The Difference is Speed

Should I scan for my paper project? How do I know? What are the benefits, pitfalls or things to think about? We get this question a lot. Use the study below to help you decide if you should scan or not scan.

Numbers Dont Lie

Your digitization project doesn’t need to be complicated. The information below is based on a typical project for us – 80,000 documents. If you have more or less, don’t feel that this information doesn’t not relate to you. Use the numbers at the bottom to help you perform the math on your particular project.

Which method have you used in the past? Which method is more accurate or faster? Feel free to send us comments from your experiences. We followed SPC when they performed a study to test both methods side by side.

Study:

The Republic of Vanuatu is a small island near Fiji. They have a population around 250,000 and they needed a quicker way to process their census information. Two different methods were used to test which is better: Scanning with OCR – using software that extracted data from census forms–AND– Manual Entry – using teams of people typing in data from census forms. SPC measured results on the two following factors: Accuracy and Speed. Both approaches were tested in a real‐life scenario.

The data was extracted from the following form:

The Graphs below show the results from the study.

 

Accuracy – Since the numbers were close on Accuracy, the data was classified as insignificant.

Time – Since the project finished 10 months faster than Manual Entry, the data was classified as significant.

 

Numbers from the project:

  • # forms = 80,000
  • Avg scan time per form = 5 seconds
  • Total scan time = 111 hours -or- 18 days (6h per day)
  • Number of scanners = 2
  • Scan time with 2 scanners = 9 days (2 weeks)
  • Avg verify time per form = 2 minutes
  • Total verify time = 2666 hours -or- 333 days (8h per day)
  • Number of verifiers = 10
  • Verify time with 10 verifiers = 33 days (7 weeks)

Conclusion:

This unbiased case highlighted that scanning is a relatively inexpensive way (relative to the total census/survey budget) to considerably speed up data processing, and free up human resources. Looking at a form count of 80,000, we can expect a full set of data to be available 10 months earlier than via manual data entry, assisting a much faster turnaround.