Our Insights

A data product is a custom-built reusable asset that pulls in relevant data to provide valuable metrics and information to anyone using the product. Data and analytics products at Lewis Media are designed to help our buyers optimize their campaign plans, and to provide transparency to our clients. Although the fun part of building data and analytics products is crunching numbers and writing code, there are steps that must be taken before we get there. Having a structure to follow is paramount to delivering a valuable product. Aligning goals and taking an idea to deployment requires careful thought. Over time, we’ve refined a framework that guides us through this process, ensuring that the final product we create is not only useful but also scalable and valuable.

Creating the Vision

Every great product starts with a clear vision. If you don’t understand what your product is meant to achieve, then during the build you end up going in circles. It all begins with conversations — talking to stakeholders, understanding their needs, and identifying the specific problems we aim to solve. In some cases, I’m building a data product for the LMP team to closely monitor a campaign’s performance as it runs. Other times, I’m working with the client to create a dashboard that puts relevant information all in one place for easy reference and understanding. It’s crucial to ask the right questions: What pain points exist? What opportunities can be seized?  What do you not know now that would be useful?  What will you do with that information?

Once the problem is defined, the next step is to figure out what data is needed to tackle it. Sometimes, the required data is readily available; other times, it might need to be sourced or even created. With the problem and data in mind, I then consider the appropriate analytical techniques that will lead to meaningful insights. To bring this vision to life, I create a mock-up of the final product, giving me a tangible idea of what the end result will look like.

Building Data Infrastructure

With a clear vision in place, it’s time to delve into the data infrastructure. This involves identifying where all the relevant data lives, whether in internal databases, external APIs, or other sources. Mapping out these data sources and understanding their structure is essential for seamless integration later on.

Prototyping is a key step here — testing the process of automatic data fetching to ensure that everything flows smoothly. This not only helps in identifying potential roadblocks early on but also sets the foundation for a scalable system that can handle the product’s data needs as it grows.

Conducting Analysis

Now comes the analytical work. With the problem defined and data infrastructure in place, I dive into researching the best techniques to solve the problem at hand. This step often involves a combination of reviewing documentation, consulting with experts, and leveraging existing tools or models.

Prototyping again plays a critical role here. By testing the analysis on a subset of the data, I can quickly assess whether the chosen methods will yield the desired results. Once validated, these insights are fed back into the mock-up, turning the vision into a more concrete reality.

Delivering the Product

With analysis in hand, it’s time to present the product to stakeholders. Using real data in the mock-up, I demonstrate how the product addresses the identified problem or opportunity. This step is all about communication — showing stakeholders that their needs have been understood and met with a tangible and functional solution.

Feedback is invaluable at this stage. It’s the best opportunity to make adjustments before the product goes live. By incorporating stakeholder input, the product can be polished and fine-tuned to ensure it aligns with expectations and delivers value.

Deploying the Solution

Finally, the product is ready for deployment. But before launching, it’s crucial to address any potential scaling issues. This means ensuring that the data infrastructure can handle increased loads and that the analytical processes remain efficient as the product scales.

By leveraging cloud resources, both the data infrastructure and analytical techniques can be scaled to meet demand. Sometimes, the data infrastructure needs to be adjusted, or a different technique that is less computationally intensive needs to be honed. The final step is to automate the flow of insights into the product.

Summary

Before we begin any analysis or data product creation, we always refer to this type of framework. Having a strong foundation and established processes in place leads to faster delivery time with a product that provides value to our clients and media buying teams.