How to Validate a Product from Start to Finish

Using Backwards Interfaces to Execute on Data Collaboration

Previously, I worked at an enterprise SaaS company in the ad tech industry. Simply put, the company's goal was to make it easier for organizations to use data effectively. During my time at this company, I created an audience segmentation product that enabled marketers to extract the maximum value from their data, resulting in a higher return on investment (ROI) on their marketing spend.

"I created an audience segmentation product that enabled marketers to extract the maximum value from their data, resulting in a higher return on investment (ROI) on their marketing spend."

I was exploring an idea that allowed customers to join different data sets; the hypothesis was that if we allow customers to join datasets, then it would increase the amount of data that customers distribute from our platform, which is directly tied to product usage and revenue. The amount of data distributed was directly proportional to revenue and product usage in our organization. Customers had been previously asked for this tool and had been enabling it through a "hacky" process using our platform.

Steps for Validation and Understanding Use Cases

Understanding the Customer:

When I began exploring this new product idea, I wanted to get into the mind of the user, learning what matters to them. I worked closely with product design and account teams to understand the customer use cases, but I first broke these down into more specific user personas. There were two major personas that would be impacted which were very different from one another.

The first type of persona was the marketer, and the second was a data operations manager. While they sound similar, they work to accomplish very different goals. The marketer is trying to advertise and sell more products for their company, whereas the operations manager wants to sell more data for their company.

Learning the Different Use Cases:

After segmenting these two personas, I wanted to break down the different use cases between them. After extensive user research, I learned that both the marketer and data operations manager had similar use cases. Examples of these use cases were data insights, segmentation of audiences, expansion of data sets, collaborating, and accommodation of overlap reports. Determining the use cases for each persona was vital because I was trying to decide which one I should focus on.

Ultimately, the user needs of both of these personas were being solved by in-house data teams or leveraging external tools. The customer experience was poor, and I believed that my company could engage with these users more effectively. We already had the central asset of identity, and we wanted to streamline the customer experience.

Delving into the Problem:

After some initial research, I learned that what mattered to users was time and money. Specifically, while the users enabled their needs using other methods, it generally took around three or four times longer than using our company's in-house method. The time spent using data-science resources or external tools was expensive and increased organizations' budgets. While this may not directly affect the operations manager or marketer, it affects the brand as a whole, meaning I needed to validate my product for the entire brand.

Validating the Brand:

I began the validation process by ranking each of these use cases by importance. I learned that partner collaboration and the idea of combining data were most important to the users as they played into one another. With privacy policies in place, it is difficult for companies to share data and collaborate without a neutral third party.

Experimentation:

Holding onto this sliver of the market, I decided to start running two different experiments with each persona. The first experiment ran an overlap analysis where a customer could overlap data with another partner to receive a report back. The second experiment involved providing tools for partners to combine their data without sharing their individualized data. At the end of these experiments, the overlap reports came back to be more beneficial than the collaboration tools.

After digging deeper into this, it came back to each persona. The combination of data wasn't solving the needs of the operations manager and data provider, meaning that my experiments targeted the wrong persona. On the right hand, these tools benefited the marketer more, and they were more willing to pay. At the end of these experiments, I learned to satisfy the brand persona while targeting accommodations and overlap reporting in a collaborative environment.

Incremental Releases:

Moving forward, our company decided that the best way to continue was to incrementally release our product. We knew that we needed to produce a self-service UI because of the way the brand would interact with the product. Our company received feedback that the initial test interface was too slow and didn't effectively increase the time span. After understanding the feedback and user needs, our team began executing on the product.

Our new product resulted in a double-digit increase in data usage for the impacted customers. The time frame to distribute data for these customers was decreased by seven days (from nine days), and in turn, our company increased its revenue.

Producing Top-Notch Products

• Some of the best ways to find new product ideas in your company are to find out how your customers are using your product in unintended ways.
• Figure out how your customers are using your product inside and outside of it. Understand their entire world view to simplify their workflow and collaboration with your products.
• Understand the problem, user personas, and use cases before thinking about products or features that would solve them.