In a previous post, I discussed how business team’s questions of “why” are an essential driver of data science value. But because of the greenfield of opportunities data science presents, we should guide those business teams toward valuable outcomes.
Guided analytics, I concluded, can help business teams take advantage of advanced analytics techniques while driving them toward greater business value rather than down potentially interesting but less consequential paths.
In this post, I will discuss how my team uses guided analytics in a customer-centric approach to development – for what the customer is trying to solve – rather than a traditional product-centric approach based on product features and capabilities. Using the Path Analysis Guided Analytics Interface as an example, I’ll explain how we work with business teams to help them answer the next “why” through advanced analytics while driving toward business value.
Teradata has a wealth of experience in path analytics. After solving hundreds of path analysis challenges for customers and partners, we wanted to make path analytics much more accessible to business teams and intuitive for people without data science backgrounds.
To do this, we developed the Path Analysis Guided Analytics Interface.
If you’ve read The Lean Startup by Eric Ries, you are familiar with the minimum viable product (MVP) approach to product development. In short, you initially roll out a very basic version of a product. In this case, we identified a few components of path analysis projects that were “essential” across customers.
After releasing the MVP, you then measure what works and what users request to prioritize feature development. By introducing the Path Analysis Interface MVP to several accounts, we quickly received feedback on what else users wanted to be able to do. Some users were previously exposed to advanced path analytics, but to others, the opportunities were wide open. Often, users asked, “Can the product do this?” With careful questioning, my team honed in on not customers wanted the interface to do, but the actual problems that they were trying to solve.
Initially, the Path Analysis Interface included a couple parameter dropdowns and a “Run” button. After users selected their parameters and ran the query, the interface pushed the query to the database, which crunched the results and returned those to the interface in the form of several visualizations.
One of the great things about the MVP of the interface is that it was incredibly intuitive. If someone wanted to run a path analysis, they could do so almost immediately.
By design, the interface enabled a new group of users to leverage advanced path analytics. Their questions would help our team decide how to prioritize new capabilities of the interface. The potential opportunities looked like a greenfield for the recently-enabled users. However, the purpose of guided analytics is to direct this potential toward use cases we know will provide business value. (Note that data scientists can – and should – still be exploring the greenfield opportunities.)
As we engaged with customers using the interface, we heard some common questions:
- What about customers who didn’t complete a path?
- Can the interface show me who is likely to complete an event such as churn or upgrade?
- Can the interface tell us what a customer is likely to do next based on what they have done before?
There are many rabbit holes we could go down to answer those questions. However, guided analytics kept us focused on providing business value. By dissecting what customers wanted to solve, rather than what they wanted the interface to do, we developed a “predictive paths” feature that allows users to highlight a portion of a path and build a list of customers who have completed that portion but not reached the event of interest.
Predictive paths is useful for marketers hoping to nudge customers along their journeys or entice them toward more rewarding paths. Based on our experience with data science teams, we also know paths can be significant predictors of future action or inaction, especially when combined with some other advanced analytics techniques. But for business people hoping to prevent churn, for example, simply knowing which customers are on those common paths to churn is of immense value.
Hopefully, this example explains how guided analytics can keep business teams focused on business value rather than data science rabbit holes. As we continue to develop the Path Analysis Guided Analytics Interface and other interfaces for business teams, it’s important to remember that we are constantly trying to help these teams answer “why” to drive business value.