ll also need to have a strategy in place to sort that feedback, and then you’ll need to make sense of it so that everyone in your organization can understand and benefit from the insights obtained from your customer feedback:
First, you’ll need to go out into the wide world of data and proactively search for customer feedback. It’s no good just waiting around for the next email to drop into your inbox; there are many sources out there that provide valuable customer feedback:
Sending out NPS surveys is one of the best ways to discover how customers perceive your product or service because it consists of one simple question, for example, ‘How likely are you to recommend MonkeyLearn?’, followed by a 0-10 scale. Respondents are grouped as Promoters (score 9-10), Passives Passives (score 7-8) or Detractors (score 0-6) and your Net Promoter Score is calculated by subtracting the percentage of Detractors from the percentage of Promoters.
First, you’ll need to gather the NPS responses you want to analyze by exporting or downloading the information as a CSV or Excel file. Or, you can use integrations from popular tools such as Zapier or Google Sheets. Promoter.io has also developed an intuitive interface for exporting feedback data and other tools like Delighted allow you to select the data range you want to analyze and export it as a CSV file.
NPS surveys deliver both scores and open-ended responses. However, to understand the granularities – why customers are happy/dissatisfied, what they’re talking about and why – you’ll need to focus on the open-ended responses (text data). While an NPS score gives you a quick overview of how your company is performing, it doesn’t provide actionable insights.
The best part is that these tools are easy to tailor. Customer surveys can be customized according to your defined target audience, and there are various ways in which you can send them, including via email, Facebook, Twitter, and by embedding a link on a web page, etc.
Moreover, SurveyMonkey also conveniently provides tools to create custom reports and charts that can be shared directly with your teams.
Would you ever book a hotel or buy a new computer without checking out the reviews first? If you said yes, then you’re far from alone. In fact, you’re a minuscule proportion of the 90% of consumers that read online reviews before visiting a business. Here’s a review that Lenovo posted about a gaming laptop via Twitter:
Now check out some of the responses:
Customers are reading reviews and feedback from other users to decide which products to buy.
They’re also leaving feedback about other aspects, in this case, Lenovo's delivery service, which could dissuade customers from buying their products.
Reviews are everywhere and your brand’s name will probably crop up on sites like Capterra, G2 Crowd, Siftery, Yelp, Amazon, and Google Play, just to name a few. Reviews are a great source of feedback because customers tell you exactly what’s great or not great about your business in as many words as they like.
The hard part is gathering all this data in one place and making sense of it. Thanks to web scraping, however, you can collect customer feedback about your business, and that of your competitor(s) if you like, in mere minutes. Tools, such as Dexi.io, ParseHub, and Import.io offer easy-to-use interfaces that allow you to import all this data, and package it into a downloadable file. If you know how to code, you might opt to use open-source libraries such as Scrapy or Goutte to get these reviews automatically.
By analyzing customer feedback on social media, you can get real-time insights about your brand. Facebook, Instagram, and Twitter can be a fertile source of high-quality customer feedback allowing organizations to take instantaneous action on more urgent issues. Most brands nowadays have meaningful conversations with their customers on social media, allowing them to get to the root of certain problems, deal with issues swiftly and more effectively, and avoid a PR crisis.
When Mailchimp recently launched their new all-in-one marketing platform, there was some confusion and anger about pricing for existing subscribers, which Mailchimp responded to immediately:
Not only was Mailchimp able to keep on top of customer frustrations (urgent issues), and reply immediately, the email marketing platform received some insightful feedback about its pricing plans – they’re hard to understand. Which would suggest customer service needs to simplify their pricing plan documentation, so it’s easier to follow.
Earlier, we mentioned that some customer feedback is insightful while some is, well, pretty useless. That’s where good categorization can help. When grouping your customer feedback into predefined categories, you’ll come across the odd piece of feedback that doesn’t fit into your hierarchy of tags. That’s probably because it’s not very insightful.
Categorization also highlights some less noticeable features that, nonetheless, can play a significant role in the overall process, making the company perform better. Remember Blackberry? The mobile phone with a miniature keyboard that once dominated the U.S. The majority of big businesses and government organizations relied on BlackBerry’s superb security, reliable email, and utilitarian functionality. However, this wasn’t enough. Had the company realized that something as simple as a ‘bigger touchscreen display’ could have won them new customers, they may have adapted and still be in business today. Today, this is something that text analysis can detect, as well as influence decisions that make or break a product or service.
