Hawkers had a complex data import process for their restaurant POS system that relied on logic in Power BI and Excel that was based on inaccurate and incomplete data. They needed to automate the data import, fix syncing issues between locations and be able to auto-adjust data based on thresholds.

The Kopis BI team sat down with the Hawkers team for an in-depth discovery and analysis of their current system. There were several main issues they were facing.

1. The toast file synchronization process took a long time to run. Sometimes the follow-up data processes ran before all of the data was pulled down because multiple processes all ran on different schedules. This often caused them to miss data for previous days for some locations. Kopis found the cause of the slowness: the sync process read all data at the destination first (and there were years of files).Kopis rewrote the sync process to sync the last 7 days of files (all that’s available at the source) to an empty folder and then merged those files into the years of files. This dropped the sync process to 30 seconds rather than 30+ minutes. Kopis also moved all the steps into a single import process to eliminate any workflows from running out of order.
2. When stores were missed, the process to fix the data was manual. Hawkers had to login to the server, delete the bad day and reimport the data. Kopis enhanced the import process to always do the last 3 days, every day, so that if data was ever missed the system would automatically recover.
3. Hawkers needed an alert system. The only way to find out there was an issue was by checking the data in reporting, which was a menial, manual task. Kopis created multiple alerts for them:
   1. 1. A job failure notification in case the import process fails.
      2. A location alert to let them know whenever they get a day of data with less locations than the previous days.
      3. An alert to let them know when they need to MAP values from the POS system to a more user friendly value.
4. To add a new restaurant location, Hawkers had to modify the import process to hard code the new location code. Kopis modified the process to be location-driven based on their location dimension table. All they need to do now is add a location to that table to get the process to import it.
5. Hawkers had a bunch of complex logic in Microsoft PowerBI datasets and in an Excel file that they then used as their report. Kopis simplified the logic into a SQL server so the resulting tables had everything they needed already done.
6. Hawkers had no way to manipulate data after the fact (except changing it in Excel). Kopis created a process to auto-adjust some numbers based on thresholds in the database; likewise, Kopis created a way to add manual data adjustments in the database.

After implementing new data workflows, storage, reports and new alert, the Hawkers team is in a much better position to analyze and leverage data to continue growing their restaurants.

• Improved sync process to 30 seconds rather than 30+ minutes
• Setup several new alerts:
  1. Job failure notifications in case the import process flat out fails;
  2. Location alert to let them know whenever they get a day of data with less locations than the previous days; and
  3. An alert to let them know when they need to MAP values from the POS to a more user-friendly value.
• Implemented an auto-recover process
• Eliminated the need for hard coding by changing the process for adding new locations with a location dimension table
• Created a Powershell API process to sync GL data from their accounting system to the SQL Server