Implementing Self-Service BI with Sisense: Key Steps and Insights

[Priscillareagan]

Implementing Self-Service BI with Sisense: Key Steps and Insights

Our Story: How We Implemented Self-Serve BI Using Sisense

In a team set-up with only a few tech-savvy members but numerous data-hungry stakeholders, building and delivering a dashboard used to take weeks, if not months. The backlog of requests was ever-growing. To address this, we decided to roll out a self-serve BI platform using Sisense.

Starting Point: Gathering Requirements

The first step was engaging with the different stakeholders to gather their requirements. We then prioritized these requirements based on their impact and identified overlapping requests to avoid duplication of resources and efforts. Each requirement was broken down into specific metrics, and we documented the sources for each metric. Leveraging our existing BI platform, Sisense, we began setting up the self-serve solution.

Building the Foundation: Elasticubes and Live Models

One thing worth mentioning is that we translated the requirements into dbt models and made it a best practice to create and maintain most of the metrics in dbt. These models are the sources for our Sisense models. This approach ensures that we maintain a single source of truth, which is essential for self-serve analytics. By leveraging dbt, we can standardize the metrics across the organization, reduce discrepancies, and facilitate better decision-making

Based on our requirement analysis, the team built Sisense data models for each category or department data. For instance, the Sales model included information on complaints received. A separate model was dedicated to complaints and refunds linked with customer care data. These smaller models help stakeholders navigate easily and find what they need, rather than getting lost in a massive, complex model.

We preferred Live Models over Elasticubes because they provide the most recent data from the lake without the need for frequent scheduling. This approach also saves a lot of memory.

Creating Template Dashboards

The next step was to create template dashboards with basic metrics from each model. These templates included simple to complex widgets showcasing metrics based on stakeholder requirements. The purpose of these templates was to give stakeholders a starting point and a reference.

Training and User Adoption

We conducted several workshops to ensure that all teams could effectively utilize the new BI tool. These sessions were designed not just as instructional meetings but as interactive workshops where stakeholders actively engaged with the tool and provided real-time feedback. Sisense’s intuitive UI and easy-to-navigate menu significantly contributed to the ease of use, enabling our non-technical counterparts to quickly get up to speed.

Throughout the training, we fostered a collaborative environment where stakeholders were encouraged to explore the tool and share their experiences. This two-way engagement ensured that the BI team could address any concerns and adapt the training content to meet the users’ needs better. 

We’ve also implemented usage analytics and data governance dashboards to track how effectively teams use the self-serve BI platform. This dashboard allows us to monitor adoption, and identify areas where additional support is needed. This proactive approach has been key to optimizing the platform’s impact across the organization.

The dedication and enthusiasm of our stakeholders were instrumental in the success of this rollout. By working closely with the BI team, they not only learned to navigate the tool but also contributed to refining its use within the organization. This collaborative approach ensured that the deployment of the BI platform was smooth and well-received across all departments.

Handling Feature Requests

As we rolled out self-serve, some feature requests were not native to Sisense. For example, stakeholders wanted to perform 'What If' analyses, a feature available in some other BI tools. Another request was the ability to pick widgets from various dashboards to create a custom personal version. Thanks to Sisense's versatility and customizability, we achieved this by installing third-party add-ons.

Ensuring Data Security

Ensuring robust data security is paramount in rolling out a self-serve BI platform. We have implemented stringent measures to safeguard our data and prevent unauthorized access. One key aspect of our approach is the use of row-level security, which allows us to control data visibility at a very granular level. Additionally, the ability to create and manage user groups in Sisense provides us with the flexibility to assign appropriate access levels to different users, ensuring that each user has access only to the data they need. This comprehensive approach to data security helps us maintain a secure and reliable BI environment.

By sharing our experience, we hope to inspire others to leverage Sisense to implement self-serve BI solutions efficiently and effectively.

Testimonials:

‘We had a training today and I’m feeling super excited about it - it’s going to take a bit of upfront investment from my team to familiarise ourselves with how to use it but I think after that it will save us SO much time (and so much frustration - I was battling with pivot tables in excel for hours this week!).’

-Head of Food Development & Trading

‘When I first started using Sisense I was a complete newbie but I was surprised by how intuitive the tool actually is! Self-serving in Sisense allowed me to explore different data visualizations and patterns and understand more about our data structures as well. I strongly recommend that people use it (not just for dashboards but) if they are interested in a quick data download or for quick visualizations to check a hypothesis! And the data platform team is great in helping out and giving advice when needed!’ 

-Senior Analyst

About the author:  I am a Senior BI Developer at Gousto with nearly 14 years of industry experience. I am passionate about data and a strong advocate of Ronald H. Coase’s quote: “If you torture the data long enough, it will confess to anything.” In my three years with Gousto, I have played a key role in shaping the company’s BI platform, driving innovation and data-driven decision-making.

About the Company: We are a data company that loves food. We combine our passion for culinary delights with cutting-edge data solutions to deliver an exceptional experience to our customers. At Gousto, we harness the power of data to innovate and improve every aspect of our recipe box service.