Analyzing data in a spreadsheet can be a nightmare. It’s time-consuming and monotonous.
Carrying out a customer survey, for example, can be useful to obtain crucial insights into the overall customer experience of your clients. But the data obtained from these surveys can be incredibly difficult to process, even after you’ve added all the results to a spreadsheet and especially if you receive a high volume of responses.
How do you process this information, then? Should you read the answers one by one? What if you want to know what people are saying about your brand on social media?
It’s inefficient for a person to read rows and rows of comments and survey responses in a spreadsheet, and impossible to do it in real-time. Customer experience teams would take forever to read them, and by the time they’ve finished more mentions and responses would appear! In addition, this is tedious work and boredom can lead to inconsistencies. To solve this problem, text analysis with AI is here to help.
If you have responses to customer surveys, social media comments, or product reviews in Google Sheets, you can automatically analyze this data in that very same spreadsheet. Not only is it easy to get started with text analysis for Google Sheets, but it will also save you precious time and resources.
In this guide, we are going to take a look at how you can carry out text analysis in Google Sheets using AI, including:
- What is text analysis with AI?
- How can it be useful?
- Text classification
- Text extraction
- How to use text analysis with AI in Google Sheets?
- Create your own machine learning model
- Use cases and applications
Let’s get going!
What is Text Analysis with AI?
In text analysis powered by artificial intelligence, machines are able to sort large amounts of data using extraction and classification. The first model will extract important information from data in a spreadsheet, while a classifier will sort the data into categories. Some examples of texts that a company might want to analyze in a spreadsheet are Facebook posts, emails, product reviews, NPS responses, and more. If human beings were to carry out this process, they would find the task boring, repetitive, and time-consuming.
In addition, manually processing data is expensive for businesses, especially for large ones. How much time (and money!) would a company like Tinder have to spend, to analyze the hundreds of reviews they get every day? The answer: a lot.
Thanks to text analysis, businesses can carry out analysis tasks much more efficiently. They’re also able to gain valuable insights that human agents might miss.
Now, for the algorithms to be able to derive meaning from the text, you have to train them. But don’t fret! It’s not difficult at all. The most crucial thing you’ll need is tagged samples of data. Once the machine is trained with enough examples, it can start associating data and differentiating pieces of text from one another.
How is this process useful for a company? Well, this technique comes in handy when trying to figure out the sentiment behind a product review, the language of a certain email, or if an issue needs an urgent answer. Let’s take a look at some examples in which AI helps examine text within spreadsheets.
How can Text Analysis Be Useful?
Every day, your company receives emails, social media comments, answers to surveys, product reviews, and more. All this data is crucial for your business, and it’s all in the form of text. You might store that data in Google Sheets; after all, it’s a great way to organize information. Now, when a person wants to analyze these texts in Google Sheets, it’s likely to be a slow and tedious process (especially if the volume of data is high). On top of that, human agents might fail to miss valuable insights.
Text analysis, on the other hand, can analyze data in Google Sheets in seconds, and gain interesting insights that can help a business make data-based decisions. Let’s take a look at some more benefits of text analysis in Google Sheets:
- It’s scalable. To manually analyze data in Google Sheets, a person would have to read each row and try to make sense of all the information. It could take hours or even days, and the task becomes especially tedious if it has to be done every day! Text analysis can scale down these hours to minutes, or even seconds, so you and your team can spend time on more fulfilling tasks.
- Analyzes in real-time. The digital world, and social media in particular moves at the speed of light. Imagine that you have just compiled all the tweets that mention your brand into a Google Sheet. All of a sudden, you realize that there are a lot of users complaining about the slow speed of your site. They are furious because it’s affecting their business. By the time you have finished reading all of these comments, it might be too late.
It’s claimed that you have less than 60 minutes to prevent negative comments from becoming a crisis. Is there something you can do about it? Sure! Text analysis can detect these issues in real-time by spotting topics such as Problem, Bug, and Slow and classifying the sentiment of comments. So, for example, comments that are classified as Negative may also be classified as Urgent, and get sent to the front of the ticket queue. In this way, you can start responding to these negative comments in no time, and prevent a small issue from becoming a nightmare.
