The president of Realtor.com announced yesterday that the company would be ending the test marketing of a controversial web application that ranked Realtors on performance metrics.

"Realtor.com is the consumer website for the National Association of Realtors. Its primary purpose is to connect Realtors with consumers in the digital world. This is a mandate we take seriously," wrote Realtor.com President Errol Samuelson in a letter posted to the site yesterday. "We have concluded the AgentMatch pilot and are working with Realtors across the country to evolve the program and find a better way to highlight their unique attributes."

The "Agent Match" app allowed visitors to Realtor.com to find active agents in neighborhoods they were house hunting in, displaying criteria such as the number of homes the agent had sold in the area in the past several months and how closely an agent’s listing price matched the sale price, using information pulled from local multiple listing services.  It was being tested in two markets, Las Vegas and northern Colorado.

Ever since the site announced they were developing the new product at the annual conference of the National Association of Realtors in November, a chorus of criticism from agents has swelled. Many agents said that the metrics used by the tool to rank agents were not reflective of the true level of service provided by a given agent, and worried that if its use became widespread, newer agents or those who work primarily with buyers would be particularly disadvantaged. Many felt particularly betrayed that Realtor.com, a for profit site which is operated by Move, Inc. but which licenses the name and logo of their national trade organization, was behind the idea.  

Samuelson specifically backed away from the primary features of the Agent Match app in describing the decision to end testing on it. 

"We learned a few things in this experiment. We learned that using an algorithm to ‘match’ consumers with Realtors is misguided. A computer cannot find the best Realtor for someone, just like a computer cannot place an accurate value on a home," he wrote, adding "We also learned just how much lies beneath metrics like days on market or list price to sale price ratio. Numbers don’t lie, but they don’t necessarily tell the whole story, either."