Customer churn, or users ceasing to use a product, destroys businesses every day. According to the CallMiner Churn Index 2020, US businesses lose nearly $140 billion a year due to avoidable customer switching.
If businesses could only predict what customers will churn, sales representatives could be alerted to take appropriate actions for those customers, such as creating engagement, renewal, and retention plans.
There are a wide range of factors that impact churn, such as a user’s tenure with the company, the services and add-ons they’ve selected, their contract type, or even their payment method. These factors are complex, and interrelated in non-linear ways, making it practically impossible to predict with traditional statistics, let alone the “gut feeling” that’s typically used.
Machine learning is the answer, which lets you comb through historical customer data to find patterns in churn, and predict what customers will churn next. With Akkio’s no-code machine learning, you can easily build and deploy churn models, and improve your bottom line.
In this case, our historical data is a Kaggle Dataset called “Telco Customer Churn.” This file has 7,043 rows of customers and 21 columns, or 20 features excluding the target column.
Because real customer data is highly sensitive, and protected by a number of data regulations like GDPR, this is a synthetic dataset generated by an IBM data science team.
The data includes a column called “Churn,” which is simply “Yes” or “No.” This is what we’re trying to predict.
This dataset is already uploaded to Akkio for demonstration purposes, which you can see on the homepage as “Churn Prediction Demo,” but you can upload any dataset you want by simply clicking “Upload Dataset.”
Now, we can click on the second step in the AI flow, which is “Predict.” Scroll all the way down under “predict fields,” and you can select the column to predict, named “Churn.”
Then just hit “Create Predictive Model,” and you’re done! In the churn prediction demo, the model has already been made, but it takes as little as 10 seconds to train a model from scratch. You can also select a longer training time—from 1 to 5 minutes—for potentially more accurate models. Keep in mind that longer training times aren’t always better, and may lead to overfitting
Also note that you don’t pay for model training time, unlike with many typical automated machine learning tools, so feel free to build as many models as you’d like.
After a model is created, you get a simple overview, which highlights the top fields, model prediction quality, sample predictions, and more.
We ended up with a very high-quality model, with a raw accuracy of just over 80%. Among 784 cases of no churn and 273 cases of churn, there were 134 false positives of “churn” and 75 false positives of “no churn.”
There’s also the option to “See Model Report,” which lets us easily collaborate with others and share these model details with anyone, even if they don’t have an Akkio account.
In short, we’ve discovered that it’s very easy to build a highly accurate model to predict customer churn.
Now that we’ve built a model, it’s time to deploy it in the real-world.
With Akkio, it becomes trivial to deploy complex machine learning models. The churn prediction demo shows deployment via Zapier as the third step, which is an easy way to serve predictions in a production setting. Zapier is a no-code automation tool that connects tools with “Zaps,” and you could use a Zap to send new customer transaction data to the Akkio model automatically.
Zapier allows you to link up thousands of different tools, so no matter where your customer data is coming from, it’s easy to connect it to Akkio to get a prediction, even if you don’t have any technical expertise.
If you’re looking for more technical power, you can also use Akkio’s API, which is formatted as a Curl command, allowing you to send a GET request that contains your flow key, API key, and input data, and get a prediction back.
Have you ever received a sudden discount or bonus from a company? This is common practice at a number of large firms when it’s predicted that you’ll churn. As of writing, Zapier, for instance, gives users who are likely to churn the option to continue their subscription for $10 a month for 3 months, instead of the usual $25/month.
Users have also received discounts from leading companies like Amazon’s Audible, Spotify, and many other SaaS companies working to reduce churn.
However, not all SaaS companies have the technical and financial resources to implement churn prediction the traditional way. Indeed, there are around 8 million businesses in the United States—according to the most conservative figures—and just around 6,300 data scientists in the US. We hear all about AI at industry-leading companies, but the reality is that the vast majority of normal companies have no AI talent.
With Akkio, any firm can effortlessly build and deploy churn prediction systems. Let’s walk through a simple example in Zapier. First, we create a “trigger” to pull in new customer data.
For this example, we’re using a Google Sheets file as the source of new customer data. Note that any database source could be used.
We’ll then use the churn prediction model we’ve built to predict whether a new customer is likely to churn. To activate Akkio in Zapier, visit this link, and then select “Make Prediction in Akkio.” Make sure that the data you input exactly matches up with what was used to build the model.
Now, Akkio will predict whether a new customer is likely to churn. We want to capture that prediction, and if the prediction is that the customer will churn, we can send a Slack warning to a salesperson. To do that, we’ll need a filter that runs if the prediction is simply the word “Yes.”
Finally, we can send a Slack notification to a salesperson for any customer that Akkio predicts is likely to churn. Alternatively, you could send an email, a text message, or even a device notification.
We’ve explored how it’s easy for even non-technical people to deploy machine learning models with Akkio and Zapier.
That said, there are a wide range of deployment options, including via the Akkio API, Salesforce, Google Sheets, Snowflake, and a web app, with many more methods coming soon. Akkio’s API, in particular, makes it possible for businesses to plug the power of predictive analytics directly into their data pipelines.
We’ve explored how churn costs companies billions of dollars a year. Large enterprises employ expert AI teams to build churn prediction models, but these efforts are too costly and time-intensive for millions of regular firms.
Using Akkio’s no-code AI, you can effortlessly build and deploy churn detection models, and warn sales teams via Slack when a customer is at risk of churn.