Case Studies
Data reconciliation for a Fortune 500 events team
We cut the events team's manual reconciliation between Salesforce, Cvent, and Workfront by 60 to 80 percent, worth $19,500 to $54,600 a year in labor.
Data reconciliation for a Fortune 500 events team
A Fortune 500 pharmaceutical company. The corporate events team plans and runs events, and its data lives in three systems: registrations in Salesforce, event management in Cvent, budget tracking in Workfront. None of the three talk to each other, so the team reconciled them by hand.
The problem
Every week, each person on the team spent two to five hours on this loop: download CSVs from each system, clean them up in Excel, match records across the three by hand, upload the results. Formats changed from export to export, so the cleanup steps had to be redone each time. The reporting that executives saw was assembled from those spreadsheets, which did nothing for anyone's confidence in the numbers.
The build
We built a file-based reconciliation platform. It works on the exports the team already pulls, so it needed no API access to any of the three systems.
- It reads each CSV and interprets the fields in context, so a renamed column or a changed export format does not break the match.
- It reconciles records across Salesforce, Cvent, and Workfront automatically.
- It scores its own confidence on each match and flags the low-confidence ones for a person to review. The team keeps the final say on anything unusual.
- It runs as a cloud file-processing service, which the client procured as data business processing rather than a software integration project. That kept it off the IT integration critical path.
The team spent about ten hours total on the project, mostly on validation calls. Their operations kept running while it was built.
The results
- Manual reconciliation dropped from 5 to 14 hours a week across the team to 1 to 3 hours. That is a 60 to 80 percent cut.
- Error reduction we estimated at 88 percent against industry automation benchmarks; we did not measure it directly. The confidence scoring and exception queue are what actually protect data quality.
- At a fully loaded labor rate, the savings run $19,500 to $54,600 a year against a $36,000 build.
The savings scale with team size. The build pays for itself at a four-person team:
| Team size | Annual savings | ROI vs $36K build | | --- | --- | --- | | 3 people | $15,600 | -57% | | 4 people | $20,800 | break-even | | 5 people | $26,000 | 39% | | 7 people | $36,400 | 101% |