By thinking strategically about your business and customers, and establishing a consistent and controlled structure for categorizing feedback, you’ll be able to get valuable insights and detect trends over time.
For example, companies deal with customers in various ways. While some support teams try to be friendly and conversational on social media, others use a more professional tone. So how can you decide the best way to talk to your customers? You could compare your business tweets with those of your competitors’ by analyzing and classifying language used by customer service teams and customers to help define your tone of voice. To find out which approach customers of telecommunication companies prefer, we categorized 200,000+ customer tweets with Verizon, T-Mobile, AT&T, and Sprint using sentiment analysis. First, we analyzed how each company interacts – T-Mobile has a friendlier and more personal approach, while Verizon’s tweets are very formal. Then, we performed sentiment analysis on the data. The results showed that a friendlier take on social media received more positive responses:
Categorization also helps improve your workflow because employees will have a better idea of how to tag pieces of customer feedback. If teammates know exactly what each tag means and which types of customer feedback go with each tag, then the tagging process is going to be faster and more accurate.
Once you’ve categorized your customer feedback using your chosen list of tags, you’ll need to summarize the results and share them with the wider team, so that they can take action. We’ll go into more detail later, on how to summarize and share results in a way that’s easy for everyone to understand.
Now, let’s look at how to categorize all that customer feedback you’ve gathered.
By categorizing data, you’re structuring data. Why structure feedback, though? Well, to start with, it’s a lot easier to analyze, and means you can gain more accurate insights based on information that’s relevant to your business.
However, there’s some pretty hefty legwork involved since there’s no one-size-fits-all solution for categorization. Before defining a way to tag or categorize feedback, you’ll need to ask yourself what you hope to achieve from your customer feedback analysis, in other words, what are your business objectives?
Perhaps you want to know which features of a product need improving, or you might be interested in how customers perceive your customer support or delivery service. Maybe you’ve noticed a sharp fall in sales, and want to get to the bottom of it.
Here are the steps you’ll need to follow before defining a way to categorize feedback effectively:
Once you’ve followed these steps, you’ll be ready to define a categorization system!
A good place to start is with a pen and paper. While reading through customer feedback, try to spot recurring patterns and themes and jot them down to get a rough overview of ‘what’ your customers are talking about – for example, Usability, Pricing or Reliability. Then look for sub-categories or ‘why’ your customers are talking about these aspects – for example Navigation, Expensive or Bugs.
How Other Companies Are Structuring Their Feedback?
To help you understand how you can categorize your customer feedback, let’s take a look at how other businesses are going about it.
HubSpot recommends splitting customer feedback into three main categories:
Below each of these main categories come the sub-categories. Product teams, for example, would be interested in feedback about Bugs and Feature Requests, while Customer service would be more interested in feedback about Accounts and Billing. Marketing and sales teams, on the other hand, would focus on feedback about Pricing and Features.
Sounds straightforward, but you’ll need to think strategically. Atlassian used to have many tags for many things. They even had one-off categories that were used once and never again. Imagine being faced with a long list of tags, as well as a long list of emails, survey responses or social media messages to trawl through; it’s highly unlikely that you’ll tag every piece of customer feedback correctly. And if you’re not confident that feedback has been correctly tagged, then you can’t trust the insights you receive after analyzing this feedback, in which case it becomes a pointless task.
So what did Atlassian do to turn their customer feedback into valuable insights? They came up with a new categorization system that focused on their product. First, they came up with a set of criteria for their tags. They needed to be scalable (long-lasting), consistent (applicable to multiple products), specific (explicit about the origin of issue) and accurate (appropriately mapped to each issue).
Then they whittled down categories to just three – Reliability, Usability, and Functionality (RUF), allowing them to quickly pinpoint what customers were complaining about and why. Like HubSpot, Atlassian also uses sub-categories:
Main tag: Reliability
Sub-tags: Performance, Bugs.
Main tag: Usability
Sub-tags: Complexity, Content, Navigation.
Main tag: Functionality
Sub-tags: Tracking, Collaboration, Content Management.
Intercom uses a similar approach by focusing on primary, secondary tags, and closing tags. Their primary tags describe the kinds of conversations customers are having with their teams, covering everything from Questions and Feature Requests, to Bug Issues and Negative/Positive Feedback. To help teams understand customer feedback tags, Intercom created a table outlining each tag – a very useful practice when defining your own tags so that there’s not too much crossover:
Obviously, some customer feedback is going to need more than one tag, such as a review for a new feature that talks about a bug issue in a negative way. But by limiting the number of tags used for categorizing feedback, teams can get a clearer picture of issues and deal with them more efficiently. Secondary tags describe the topic or feature in question, such as Pricing and Integrations, and closing tags, categorize customer feedback that’s been Resolved. Try out this categorization method for yourself using Intercom’s useful tag cheat sheet.
