It’s said that 95% of new products fail, and less than 3% become “highly successful.” Product feedback is a critical part of the product development process. The goal is to improve a product and the user experience, and product feedback can help you do just that.
Product testing is an integral part of product development, but it’s time-consuming and inefficient to manually sift through hundreds or even thousands of pages of feedback to extract insights for your product innovation teams. Traditional, code-based AI models can partially automate this process, but they take teams of technical experts, like data scientists and engineers, to get to production.
What if there was a way to take any type of real-life feedback—from “I wish you would add more features” to “your product is pricey, but it’s worth the cost”—and automatically assign it a positive or negative sentiment analysis score, or even classify the type of feedback?
What if this scoring happened instantaneously across an entire corpus of user-generated feedback, so you could quickly identify themes and insights that matter most for your product innovation teams?
That’s exactly what Akkio can do for you. Using no-code NLP (natural language processing) techniques, our platform can analyze any kind of text input—whether it’s Amazon reviews, comments on Product Hunt, or even tweets about your company—and determine whether each piece of feedback is positive or negative in tone.
And because you can do this at scale across thousands or millions of pieces of user-generated text all at once, you get massive speedups in terms of how quickly you can go through review rounds and incorporate valuable feedback into your products.
To get started, simply sign up to Akkio for free. As with any machine learning task, the first step is getting historical data and picking a column to predict.
For demonstration purposes, we’ll use the Amazon Reviews: Unlocked Mobile Phones dataset from Kaggle. The historical data is more than 400,000 reviews from Amazon's unlocked mobile phone category. The file includes a target variable called “Rating.”
We can use machine learning to predict a review’s sentiment, and then prioritize feedback with negative sentiment for analysis by product innovation and improvement teams.
Below, we’ll explore how to connect this CSV file to Akkio. First, we simply create a new flow, and hit “Table.” You can drag-and-drop the CSV, or simply select it from your computer.
After the dataset is connected, Akkio will automatically figure out the data types of each column and show you some details, like the number of rows and columns. You’ll also get a scrollable preview of the dataset.
Now we can click on the second step in the AI flow, which is “Predict.” Under “predict fields,” you can select the column to predict, named “Rating.” Then just hit “Create Predictive Model,” and you are done. In our example below, we have quickly created a model with sufficient training data.
Once the model is created, you can get a quick overview of your new predictive model.
When you click “See Model Report,” it will open up a new tab in your browser with an interactive report which allows you to quickly understand what was predicted for each field.
The model report will highlight the following:
You’ll be able to see specific details like the accuracy, model type selected (in this case, Deep Neural Network with Attention was used), and so on. The best way to improve model accuracy is to get additional, high-quality training data. With Akkio, you can easily merge on a new dataset, such as with another source of feedback data.
You can also try increasing training time, which may help increase accuracy, particularly if you have a large dataset to begin with. In this case, we can see that our model did particularly well with predicting 5-star and 1-star reviews, which are good for finding out what to double-down on, and what to revamp, respectively.
Now that we have our classification model trained on some historical data, let’s deploy it. The final step in the flow is to select where your prediction should be made.
In this example above, we’ve built a model to predict the sentiment of customer feedback, but we could just as easily build a model and deploy it on product testing feedback, tweets, or really any other text source. Let’s say that we wanted to predict the sentiment of tweets, and notify our teams about negative tweets.
First, we’ll want to select Zapier as the output step. We can then set up Akkio in Zapier, where we can deploy model predictions in thousands of different tools.
Our Zapier flow would look like the below, where we’d search for mentions of our brand, product, or service on Twitter, and send those mentions to our predictive model. We can then use a filter action to send only negative feedback as a Slack message to our product innovation team.
Ultimately, it takes just a few minutes to build and deploy an AI model in the real world.
Using NLP techniques such as sentiment analysis has become increasingly important when it comes to product feedback. By utilizing NLP, you can analyze how users feel when they interact with certain features and functions of your product.
The stakes are high for any company that relies on delivering a great customer experience every time, and manual analysis doesn't cut it when you're dealing with a high volume of product feedback.
With Akkio, teams can scale globally without needing to build or maintain any code or infrastructure themselves. As a result, deployments take seconds instead of days like other machine learning platforms.