• Maciek Lasota

How to add discovery phase and create Dual-track Agile in Analytics.


Implementing design thinking methodology in building BI solutions with Kanban.


During five years of managing the Data quality Dashboard, we moved from a centralized old-fashioned BI tool to a modern user-centric platform that allows self-service analytics. The mission stays the same - to help improve data quality by showing relevant metrics and dashboards.


Kanba Agile flows for Analytics and Data

Besides the shape of the tool itself, the development process also went on a long journey. Combining design thinking, Lean and Agile is currently fundamental in the delivery process.

Why this is important:

1. Better product - Better validation with end-user

Only validated features are delivered to customers. It minimizes the chance of releasing a feature that is barely used

2 Efficiency - less waste and lower costs

  • Breaking flow into two means that is more likely to make it right the first time instead of going through a few iterations to calibrate the product this results in faster development and release cycles

  • As only valid features are allowed in the product backlog so there will be fewer wasted resources.

We started with Agile and development with users clearly defining what they needed in form of functional requirements. The product built the right way to be responsive and dynamically adjust to customer's needs. With grow of maturity and skills at some point, there was a feeling that we can do more by being proactive in data quality aspects.

The goal was to deliver features beyond ready requirements and start creating an analytical solution that brings even value and using data profiling. Our expertise and skills grow but we still didn`t quite hit on customers' needs.

This is when Design101 workshops took place.

The key learning was to empathize with our users and let them talk about the problems not just wait for them to provide requirements. We started using our experience to find and

tackle problems instead of just developing the software.

Our Approach changed and we were creating micro products that we showed to our stakeholders (Small MVPs or how we call it Minimum Viable Dashboards) then we follow up on initial feedback and either go further with development or just cut our losses.

Concept is a combination of Design Thinking, Lean Startup, and Agile development It contains 2 coexisting tracks - Discovery and Development. Both tracks in the kanban system are working parallel but have different scopes and goals.





Dual Track agile flow  for business inteligance and analytics soultion



Delivery is about building products the right way. Using Agile we can be dynamic responsive and allows our team to quickly adjust to customer's needs.

Discovery is about building the right product using:

  • Empathy: Getting to grips with a real user problem and building empathy for the target users/customers. It`s about research and looking for opportunities. The main tools are Interviews Survey and Analytics

  • Ideation, innovation, and problem-solving: Generating as many ideas and potential solutions as possible.

  • Prototyping and testing: Building low-fidelity prototypes of the ideas generated, ready for testing on real or representative users. It allows to create validate ideas before committing to build them.


One of our favorite things is to fill in the Value proposition canvas whenever we perform user interviews. It really helps to get off the right foot.

Discover Track should answer 3 questions plus one extra:

  • Will users use it? (Value)

  • Can users use it? (Usability)

  • Can we build it? (Effort)

Those 3 were standard but there is one extra we added for functional features

  • Could we reuse it ( reusability)

So if we manage to create a useful statistical algorithm for deduplication or functionality like Wrong data alerts potential value for another content dashboard should consider. This can tip the scale in many cases.

To be decentralized and more agile we split the product which is now more of a solution or platform into smaller products with different owners and customers so we can easily perform the design phase as applications were supposed to serve very different groups of users

Technology change matters.

As started the whole application was custom developed backend database level plus a C# webpage with scorecards. At some point, we started using PowerBI which was a gamechanger because instead of processing each small change with sprint planning we could have tweaked the report with quick slice and dice.

We also had our first super users that we're able to create their reports using cleaned datasets that we prepared.

Roles

With the team being the mix of developers and analysts we tried to avoid setting lines between and trying to involve everyone in both tracks but devs are going to be more involved in backend development on the database level plus some application while the analyst is more focusing on the design track with some additional power bi report creation but it`s just on paper in practice there is a lot of cooperations.


Dual Track allows teams to put the effort into the right features and launch products that will bring value. Subscribe and follow me on Twitter not to miss the next part of how we implement this is on our team.

Recent Posts

See All
Suited Data Management Solutions

Solution Tailor

Blog about Data Management and Quality  Analytics Solutions.