- It’s consistent. People naturally have different opinions, feelings, and points of view. Let’s imagine your team needs to tag product reviews in a spreadsheet to understand if they are positive or negative. One person may think that a comment such as “I absolutely loved the camera, though the pricing was a bit high” is a great review and tag it as Positive. However, another member of your team may see it as Negative because of the comment about the pricing.
Moreover, those opinions might change over time! The people tagging feedback might have felt tired, bored, or even hesitant at the time. They might have even made a mistake because of the tedious nature of the task. That’s only human. Of course, this means that the insights your company is getting are not as accurate as they could be.
How can you prevent this from happening? Text analysis can keep your criteria constant. When you train a model, you teach it to follow certain criteria that will never change. The model will never feel tired, bored, or hesitant.
After analyzing how text analysis is useful for Google Sheets, it’s time to take a closer look at the different types of text analysis that you can run within Google Sheets, depending on your needs and requirements.
Text classification refers to tagging a text (or pieces of it) by category, including topic, sentiment, and language.
Conducting topic analysis means to tag texts depending on their theme. If you are taking a look at social media comments, for example, you might read one that says “I enjoy how simple the dashboard is”. Using topic analysis, this comment would be tagged as Ease of Use. If you receive a comment that claims, “The company takes too long to answer”, it would be classified as Customer Support.
Why would you want to carry out topic detection? Picture this scene: you want to carry out an NPS survey to understand how satisfied your clients are with your software. You gather all this data in a Google Sheet and run a topic classification. After sorting your data into categories, you’ll be able to see which topics your customers mention most, for example, UX, Bugs, Pricing, and Integrations.
You can try out this pre-trained text analysis model that classifies responses for SaaS products by Customer Support, Ease of Use, Pricing, and Features.
If you are monitoring social media and want to know how users feel about your business, you no longer need to read each comment one by one. Automatically tagging texts based on the feelings of the people who wrote them is now possible, and it’s called sentiment analysis.
Sentiment analysis is faster and more accurate than going through a Google Sheet and manually tagging hundreds of negative, positive, and neutral comments. See for yourself! This pre-trained model will automatically categorize your texts into Positive, Negative, and Neutral.
If you combine this technique with topic analysis, you’ll get even more insightful information. Let’s imagine that you are launching a new product and want to know how your customers feel about it. You can conduct a survey, gather all their responses in a Google Sheet, and run the sentiment analysis model plus the topic analysis one. Voilá! These models will give you accurate information about your clients’ opinions about your new product, and also about the topics they mention in their responses. This way you can understand, for example, if users love the UI but hate the pricing of the new product.
Today, it is very important to truly understand what your customers want to stay ahead of the game. It’s certainly not easy to get (and process!) this information. Yet, even if machine learning cannot give you the precise answer you are looking for, it can lead you in the right direction when finding out what your clients desire.
For example, you might get a lot of emails from potential clients every month. Some of them may be asking for more information about your product or service. You can compile all these emails into a Google Sheet, and then apply this pre-trained model to identify those messages and tag them as Interested. This way, you can follow up on emails from customers that have shown interest and close the deal. Other tags this pre-trained model use include Unsubscribe (which lets you know that someone wants to stop receiving your emails) and Email Bounce (which suggests that an email bounced due to a typo in the email address).
If your team receives product reviews or survey responses in different languages, it’s possible that you have some routing issues when deciding who to send them to. Perhaps your English-speaking team starts reading comments gathered in Google Sheets and realizes they’re all in German, so sends them to the right team. Between sending the comments from one agent to the next and working out which team is best equipped to understand them, a lot of time has already been wasted!
Even though detecting a language and deciding which team needs to deal with an issue is straightforward, it takes time that could be better spent elsewhere. Luckily, language detection can tag a message with the appropriate language e.g. Russian, French, Spanish. This way, the next time you spot a review in your spreadsheet written in Russian, it will be automatically directed to the team that speaks this language. This pre-trained model can recognize 49 languages, so you might want to check it out!
Different from text classification, text extraction refers to the process of extracting information that is already within the text; for example, keywords, prices, countries, features, and more.
Let’s imagine you have a batch of NPS responses in Google Sheets. Running an extractor can be a great way of summarizing the content of these responses by extracting the main keywords within them. Some examples of text extraction include keyword extraction and entity extraction. Below, let’s take a look at how extractors can be used in Google Sheets.