Here at MonkeyLearn, for example, we provide integrations within our software, and they’re a big deal because they allow everyone to easily connect everyday apps to our machine learning software. Since integrations are a core part of our business, we use Integrations as a main category and the specific integrations, such as Google Sheets, Zapier, and Zendesk, as secondary tags.
Another way that organizations might define a list of tags is by performing keyword extraction on their historical data. This allows you to quickly find out which themes your customers are talking about most often. Using a keyword extraction model, you can input text, and a list of keywords will appear, like in the example below:
Above, we added a negative review about Slack to the keyword extraction tool. We can quickly see which aspects the customer is unhappy about. So, how would we define tags in this case? Well, ‘smaller fees’, ‘paid plan’ and ‘free version’ would call for a more general tag like Price:
Here’s another example:
This is a review about Slack’s Reliability, which we can deduce from keywords such as ‘glitch’, ‘issues’, ‘update’.
If you run keyword extraction on all of your feedback collected through the years, you will discover useful insights that will help you define the categorization structure for your feedback.
Above, we’ve mentioned a few examples of how companies tag their customer feedback, and you might be starting to get your head around the main category, subcategory partnership. So, let’s throw a spanner in the works by delving deeper into factors that could influence your category list:
Each organization will have a list of topics or themes that are more important to them than others and might split their feedback into topics. That way they can focus primarily on the feedback that mentions their topic(s) of interest (for example, Customer Service, Ease of Use, Pricing).
Businesses might want to distinguish types of customer feedback that appear more frequently. For example, if you wanted to focus on types of negative feedback, you could create tags for Complaint, Improvement, and even filter this type of feedback by topics of features, so you know exactly what you need to improve.
Often, customer feedback mentions specific features about a product or service, and the features that appear more often will be the features that organizations want to focus on, whether positive or negative feedback. However, if you wanted to focus on negative reviews, you could first run a sentiment analysis on your customer feedback, followed by topic analysis, otherwise known as aspect-based sentiment analysis. In our case, we’d use the tags Text Classification, Text Extraction, Integrations, Batch Processing, since these are important features of MonkeyLearn’s platform.
Some organizations might decide to classify feedback by severity, prioritizing more urgent issues over less important issues – using High Priority, Medium Priority, and Low Priority as their tags. This is a simple way to deal with customer feedback, but it’s crucial that teams understand what classifies as urgent and what classifies as low priority, which is why outlining these tags, like in the example above, is important to denote different severity levels. For example, feedback from the CEO of an organization or a bug issue affecting numerous customers might be deemed as high priority.
Businesses that offer premium services might categorize feedback by paying and non-paying customers, in tag terms – Subscribers and Non-subscribers. These categories are also likely to inform issue priority, with premium members’ requests being dealt with first.
Other companies might classify and route customer feedback depending on the channel. Perhaps they have different agents that deal with different lines of communication – social media, on the phone, live chat, etc.
Organizations offering different products often have different teams specializing in each product. In this case, customer feedback needs to be categorized, first and foremost, by product. For example, Google might have a tag for Google Smartphone, Gmail, Google Store, and so on.
While these are all great ways in which you can categorize your feedback, you’ll find that some work better for you than others. But before you decide to try out all of the variants above, at the same time, take a moment to think about the last time you did three or more things at once. Tricky, right? Well, categorization is similar, if you try to categorize using too many variants, then your results will be confusing. That’s why we recommend choosing one or two criteria to shape how you categorize feedback. Then, once you’ve got the hang of it, you can gradually add more.
Now that you know how to categorize customer feedback, you can put it into practice. However, first you’ll need to find a way to scale your customer feedback analysis, and that’s where artificial intelligence can be your knight in shining armor.
Up until now, we’ve focused on manual categorization of customer feedback but, realistically, manually sorting through thousands of pieces of text data every day is impossible. Just take a look at Domo’s Data Never Sleeps 5.0 report to get an idea of the vast quantities of information that’s generated every minute of the day. And the numbers keep growing, which means your customer feedback is likely to follow suit.
