Embedded dashboard of the month: Inspyrus
Feb 26, 2021
Welcome to our first embedded dashboard of the month blog post. We’re kicking things off featuring a Looker dashboard by Inspyrus, a company that helps their customers automate invoice processing and payments in the cloud.
I spoke with Prashant Soral, CTO, who heads up development and anything analytics related for Inspyrus. He told me that they switched to Snowflake and Looker for embedded analytics to achieve faster load times, greater visibility into the customer base, and the ability to quickly make and deploy improvements.
How Inspyrus uses Looker embedded dashboards
Tell us about your dashboard. In general terms, what is its purpose? Who uses it?
Our analytics provide customers with a comprehensive view of how many and what kinds of invoices they are processing, the invoice processing cycle time, processing automation rates, and payments.
The dashboards are used by our customers (Accounts Payable, Controllers) and internally (to view data across customers).
We currently have 9 dashboards. Here’s an example of one:
What were your goals in setting it up?
We had the following goals in moving to the new platform:
Solve the data load issues we were seeing with our previous data solution.
Loading data from the transactional database to our previous solution took an hour for each customer, and the loads did not complete successfully in a consistent manner.
Ability to provide a unified view across customers to our internal audience.
Our previous solution required a separate instance of the dashboard for each customer. This prevented us from seeing a holistic view of our customers.
Ability to make improvements and fix issues rapidly.
The old platform required us to replicate each change to each customer manually. Making a simple modification took 2 weeks of a developer’s time and was error prone. We got to a point where we were not able to make any improvements and had to resort to fixing issues only for the customers that reported them. Now, we can deploy improvements for all customers simultaneously by making changes in the modeling layer.
What results have you seen since implementing Looker?
We have been able to reach all of our goals. Data loads are near real-time with Striim. A single centralized data warehouse in Snowflake provides us the ability to query all of our customer data together. The developer-friendly process in Looker enables us to make improvements rapidly.
Creating their embedded dashboards
Which team members were involved in creating your embedded analytics?
All the work was done by 3 individuals:
- Me, CTO (warehouse design, data transformation, visualizations in Looker),
- A data engineer (loading data from OLTP to Snowflake using Striim)
- A UI Developer (to expose embedded dashboards through our UI)
How long did it take to build your first iteration (MVP)?
- 3 months to select the individual components of the platform (Striim, Snowflake, Looker)
- 2 months to build the data load application in Striim.
- 2 months to design the dimensional data model and build the dbt model to transform from transactional to dimensional model.
- 1 month to design and build visualizations in Looker.
- 2-3 weeks of testing.
We did all our development using production data. This allowed us to resolve issues related to data cleansing, report design and performance during development instead of finding them later during production use.
Did you have a “wow” moment when making this dashboard? If so, what were you surprised that you could do or achieve with Looker?
It was not really a surprise because these were the main reasons I selected Looker — the way Looker allows conventional development practices to be applied to visualizations, and the ease with which one can embed row level security using user defined attributes is amazing.
How the dashboard has impacted everyday operations
In what way has embedded analytics impacted the way you sell your product?
The old dashboards did a great job of helping us to sell. As a result, the sales cycle has not seen much of a benefit. However, our customers enjoy much faster data load times now, and our newfound abilities to see a comprehensive view of our customers and to rapidly deploy improvements has felt like a huge upgrade.
What to look for in an embedded dashboard platform
How do you measure the success of this dashboard?
There are two factors:
- Is the data getting loaded properly and in a timely manner?
- How quickly can changes be made?
If you could go back in time to the beginning of this embedded analytics journey and give yourself one tip, what would it be?
If I was to select one, it would be to hire professional services from Striim to work through initial configuration. Configuration items in Striim have a major impact on performance. Knowing them upfront would have reduced our development time further and would have avoided a few data load issues we encountered post-production.
If you’d like to see Looker’s embedded analytics and reporting in action, you can join a group demo.
For a more personal touch, we can help you explore what an embedded analytics deployment would look like for your specific organization. You can get started here.