This type of extractor will identify and obtain the most crucial word, or set of words, from a text. These keywords act as a sort of summary of the piece.
For example, you may have a Google Sheet with product reviews to analyze. If you want to figure out the main topic each review talks about, then keyword extraction can give you this information in just a few seconds. When analyzing a batch of reviews in a spreadsheet, you’ll get the main keywords for each review in a new column. Try out this pre-trained keyword extractor to get a feel for it.
If you want to obtain the name of a person, a company, a brand, or a location, you can do it with an entity extractor. Imagine that you want to know how many times your competitor is being mentioned in media outlets. What you can do is to put all those mentions in Google Sheets, and then run it through a company extractor. You’ll get the results in no time with named entity recognition.
How to Use Text Analysis with AI in Google Sheets
If you want to get started ASAP, you can use MonkeyLearn! We are a machine learning platform for text analysis using AI that will allow you to get valuable insights from raw text.
To start analyzing the data in your Google Sheets, you can use one of our pre-trained models. As they have already been trained, you just have to run this model to analyze your data in Google Sheets, and AI will do the rest!
Follow these steps:
1- Install MonkeyLearn’s add-on for Google Sheets
MonkeyLearn will need your permission to run:
And ask for the following permissions:
Finally, you have to connect the MonkeyLearn API to your Google Sheets. First, go to ‘Add-ons’. Then, click on MonkeyLearn —> Start. You’ll see a menu to the right. There, you’ll have to set up the API key. Click on the link, which will take you to MonkeyLearn’s page. Just copy and paste the API key, and you are ready to go!
2- Decide on a Model to Use
On MonkeyLearn’s dashboard, you will find different options: a sentiment analysis model, a keyword extractor, an urgency detection model, and more. Choose the one that suits your needs. In this guide, we are going to show you how to use the keyword extractor in detail, but the process is the same for all of the models.
In MonkeyLearn’s add-on for Google Sheets, you’ll see a window on the right-hand side of the screen where you’ll be able to select the model you want to use. In this case, we chose the Keyword Extractor:
3- Run the Analysis and See the Results!
The only thing left to do to analyze data automatically in Google Sheets is to specify the Column or Range of the pieces you want to examine with the model. In this case, we have reviews that go from cell B2 to cell B7. So, you need to add B2: B7 in the Column or Range field:
After this, just click ‘Run’ and that’s it! The model will take a look at all our product reviews and give us the keywords in column C:
And that’s it! Quite simple, right? You can follow the same steps for any of our pre-trained text analysis models available in your dashboard.
Create Your Own Machine Learning Model
You may use pre-trained models, like the ones above, to analyze your data in Google Sheets. But you also have the option of adapting models to the specific requirements of your business. For example, you can add categories to the sentiment analysis model to see if your customers are angry or happy. You can also train your model to recognize industry jargon. By using a custom model, you’ll get more accurate insights than if you used a pre-trained model.
1- Create Your Own Model
Sign in to your account and access the dashboard. Then, click on the button ‘Create model’:
This time, let’s create a classifier:
2- Upload Your Texts
In this step, you’ll need to upload the information you want to analyze. You can upload information in various formats, including CSV and Excel files:
3- Choose Your Tags
Now it’s time to choose your tags. In this case, we went for Ease of Use, Customer Service, and Pricing. These are the tags that the model will later on use to categorize data and make predictions. Keep in mind that if you decide to use many tags, you will need further work to train the model:
4- Start Training Your Classifier
It’s time to train your model so it can recognize the information you want to classify. For that, you will have to manually tag some of the pieces of text you uploaded, as shown below:
5- Test it
After tagging a certain number of samples, you’ll be ready to test your model. You have to give your model a name, and click ‘Test’:
Then, type something in the box and see how it works:
In this case, it worked very well!
6- Use the Machine Learning Model in Google Sheets
If you want to use this custom model in Google Sheets, then you need to follow certain steps. As shown below, you have to access the sheet with the information you want to analyze. Then, you have to go to ‘Add-ons’ in the toolbar and choose your custom model.
Finally, decide the columns and rows you want to analyze and your model will show you the results!
If your model is not as accurate as you’d like, don’t worry, you can go back and train it by tagging more data to make it smarter.
Want to do the same with an extractor? Follow this guide.