Now, you might be thinking that you don’t need to read all your customer feedback. Well, actually, you do. You might not use all of it but you do have to take into account every customer’s qualms. Why? Because, for every customer who complains, it’s likely that there are 26 customers who don’t say anything. Customers need to feel like they're being listened to; if they don’t, they’ll do business elsewhere – bad news considering that reducing churn by just 5% can increase profits by 25-125%.
So, now that you’re faced with the daunting task of analyzing thousands of pieces of customer feedback, let’s introduce you to auto-tagging with machine learning. At MonkeyLearn, we use two machine learning models to auto tag and analyze text – classifiers and extractors. They’re very simple to use and can help tag your customer feedback in seconds. If you want to build your own models and tailor them to your business, you’ll need to spend time training your models first. This can take time but the results are worth the effort. On the other hand, if you want to try out our machine learning tools before building your own, we offer some ready-to-use machine learning models, including sentiment and aspect classifiers.
How do they work? Through machine learning. You teach models how to differentiate texts, first by creating tags and then by tagging a certain number of texts manually. Once you’ve tagged enough examples, machine learning models start to notice patterns and make their own predictions using your predefined tags. And there you have it – the process of auto-tagging your customer feedback with machine learning.
To help you further understand how machine learning works, let’s take a closer look at how to build and train AI models we mentioned above – sentiment and aspect analysis.
Before you even start tagging your feedback and training your AI model, it’s helpful to split text into smaller parts depending on length and complexity. These parts are called opinion units and usually contain multiple opinions, for example, this review about Google Sheets:
I enjoy google sheets because I can access it from anywhere and update it from anywhere. But for my family budget, I still prefer Excel, especially with the ease of conditional formatting.
Here, we have multiple aspects (ease of access and formatting) and sentiments. Using our opinion units extractor model, you can separate your feedback into smaller units, otherwise known as opinion units:
Machine models that have been trained to detect opinion units are much more precise because it’s notoriously hard, even for humans, to analyze free-text responses. Let’s face it, it’s easier to understand a simple sentence like, ‘I enjoy google sheets because I can access it from anywhere’, which has one sentiment, than a more complex text that has multiple sentiments. Even humans struggle to classify sentences with more than one sentiment!
Once you’ve broken down text, making it more manageable for machines to resolve, you’re ready to analyze your customer feedback, in this case, using aspect and sentiment analysis models:
You can either use MonkeyLearn’s pre-trained models or you can create your own. If you choose to create your own, here’s a comprehensive guide on how to create your own classifier. We’re just going to outline some of the more important steps below:
Once you’ve read some of your feedback, you’ll be able to reel off a list of tags in seconds! In this example for SaaS products, we’ve used Pricing, Features, Ease of Use, and Customer Support.
Alternatively, you can use our Keyword Extractor to help you detect the most frequently talked about topics in your customer feedback to define your initial tag list. We used the same example above in the keyword extractor below. Take a look at the keywords that were extracted, and how you’d use them to define your own set of keywords:
Pricing is definitely a key theme, as well as Features and Licences.
For a sentiment analysis model, coming up with tags is a lot easier – Negative, Positive / Good, Bad:
For all classifiers, we recommend using a maximum of 10 tags to start with and making sure you have enough pieces of text per tag – around four for a basic machine learning model.
You can either upload your customer feedback as a CSV or Excel file, or use one of the available integrations to import data to train your models.
You’ll need to manually tag customer feedback to help your machine learning model understand what each tag means. For example, sentiment classifiers need to be able to distinguish the difference between a Positive and Negative comment by understanding words and phrases such as ‘issues’, ‘complex’. and ‘easy-to-use’. While aspect classifiers using the tags Pricing, Bugs, and Ease of Use will need to understand words and phrases like ‘expensive’, ‘glitch in the software’, and ‘navigation’. Once a machine has learned how to differentiate texts, it will be able to process and analyze them.
Once your machine learning model(s) have learned how to automatically process language and deliver insights, you’ll need to test your model by uploading a new batch of customer feedback. You can do this in three ways:
a) Manually uploading a CSV or Excel file to classify data in a batch:
b) Using one of the third-party integrations that connect seamlessly with MonkeyLearn, including Google Sheets, Zapier, and Zendesk.
Once you’re able to upload large datasets for analysis, you’ll need to be able to easily interpret the results from this data in a way that’s fast, effective, and easy to understand.
