Predictive Insights

Effortlessly Segment and
Surface The Key Drivers Of Your KPIs

With Predictive Insights, you simply tell our system what you care about and you'll then be provided with a variety of different segments and conditions that lead to an improved KPI. You can also train and host A.I. models, using them to generate your own predictions, which you can consume straight on Apteo or integrate into your own apps via our API.

Watch: How to Surface Insights In 5 Minutes
How It Works
Visualizations let you explore and understand your data using
beautiful charts and graphs. Easily filter, group, and visualize
your data with just a few clicks.
Connect Your Data
Using our data connectors, you can plug in your data whether it lives in a CSV file, database, data warehouse, or API. Once your data's connected, it's easy to keep it up-to-date. No need to download extra software just to create a graph.
Analyze Your Data
Apteo's platform makes it easy for you to explore your data in lots of ways. You can graph it using conditions and groups that you specify, run statistics on it, and build your own A.I. models to create predictions, all without knowing data science.
Share And Collaborate
It's easy to build dashboards out of any of the visualizations that you create. Simply create a new workspace for your team, add your visualizations to your dashboards, invite your team, and set their permissions.
Auto ML

Automatically Create
Predictive Models

Apteo's platform lets you create your own machine learning models using a few points and clicks
of the mouse. Plug into these models with a browser or an API.

ML Models

Forecasts and Predictions

Identify Key Drivers of KPIs

Generate Predictions Via API

Try It Now

Common Questions Before Getting Started

Here are a couple of tips to help you make the most of the predictive insights feature

What kind of things can I do with predictive insights?

Our Predictive Insights module uses automatic machine learning to help you do three key things:

  1. Segment your data into groups based on high and low values of your KPIs (i.e. churn, sales, etc)
  2. Automatically set up a machine learning model that you can use to predict your KPIs going forward
  3. Identify the key drivers and factors that impact a KPI

When you use predictive insights, you'll be able to analyze any KPI for which you have descriptive data. For example, you can identify which of your users are likely to churn, what's causing them to churn, and who's likely to churn in the future. Or you can use it to analyze the key drivers of sales and find leading indicators of ARR. The possibilities are endless!

How do I make sure my data is set up properly?

The best way to structure your data is to think about your rows as individual examples that represent a single thing (like a user, if you're trying to analyze churn, or a transaction, if you want to predict fraud).

Your columns should be individual attributes and factors that provide some interesting, and hopefully relevant, information about your row. In general, it's better if your columns include information that's relevant to your KPI. When analyzing churn, for example, this would include demographic data, data about which plan a user is on, and behavioral data (for example, how many times did the user log on to your service, or how many times have they purchased something from your site).

You can see an example of a sample dataset we have available around churned users to give you an idea of how you might want to get started. You can also schedule time with one of our data experts to ask how to get up and running. Finally, you can read our How To guide to get started yourself.

Do you have an example of what my dataset should look like?

Yes! If you'd like, you can download an example of a dataset that's useful for churn analysis here, and if you'd like to learn how to do it yourself, you can read our how to guide below.


How To Create The Best Dataset For Predictive Insights

Learn how to create an optimal dataset for predictive insights

How do I know your system is accurate?

Segments in our predictive insights model are created from the data that you connect to the system. You can dive deeper into them and run your own calculations to verify that the average value of the KPI in each segment is as we report it to be.

A.I. Models
When we set up an A.I. model based on your data, we use several different learning algorithms and we cross-validate each model to find the one that has the best accuracy (basically we use a process of training a model on part of your data and evaluating it on data it hasn't seen yet). You can always see the list of models trained, the accuracy metric for each cross-validated fold, and the overall accuracy of a model by clicking the "Model Info" tab at the top right of the Overview page for a Predictive Insight.

Said differently, we evaluate your models and give you the results, but you can also do this yourself by creating two datasets, one that you use to train a model on Apteo, and the other that you use to test the predictions of that model. You can then evaluate the performance of our model yourself using the dataset that you didn't use to train the model.

Want some help getting started?

Schedule A Chat With Our Experts