Use Cases and Applications
You now have all the info you need to start analyzing texts in Google Sheets and save countless hours of tedious work. Now, how is this helpful for your business? Well, it depends on your goals and objectives. Read on to see how artificial intelligence for Google Sheets is helping many businesses today.
Get Insights into Customer Feedback
Public reviews are becoming increasingly important. They are crucial for clients (more than 80% of buyers report they would never buy something before reading comments online), and for brands, too, as they give incredible insights into the overall customer experience.
Let’s picture this situation: you want to know how much your clients like your new product, so you ask them to fill in a survey. You end up getting thousands of reviews and compiling them into a spreadsheet… and that’s far more than you expected. Now you have to derive meaning out of this pile of reviews sitting in your Google Sheets. Luckily, text analysis can help you sort them quickly and easily.
You can use this sentiment analysis model to understand if customers liked your product or not. For example, we ran a sentiment analysis model to analyze TripAdvisor reviews and figure out how people felt about hotels in different cities around the world. As you can see from the following graphic, London received the worst reviews when compared to other major cities:
Understand Your Brand Perception
Today, there is a lot of information online about what people think about your brand. You can read tweets, Facebook reviews, blog articles, Amazon reviews, comments in the Google Play Store: the possibilities are endless. Text analysis can help figure out how your brand is perceived online.
In 2016, the UK voted to abandon the European Union, a process that would be known as Brexit. At the time, we wanted to know what people were saying about this historical event. That’s why we gathered more than 250,000 tweets in English and analyzed them with our sentiment analysis model. This helped us find out if the tweets were in favor, against, or neutral about Brexit.
We also wanted to quickly understand which aspects people were discussing, so we analyzed the tweets with a keyword extraction model. This way, we obtained the main words and phrases that people use when referring to this event.
Not surprisingly, the opinion was hugely divided:
Those in favor used the following expressions:
Those against Brexit used these expressions:
Applying these text classification and extraction models was quite simple. You can do the same with your business, to analyze what people on social media are saying about your brand, make informed decisions based on that information, and step up your game.
Do Valuable Market Research
Doing market research is a crucial activity for any business. Would you like to know more about your target audience? Is spotting a new business opportunity one of your objectives? It’s particularly useful to get insights on problematic areas and to identify unaddressed customer needs.
One way of applying machine learning to this process is to examine product reviews, social media comments, articles, or survey responses that you’ve collected in a spreadsheet. Using an extractor for keyword analysis, for example, can save you a lot of time and effort when taking a look at these texts. It’s a lot faster than reading every review yourself and you won’t have to spend hours creating filters or complex formulas in spreadsheets to examine your data. Instead, this keyword extractor will organize all the information with simple, easy-to-read tags.
When running such a model, you may discover certain topics your target clients are discussing, and come up with a great idea to solve their issues or continue doing what you’re doing if they’re mentioning topics in a positive light.
As you can see, Slacks perform well in a lot of aspects (such as Ease of Use, Integrations or Groups). However, people complain a lot about their Notifications, Pricing, and Performance-Quality-Reliability. This analysis gives the Slack team crucial information on aspects to improve their service.
Outselling competitors is also possible when applying AI to your Google Sheets. You may realize that many customers are complaining about a certain feature in your competitor’s product, an insight that you can use to your advantage by improving that particular feature in your own product.
We all know that analyzing texts in a spreadsheet can be a hassle if you have large volumes of data to work with. It’s tedious, boring, and expensive. When somebody buys a product, they will often write a review, which leads to new information. Now imagine that you’re receiving hundreds or even thousands of those reviews.
The good news is that your business can take advantage of all the insights contained in these reviews, emails, and survey responses. However, to be able to analyze them efficiently, you’ll need to implement text analysis with AI.
It’s very easy to analyze information in Google Sheets by integrating MonkeyLearn. Forget about manually tagging rows of text – text analysis is scalable, works in real-time, and always follows the same criteria, making it faster, more efficient, and accurate than humans.
The only thing you have to do is to choose a suitable pre-trained model to process your texts. Or, you can train your own model, which is just as easy. Even if it takes a little bit more time to train your custom model than using a ready-made one, it will give you results that are tailored to your business.
Don’t hesitate to ask for a demo. Our team will help you to carry out text analysis in Google Sheets and show you how to use this technology in your company.