Good Apple: Scaling Analytics and Unlocking New Value With Dataiku

Good Apple, a highly specialized media agency, delivers proprietary data-driven solutions to help clients optimize performance and maximize impact. Known for its scalable results, Good Apple combines analytical expertise with advanced tools to stay ahead in a competitive market. As client demands grew, the team needed to advance analytics processes and transform manual, one-off projects into repeatable, high-value products.

The Challenge: Overcoming Manual Effort to Scale Insights

Prior to Dataiku, Good Apple had built a solid foundation using Snowflake for data storage, ETL processes for data transformation, and reporting dashboards to provide insights for clients. With this solid infrastructure in place, they looked to streamline deeper analyses — like customer segmentation and attribution modeling — which remained highly manual and resource-intensive. By integrating Dataiku with Snowflake, they accelerated these efforts and unlocked greater efficiencies.

• Custom Analyses: Insights for individual clients were delivered as one-off, manual projects that couldn’t be reused or scaled.
• Capacity Limits: The small analytics team struggled to keep up with growing demand for consistent, repeatable insights.
• Model Complexity: Expanding model sophistication without directly managing extensive code became increasingly difficult.

To solve these challenges, Good Apple needed a platform that combined automation, scalability, and explainability to streamline workflows and empower their diverse analytics team.

Why Good Apple Chose Dataiku: Transparency, Usability, and Scale

In 2019, Good Apple selected Dataiku for its ability to simplify workflows, foster collaboration, and scale their analytics capabilities. At the time, their team included analysts from diverse backgrounds — physics, math, computer science, and economics — each with different tool preferences. They needed a solution that could serve as a common platform, be easy to learn, and support the capabilities of their preferred tools and languages. Dataiku’s visual flows and low-/no-code environment made it easy for analysts — regardless of their data science background — to understand, manage, and contribute to projects without relying on complex code. Creating a common platform helped the team collaborate more effectively and accelerate project execution.

Equally important was the need for transparency and explainability. Good Apple prioritized clear, interpretable models to validate results internally and communicate effectively with clients. Dataiku’s explainability features — such as feature importance and variable dependence — provided deep visibility into how models reached their conclusions. This clarity built trust and confidence both within the team and with external stakeholders.

Finally, scalability also played a key role. With Dataiku, Good Apple transformed manual, one-off analyses into automated, repeatable workflows. By streamlining processes and improving efficiency, the team delivered consistent, high-value insights across industries while maximizing their existing resources.

The visual recipes of the [Dataiku] Flow helped our team understand what was happening and better leverage the platform. Collin Joseph Director of Data Science at Good Apple

How Good Apple Uses Dataiku to Drive Value

1. Customer Segmentation: Faster Targeting With Greater Complexity

Good Apple leverages clustering techniques with complex datasets and advanced performance metrics to segment target audiences — such as healthcare professionals — into lookalike populations with shared characteristics. These insights make it easier to identify the right audiences and improve targeting strategies for clients.

Before adopting Dataiku, the team relied on manual scripts for segmentation, which limited model complexity due to time constraints and challenges with reusability. With Dataiku, the team automated the segmentation process, enabling them to build more sophisticated models within the same timeframe — without the need to directly manage code. This efficiency gain allowed Good Apple to scale the solution beyond a single project. What began as an analysis for one healthcare client became a repeatable offering used across industries.

With the efficiency increase, we expanded customer segmentation to other clients and other verticals. Collin Joseph Director of Data Science at Good Apple

2. Attribution Modeling: Delivering ROI Insights Faster and More Consistently

Attribution modeling is a core competency of Good Apple and plays a crucial role in how their clients measure the business impact of their media campaigns. The process brings together diverse datasets — marketing activities, market conditions, and competitor performance — to understand how much sales can be attributed to specific campaigns.

Before Dataiku, building attribution models took a data scientist six weeks, with outputs delivered infrequently due to the manual effort required to update code and incorporate new datasets into the data flow. Today, Dataiku has cut the time required by 50%, enabling the team to develop models in just three weeks while running a much broader set of analyses.

This shift also empowered junior analysts — alongside data scientists — to build and execute models using pre-defined workflows, significantly expanding the team’s capacity. As a result, Good Apple now delivers attribution insights on a quarterly or monthly basis for multiple clients, providing timely and actionable ROI data.

With Dataiku, we’ve made attribution modeling a consistent, repeatable process. Now, we’re running it at least quarterly for several clients, biannually or annually for others, and even monthly in some cases. The throughput has improved dramatically, and we’ve gained more experience doing it, which has made the process even more efficient. Collin Joseph Director of Data Science at Good Apple

3. Programmatic Bidding: Predicting Ad Performance to Optimize Spend

Programmatic bidding is a key strategy for optimizing digital ad spend. Good Apple uses a proprietary process that creates a multi-stage algorithm to create predictive bidding models. Now built in Dataiku, they forecast the optimal bid price for display ads, ensuring campaigns deliver the highest possible ROI.

Here, Dataiku serves as a critical link in the model development process. Once built, the models generate bid lists that feed into Good Apple’s programmatic platform, where campaigns are executed in real time. This programmatic bidding product is still in beta testing, but the ability to generate outputs at scale would not have been possible without Dataiku’s automation and infrastructure.

Delivering Scalable Insights and Unexpected Benefits With Dataiku

Good Apple elevated its tech stack to match the sophistication of its analysis by automating processes, scaling insights, and delivering consistent, high-value solutions to clients. By integrating Dataiku with Snowflake, the team ensures seamless data workflows — covering everything from model training and scoring to visualizing results — while eliminating previously manual tasks. With Dataiku, Good Apple’s data scientists can now focus more on data science and less on logistics and coding.

An unexpected benefit of Dataiku has been its API capabilities, which allow Good Apple to streamline recurring workflows like attribution modeling and programmatic bidding. This automation not only accelerates delivery timelines but also ensures clients receive consistent, actionable insights to measure ROI and optimize strategies effectively.

With Dataiku, Good Apple successfully turned resource-heavy, custom projects into scalable solutions that align with their core business goals: empowering clients to achieve measurable results through smarter analytics. Looking ahead, the team is setting its sights on exploring generative AI to unlock new opportunities for innovation.