We asked PredictSheets founder, Adam Roach, about building an app that uses machine learning to analyze data and generate predictions for businesses. Adam has worked as a data scientist and enjoys analyzing complex data sets. He spends his freetime building apps like PredictSheets. He built his app on Bubble because the no-code platform supported all of the features he needed such as "complete front-end customization, subscription payments with Stripe, charts for reporting, and cloud services for back-end processing."

Tell us about your app and what it does.

Adam Roach:  I built PredictSheets to help business owners who aren't familiar with machine learning analyze complex data sets and predict outcomes. My vision is that by using these data sets, business owners can increase the value of their business. A few common cases where PredictSheets has helped companies are in reducing customer churn, preventing employee attrition, and boosting sales-win rates.

How does your app work?

The best part about PredictSheets is that it’s simple. Our user uploads a dataset, like an Excel file, to our platform. Our system will then automatically build a model to generate predictions.

Once the predictive model is built, the user can download the file and you can "score" the new data. For example, if you have new sales leads, you can use the model to predict the probability of winning for each lead. PredictSheets enables you to prioritize your outreach accordingly and improve your win-rate.

There is only one catch. Depending on where the data comes from, the user may have trouble formatting the initial dataset. To combat this, I currently offer Unlimited Premium Support (for launch only) to all subscribers. I will walk each user through the process step-by-step to ensure you get actionable insights from your data.

Uploading a dataset to generate a predictive model on PredictSheets
Once a user uploads a data set, a predictive model is automatically generated.

Tell us a little about your background and how you discovered no-code tools like Bubble.

As a data scientist, I already know how to code. My expertise include analyzing complex data sets and building machine learning models.

When I decided to build a SaaS application to offer my expertise in a self-service manner, I had no idea where to start. I worked on a mobile app, Artyst, but had to hire developers for it. It turned out to be really expensive, time-consuming, and stressful. I didn't want to go through that again, so I  searched for no code front-end development tools.

Luckily, I stumbled upon Twitter's growing no-code community and never looked back. The rest is history.

PredictSheets founder Adam Roach, Bubble user
PredictSheets Founder Adam Roach found no-code and never looked back.

Why did you build your app on Bubble?

After Twitter introduced me to no-code, I looked for a tool to build my platform. Bubble was the obvious choice. As a founding player in the no-code game, Bubble offers an expansive support community and an extensive set of plug-ins for any imaginable feature an app might need.

For PredictSheets, I needed an app that supports an administrative user interface, Stripe's subscription payments, reporting charts, and the back-end of a cloud computing system for machine learning algorithms. Bubble offered all of the features I needed. I couldn't find any other platform that would support it all.

What have been your successes or milestones so far?

After I posted a Medium article documenting my process building PredictSheets, I received positive feedback from no-coders interested in building similar applications. Though my Twitter following remains small, I received a lot of traffic to my app through my Medium article.

With the value I offered, including the one-on-one support for early PredictSheets users, my post led to new signups early on. The launch users taught me about what really drives companies to seek out predictive analytics. This helps me improve the platform to better meet their needs.

What are your next steps or future plans for your app?

PredictSheets is still pretty broad right now and can be used for all predictive analytics cases. This is a great start, but if I narrow my focus on a single use case, I believe it will bring more value to the app. Narrowing my focus will allow me to tailor the user experience with more automated processing and reporting. I want to offer more custom integrations with user data sources and more complete, pre-built reports that deliver valuable insights and drive action within the organization.

I haven't decided what this niche will be yet, but I'm actively monitoring early user feedback. I hope to deliver huge value to an underserved market.

My vision is to find someone who is wasting 15 hours every week manually pulling data into giant, crash-prone Excel sheets and running reports. I want to say to that person: "Stop what you're doing. Delete the spreadsheet. I've automated this entire process. You never have to go through this process again. You can have this all in one-click, any day, any time."

Learn more about PredictSheets: Twitter

PredictSheets logo

About Bubble

Bubble is a leader in the no-code movement. Bubble offers a powerful point-and-click web editor and cloud hosting platform that allows users to build fully customizable web applications and workflows, ranging from simple prototypes to complex marketplaces, SaaS products, and more. Over 400,000 users are currently building and launching businesses on Bubble - many have gone on to participate in top accelerator programs, such as Y Combinator, and even raise $365M in venture funding. Bubble is more than just a product. We are a strong community of builders and entrepreneurs that are united by the belief that everyone should be able to create technology.

Join the no-code movement today.


Do you want your app featured by Bubble? Reach out to us and tell us about your app. In addition to this series, we sometimes select our best apps to showcase on our website.