We are thrilled to announce that Unilytics has joined Optimus SBR
Learn more

Client Success Story – Melissa & Doug

Client Success Story – Melissa & Doug

Toys and Tableau

Unilytics markedly improves dashboard performance, enhances user experience, and delivers better insight to Melissa & Doug management.

“The results are fantastic and senior management couldn’t be happier” – Steve Brown, FP&A & Business Intelligence Leader, Melissa & Doug

Industry: Manufacturer (Children’s Toys)
Organization: Melissa & Doug
Website: https://www.melissaanddoug.com/
Technology Platform: Tableau Software


Founded in 1988, Melissa & Doug, LLC is a manufacturer of children’s toys including wooden puzzles, arts & crafts products, plush toys, and other educational toys. The company has multiple factories  and substantial sales growth every year. Its mission is to provide a launch pad to ignite imagination and a sense of wonder in all children, so they can discover themselves, their passions, and their purpose.

Melissa and Doug realized that to remain competitive and to deliver optimal products to their customers, a better understanding of their business was imperative. Improved analytics for their account & product sales, supply chain, and spend would help them drive better value to their partners, employees, and customers.

Tableau was purchased to create interactive dashboards to provide greater business insight across the business.


Melissa & Doug hired one of America’s top consulting firms to build their data repository and help create Tableau dashboards. However, performance of the resulting dashboards was very poor. One of the dashboards took an average of 45 seconds to open, plus another 24 seconds per click. This was unacceptable.

The moderately sized data source with approximately 45-50 million rows of data spanning multiple years, should not have created such significant slowdowns.

Steve Brown, Melissa and Doug’s recently appointed Business Intelligence leader, reached out to Unilytics in the hopes of remedying the existing Tableau implementation.

Various aspects of the Tableau implementation could have resulted in poor performance. Determining which components were causing the majority of the degraded performance proved challenging. Was performance degradation occurring in server hardware, or in data retrieval, or in computational location (e.g., in the database or in the workbook), or in workbook complexity, etc.? Any number of components can cause performance issues, and “fixing” components in the wrong place would result in minimal or at most mediocre performance increases.


Although Tableau technical teams did not have a specific proposal / solution for M&D, they did recognize Unilytics as a top and trusted partner. After Unilytics articulated how incorporating a lightweight analytical layer would significantly boost performance, how it would be created, and quantified historic successes, Mr. Brown decided to push forward, believing this was the answer.

Rather than taking a proscriptive approach to the problem and insisting Melissa & Doug fix the issue a specific way, Unilytics worked with internal team members to achieve incremental, measurable fixes. The incremental fixes were designed to minimize the impact within the organization in terms of effort and reporting. With each incremental fix additional meetings were held to review the results, plan next fixes, and prepare internal teams for the tasks necessary to achieve the next incremental fix. During this incremental fix process, the internal team also learned improved ways to use Tableau going forward so organizational knowledge and efficiency increased throughout the process.

Unilytics….what was done?

  • Evaluated various options on how to resolve performance issues, including simplifying existing calculations “in place” within the dashboard, simplifying existing data sources, splitting dashboard functionality into multiple workbooks, etc. After trying to “fix” the symptoms of the inefficiently designed data source yielded moderate results, it was determined that a full data source redesign was necessary. This decision was based on numerous “deep dive” performance investigations and evaluating the impact of various changes.
  • Evaluated whether performance was related to using Tableau Online versus an internal Tableau Server where server hardware could be modified. Using performance recordings, it was determined that performance issues were unrelated to the hardware present in Tableau Online and that relocating the service to an internal Tableau Server would not solve the performance issues.
  • Created highly optimized and minimized data sources to replace an existing “monolithic” (i.e., single, large data source). Resulting data sources had the optimal granularity and only contained the fields that were needed for their respective dashboards.
  • Pushed intensive calculations from Tableau back into the database in an analytics layer built inside the database. This removed significant processing time from the dashboard rendering process.
  • Simplified calculations in Tableau as much as possible. Numerous calculations had thousands of characters worth of excessively complex formulas due to mismatching levels of detail in the underlying data sources (these excessively complex formulas were created by the original consulting firm who implemented Tableau). This required considerable investigation of hundreds of calculated fields, including multiple nested table calculations (i.e., table calcs) and level of detail (i.e., LOD) expressions. These formulas were greatly simplified after pre-aggregating the source data to a level appropriate for the dashboards.
  • Conducted rigorous performance testing on multiple types of data source architectures (i.e. live connection joins, extracted joins, extracted multi-table joins, data blending). Compared processing characteristics of the various data source architectures to select the optimal one for their specific needs.
  • Simplified the approach to date calculations within the overall analytical environment. The company uses the “merchant calendar” or 4-4-5 calendar which can be complex to compute inside a Tableau workbook, so the computations were eliminated, and a date dimension table was used instead. This greatly simplified date calculations and resulted in much more maintainable calculated fields in Tableau, as well as correcting some logic that was skewing results for “adjustment years” in the 4-4-5 calendar.


“The results are fantastic and senior management couldn’t be happier” – Steve Brown

The dashboard improvements have been exceptional, with initial load times down from 45 seconds to 6 seconds and refreshes on click down from 24 seconds to between of 1 – 3 seconds. That makes for a great user experience. And by redesigning their environment, Unilytics was also able to deliver better insights to management teams as they can now ask more granular questions of their data.

“I was very impressed by the whole development process. The Unilytics team not only delivered extraordinary results, but they engaged us along the way providing various options to best meet our needs.” – Steve Brown

In addition to achieving excellent performance results, Unilytics also delivered a highly simplified version of the workbooks and their contents (e.g., calculated fields). This makes ongoing maintenance of the workbooks much simpler and reduces overhead costs of maintaining the workbooks for corporate use.

Unilytics also put infrastructure in place to facilitate creation of new dashboards using data source and style templates. This allows power users to extend existing analytical benefits deeper within the organization with less centralized control required and with less up-front effort to start building dashboards.




Want to know more?

Contact us