Count me in
By combining personal experience with progressive technological thinking, Herman Hollerith created an electric tabulating system that dramatically improved data processing and laid the foundation for modern computing.
8 min read
Each month, our Journeys of Innovation series tells the stories of inventors or entrepreneurs whose groundbreaking innovations have made a positive difference in the world. Delve into the past and learn about one of history’s great innovators.
The year is 1888, and the next U.S. census is two years away. Your former employer, the U.S. Census Bureau, announces that it’s hosting a competition. As an innovative statistician, you’re intrigued. The prize? Winning a lucrative contract to process and tabulate the data for the next census. All you have to do is prove you have the most efficient tabulation method for recording census data. This is convenient because you’ve already been testing your own electric tabulation system for several years and are convinced it’s the best out there. You’ve even filed a patent application for your invention. What have you got to lose?
The competition proved to be one of the most pivotal moments in winner Herman Hollerith’s career. A son of German immigrants, Hollerith was born in Buffalo, New York, on February 29, 1860. He was not a great student as a child, but he did manage to graduate from the Columbia University School of Mines with a mining degree in 1879 at age 19. Hollerith’s first job had nothing to do with what he studied in college. At the invitation of W.P. Trowbridge, one of his professors from Columbia, Hollerith started working as a statistician for the U.S. Census Bureau in 1880. Because it was a census year, he witnessed firsthand the slow and tedious process of recording thousands of data points on the country’s steadily growing population. Before Hollerith’s electric tabulating machine, census results were hand-counted and took years to process; the 1880 census results, for example, were not fully processed until 1887.
“It was a singular moment in the history of data processing, one historians could reasonably point to and say that things changed because of it. It stirred Hollerith’s imagination and ultimately his achievements.”
While working for the Census Bureau, Hollerith’s professional connections grew. He met John Shaw Billings, the director of the Division of Vital Statistics. Billings reportedly mentioned to Hollerith that “there ought to be some mechanical way of doing this job … a machine for doing the purely mechanical work of tabulating population and similar statistics.”
This conversation “was a singular moment in the history of data processing,” wrote historian James Cortada, “one historians could reasonably point to and say that things changed because of it. It stirred Hollerith’s imagination and ultimately his achievements.”
The intellectual fruits of this exchange, though, were not immediate. In fact, Hollerith stopped working for the Census Bureau shortly after the 1880 census (and before formulating his groundbreaking idea) to teach mechanical engineering at the Massachusetts Institute of Technology (MIT). It was during his short tenure at MIT that Hollerith formed his idea. Inspired by a comment from Billings about using something similar to the Jacquard loom, which utilized punch cards to make textiles, and by observing how punch cards were similarly used in train ticketing, he realized he could create an electric tabulating system by counting data from seemingly random—though completely specific—holes in punch cards.
Unique to Hollerith’s invention, the system was made of multiple parts—resulting in multiple patents. U.S. patent nos. 395,781; 395,782; and 395,783 comprised what he called the “Art of Compiling Statistics” and the “Apparatus for Compiling Statistics.” Initially filed in September 1884 and June 1887 and issued in January 1889, these patents described a punch-card tabulator that would record data through “mechanical means.”
In “An Electric Tabulating System,” published in the 1888-1889 volume of The School of Mines Quarterly, Hollerith wrote that “the records must be put in such shape that a machine could read them [through electromechanical sensing devices]. This is most readily done by punching holes in cards or strips of paper, which perforations can then be used to control circuits through electro-magnets operating counters, or sorting mechanism, or both combined.” Each hole represented a different data point—such as age, education, and birthplace—and the pattern of the holes in the card described the person being counted in the census. This system used only one card per person.
Hollerith first used his machine to count health statistics for the city of Baltimore and the state of New Jersey in 1887, just in time to fix any potential glitches and enter the Census Bureau’s competition to claim the 1890 census contract. In the competition, Hollerith competed against two other contestants to process specific data from the 1880 census and demonstrated that his machine would save the government considerable time and money. In each category, Hollerith had the fastest results: 72.5 hours for data capture (transcription) and 5.5 hours for tabulation. His competitors’ methods were roughly two times slower for transcription and 10 times slower for tabulation.
Due to his statistically significant results, the Census Bureau awarded Hollerith the contract to process and tabulate the 1890 census. This proved to be a wise choice, as he was able to process the census in under three years and $5 million below projected costs, which is equivalent to approximately $141 million today. Hollerith’s machine was also used to process the 1900 census.
