The Data Technology Stack your Customers Want is a Difficult one

David Alvarez Debrot
April 4, 2024

The article discusses the complexities of building a data technology stack that meets the demands of SaaS customers, emphasizing total integration between operational and analytics modules. It highlights three main challenges: simplifying data complexity for end-users, ensuring data level permissions align across interfaces, and seamlessly integrating UI to maintain brand consistency.

TABLE OF CONTENTS
Technology
Semantic
Analytics
All

Total Integration

The technology stack many SaaS customers need is one that has total integration between all layers - see the Value Proposition angle on this in our article What B2B SaaS customers really want from analytics and dashboards but what does it exactly mean?

Data is key in decision making, the more efficiency gains a company looks for the more direct, near-time data driven operational decisions they will want. In a SaaS context this means that the integration between the reporting modules and the operational modules is the responsibility of your technology stack. Why? Because if it isn’t your application delivering this value then it’s the customer that will need to do data extractions, analytics and then find a - usually ineffective - way back into their operational application (think Excel sheets with IDs or filter criteria that need to be then copy pasted into the operational system). Stay tuned for the full story in an upcoming article.

What this implies is that your Data Analytics and your Operational Modules are merging into a Solution in terms of data flow, security, separation of concern, UI, branding, auditing and more. And that is not as easy as we could think before having done it. 

The Top 3 Difficulties

Simplify Data Complexity

Give your end-users a consistent & accessible view of unified business concepts across the various parts of the application which then allows for specific customizations for specific roles without each individual worker needing, themselves, wasting their time drag & dropping these concepts to build charts, tables and filters while at the same time allowing the combination of data from different data sources and finally go the extra mile with dynamic concepts, advanced statistics and much more. Add to this that you might have customer specific customizations that bring in their own concepts in potentially hierarchical ways but still need to integrate with the rest of the functionality and this starts being fun to develop without making it a hot mess. 

Data Level Permissions

If your user can’t see the data in your operational interfaces they shouldn’t see them in their data analytics either. But think about it, modern data analytics are all about moving quickly from the overview into the details. This means you need both row level and aggregation strategy level ways of masking the details that are out of the scope of responsibility of the given user without preventing them to find what they need. Additionally we want a single authentication and authorization definition that is applied consistently through all layers without the need for human interventions. 

UI integration

Your brand is one software for your clients and not multiple ones so if you are convinced at this stage that the data analytics part is not your core skill and you want to integrate with a partner you need to make sure that the UI integration they offer is fully flexible so the resulting application is seamlessly working and showing up together. This constraint is not only a pure color scheme/branding question but almost every aspect of the user interface logic needs to feel more like customizable components to your technical team than a block of foreign elements squeezed into your beautiful application.

Some real-life case studies to illustrate the Data Driven operational application technical challenges

Power Plant and Energy Simulation Platform

A Energy Analytics Platform simulating and forecasting power market operations, processing highly complex data calculations on-demand to optimize machine usage on AWS.

https://www.iccube.com/use-cases/embedded-analytics-with-aws

24/7 Consumer Panel Analysis Platform

A Consumer Panel Platform delivering market monitoring and advanced analysis, needing high quality-of-service (QoS) and scalability.

https://www.iccube.com/use-cases/24-7-data-analytics