The big advantages of using AI to analyze customer feedback are that it’s faster, more accurate and more scalable than manual tagging. And these factors have many knock-on effects such as more actionable insights, faster responses and resolutions, better customer experiences, lower churn rates, and higher profit margins, to name a few. Let’s take a look at the direct advantages in more detail:
AI-powered customer feedback analysis enables customer support agents to focus on more important tasks, whether that’s delivering more personalized responses to customers or creating reports that can be shared with stakeholders. It also means that you can analyze all customer feedback in seconds rather than in weeks or months.
Let’s take a look at how auto-tagging customer feedback helped Retently scale their processes and gain actionable insights. They wanted to find out what was driving their NPS score. But, instead of manually sorting through thousands of pieces of feedback, they opted for machine learning and trained a classifier to sort their open-ended NPS responses Once Retently had finished training their machine learning model, they were able to discover actionable insights in next to no time, helping them influence strategic decisions and provide a better user experience:
We all know the drill. Monotonously scrolling through survey responses, reviews, social media comments, tagging them, resolving them, maybe skipping a few, taking a coffee break, back to scrolling – then before you know it it’s time to sign off and you’ve only got through 200 hundred more pieces of feedback. This is where machine learning can help. Once you’ve trained your machine learning model, you can let it run automatically in real-time, 24/7. Leaving teammates to focus on the results and deal with urgent issues immediately.
You’ve defined your tags, you’ve trained your machine learning model, and all that’s left to do is obtain the results. Analyzing customer feedback is simple if you’ve put the time and effort into establishing the correct criteria, and because machine learning models use the same criteria to analyze every single piece of customer feedback, you can rest assured that your results are going to be consistent, more accurate and 100% actionable.
Let’s imagine the following scenario. You’ve asked your customer experience team to share insights for a new app release because you want to present initial customer reactions to the app, to your company’s shareholders. ‘No problem’, says teammate X, ‘I’ll just ping over the analysis we ran on all customer feedback mentioning the app’.
Five minutes later, you receive a CSV file with hundreds of rows that something that looks like this:
This is probably not what you had in mind. Yes, you can scroll through the results and see which topics of the new app are mentioned in a positive or negative light, but it’s hard to establish an overall perception of each topic.
Luckily, there are data visualization tools that you can use to transform your customer feedback into a comprehensive work of art. When asking your teams to send over reports in the future, imagine receiving results that look something like this:
A visual chart that everyone across the company can understand and use to influence decisions.
Google Data Studio
With Google’s visualization tool you can combine your aspect classification and sentiment analysis results to create digestible graphs and reports. And to make things even easier, you can connect directly to your file source, whether it’s Google Sheets, a CSV file or an Excel file.
Once you’ve imported your customer feedback, you’ll be able to visualize the results in charts and graphs of your choice. Now, you’ll have something that you can share with shareholders and the wider team. For a more in-depth explanation on how to use Google Data Studio, watch these tutorials.
Like Google Data Studio, Looker also provides easy-to-use integrations and a smart looking dashboard that’s easy to navigate. However, its standout feature is its zoom function. By selecting filters, you can get a more detailed view of the areas that you want to focus on, and zoom in to get a more detailed analysis of this particular topic. Want to learn more about Looker? Check out their Youtube channel.
Featuring an intuitive drag-and-drop interface to build graphs and charts, Tableau also offers integrations with a wide variety of big data platforms, so can connect to data in real-time. You can also interact with your results by highlighting sections that you want to transform into charts.
Customer feedback analysis isn’t as straightforward as it may seem. Especially when it comes to qualitative data (unstructured text). Randomly picking tags and banging out a machine learning model in five minutes is not going to deliver the insights you’re after.
We’re all familiar with the expression ‘haste makes waste’, and this definitely rings true of customer feedback analysis. It’s a process that takes time to implement and requires strategic thinking since there’s no one-size-fits-all strategy that can be applied to every organization.
Atlassian experienced the disadvantages of a convoluted list of tags that didn’t accurately represent the issues their customers were talking about. As a result, they missed out on important insights and effectively wasted a year incorrectly sorting customer feedback while receiving non-insightful data. However, they did learn from their errors and transformed their customer feedback analysis process into a tour de force.
By carefully establishing rules for categorizing your customer feedback, and combining your process with well-trained machine learning models, you’ll be able to get the most out of your customer feedback in the fastest way possible. The results mean that you can deliver the product and/or service that your customers expect.
Check out MonkeyLearn for yourself, and see how it can help sort your customer feedback. Once you’ve got the hang of using our software, create your own models using your own tags, to get the most valuable insights out of your customer feedback. If you need help getting started, request a demo and our team will guide you on how to work with your customer feedback, to uncover the insights you need.
June 3rd, 2019