In 1889, Hollerith’s system won a gold medal at the World’s Fair in Paris. The governments of Canada, Norway, and Austria also used it to count and process their 1891 censuses. Hollerith knew that the idea of compiling data through patterned hole punches could be used in other ways. Realizing he had created something that many businesses, including railroads and governments, had a need for, he started his own private venture in 1896 called the Tabulating Machine Company. This entity eventually merged with two others to form Computing-Tabulating-Recording (C-T-R), the precursor to International Business Machines (IBM). Hollerith worked for the company as a consulting engineer after the merger. In 1921, he retired to his farm in Maryland and spent his remaining years there. He died of a heart attack in 1929.
“A census is often spoken of as a photograph of the social and economic conditions of a people. The analogy can be made, not only with references to the results obtained, but also to the methods of obtaining these results.”
Throughout his lifetime, Hollerith patented a number of other inventions in addition to his electric tabulating system, including an “apparatus for perforating record cards” and a “card testing device.” He had a keen understanding of how important the census data was for the country’s future well-being. In addition to helping businesses decide where to open new stores and which products to offer, the federal government, local governments, real estate developers, and residents use census data to determine where to build new schools, hospitals, and roads; where more federal funding is needed; and how many seats each state will get in the House of Representatives.
In “An Electric Tabulating System,” Hollerith aptly described the census as “a photograph of the social and economic conditions of a people.” Just as a photograph captures a moment in time, a census captures the country’s current social and economic status. He also emphasized the importance of the methodology for gathering this information, which his system improved: “The analogy [of photography] can be made, not only with reference to the results obtained, but also to the methods of obtaining these results.” In other words, the tabulation process can be compared to the process of developing a photograph. As he explained, “the compilation of a census corresponds with the development of the photographic plate.”
Because of his progressive tabulation invention, Hollerith laid the groundwork for the modern census as we know it today. The U.S. government used versions of his technology to complete the census up until 1960, when computers took over the job—and technological progress driven by the changing needs and challenges of a decennial census continues even today. Using the Census Bureau’s current automated system, paper response forms from the 2020 census are expected to process at an average rate of approximately one form per 35 seconds.
This year’s census also marks a new milestone. For the first time ever, respondents can complete the census online or by phone, as well as by mail. For households that do not respond on their own, census takers will even be able to use an iPhone app to record results, thereby advancing the long line of census-related innovations since Hollerith's time.
Credits
Produced by the USPTO Office of the Chief Communications Officer. For feedback or questions, please contact inventorstories@uspto.gov.
Story by Lauren Emanuel. Portrait of Herman Hollerith in 1888 at top of this page courtesy of the Library of Congress. Photo on homepage of 1890 card puncher courtesy of U.S. Census Bureau. All other photos credited in captions. Additional contributions from Alexander Phillips, Eric Atkisson, Marie Ladino, Alexander Camarota, and the U.S. Census Bureau.
References
“1100100 and counting; IBM.” (2011). Economist, vol. 399, 67-69. Retrieved from https://search.proquest.com/docview/871400011?accountid=14753.
“$5,000,000 in 1890 → 2020 | Inflation Calculator.” (December 2019). Official Inflation Data, Alioth Finance. Retrieved from https://www.officialdata.org/us/inflation/1890?amount=5000000.
Blodgett, J.H., and Schultz, C.K. (1969). “Herman Hollerith: Data Processing Pioneer.” American Documentation, vol. 20, no. 3, 221. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=edb&AN=16977381&s….
Cortada, J.W. (1993). “Before the Computer: IBM, NCR, Burroughs, and Remington Rand and the Industry They Created, 1865-1956,” 47-48. Princeton, New Jersey: Princeton University Press.
Hollerith, H. (1888-89). “An Electric Tabulating System.” School of Mines Quarterly, vol. 10, 238-255. Retrieved from https://hdl.handle.net/2027/uc1.b2978976.
Hollerith, H. (1889). United States patent no. 395,781. Retrieved from http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u….
IBM. (2007). Frequently Asked Questions. Retrieved from https://www.ibm.com/ibm/history/documents/pdf/faq.pdf.
Reid-Green, K.S. (February 1989). “The History of Census Tabulation.” Scientific American, 98-103. Print.
Smithsonian National Museum of American History. (n.d.). Hollerith Tabulating Machine. Retrieved from https://americanhistory.si.edu/collections/search/object/nmah_694410.
U.S. Census Bureau. (n.d.). Herman Hollerith. Retrieved from https://www.census.gov/history/www/census_then_now/notable_alumni/herma….
U.S. Census Bureau. (n.d.). The Hollerith Machine. Retrieved from https://www.census.gov/history/www/innovations/technology/the_hollerith….
U.S. Census Bureau. (n.d.). Importance of the Data. Retrieved from https://2020census.gov/en/census-data.html.