Crowdsourcing Data Not Without Major Pitfalls

Mark Davidson, Director of Data Operations for SaveOnBrew, the world's only beer price-search engine, discusses five pitfalls of crowd sourcing as a method of data aggregation.

  • Share on TwitterShare on FacebookShare on Google+Share on LinkedInEmail a friend
When you’re dealing with a data set of 300,000 records every day, you need to be able to manage the collection of that data.

Houston, Tx (PRWEB) June 11, 2012

The popular practice of “Crowdsourcing” occurs when a company makes an appeal to consumers to add data to their sites. One example of crowdsourcing would be Wikipedia, which relies on contributors to update its files. Another example is a website like Crowdspring, which lets freelance designers submit logo and design ideas for businesses. Pabst Blue Ribbon is dabbled with crowdsourced ownership for anyone who contributes toward the $300 Million asking price for the company. While crowdsourcing is a quick and easy way to bring in massive amounts of data or opinions, not everyone is a fan.

“Crowd sourcing is not the end-all be-all for companies trying to wrangle huge consumer-based data sets,” says Mark Davidson, Director of Data Operations for SaveOnBrew.Com, “especially data sets geared towards consumer goods.”

Davidson should know. SaveOnBrew collects data from more than 50,000 retailers across the country and uses the data to fuel their beer price-search engine. "As the data set broadens, it becomes less practical to rely on a crowd for accurate, timely data."

“Our original design spec for SaveOnBrew was a combination of data-scraping and crowd-sourcing. We quickly ran into a host of crowd-related problems and came to the realization that, for our needs, crowdsourcing just simply was not the answer.”

He shares five potential pitfalls to crowd-sourced data.

1) Damaged Reputation: “At SaveOnBrew, we’re only as good as our data. If it isn’t accurate, we’ll lose our core audience. Unfortunately, a crowd can be unpredictable. If our customers lose faith in our data and we become known as being inaccurate, they’ll leave and not come back.”

2) Unmanageable Labor. By its very nature, the crowd is made up of those most willing to help the project succeed. But at the end of the day, you can’t correct, discipline or fire them. They are inherently unmanageable. “When you’re dealing with a data set of 300,000 records every day, you need to be able to manage the collection of that data. By shifting from a crowd to dedicated data-scrapers, we build in the ability to manage our work force, including the ability to reward and discipline them as employees of the company.”

3) Sabotage. If data is your claim-to-fame, then crowd-sourcing gives your competition an “in” to sabotage that data. “We have procedures in place where we audit our data for accuracy. We literally go back and check the data entry work,” Davidson explains. Crowd sourcing tends to rely on input directly into a production database, without the necessary checks and balances. “We wouldn’t want our competition to enter inaccurate beer deals that might alienate our customers and business partners.”

4) Rewards That Lose Their Luster. Davidson explains, "Once you start rewarding, you have to be prepared to do it for the long run. If you decide to incentivize the crowd, they will grow to expect it, and may feel resentment when the incentive changes or, in the case of a drawing, they are not the person chosen for the reward. Instead of building loyal followers, you might be building a crowd full of resentful people who eventually just move on."

5) Difficult to Predict. According to Davidson, it's difficult to predict participation which leads to difficulty in business planning. "Our operation runs six days per week, and we need to have the deals up early in the day. Our tests showed that, no matter how we incented our crowd, some of them just simply wouldn’t participate to the level we needed them to participate. Back when we were a start-up, our crowd size was way too small to support the needs to the business. They simply were not dependable, and that left us scrambling for data."

But Davidson does see some advantages to the approach. "We actually do plan to come back to crowdsourcing, which we refer to internally as 'the street team.' We see that we have an opportunity to build a culture where participation is rewarded with both physical and social rewards. We want to make sure, though, that we build in checks and balances to ensure accurate data delivered in a timely manner."

About SaveOnBrew.Com: Founded in 2010 to help thirsty beer drinkers across the United States find the lowest advertised prices for one of the world’s most popular beverages.


Contact