The foundation of any successful predictive analytics program is built on consistent and correct data.
Don’t Let Your Dirty Data Get You Down
Aug 28 2017
“Dirty data” is a slang term for raw data, such as data that differs in format, resulting in duplicates, includes special characters or HTML tags, or has empty fields or errors, before undergoing an enrichment process. Consider a common example: The good old U.S. of A:
United States of America, United States, U.S.A., USA, U.S., US, 100-United States of America.
There are 196 countries in the world, and we refer to a number of them in multiple ways. If your predictive analytics program integrates feeds from multiple data sources, you may find the globe suddenly expanding by a hundred extra countries.
Date formats are not universal. One thousand can be entered like this: 1,000 or like this: 1000. And then there’s human error. Adding an extra “0” once in sales data can throw the whole curve.
On average, 24% of enterprise data is dirty and unusable. Gaining the true insight to make quality decisions relies high quality data. Manually reviewing data tables is extremely time consuming, can cost you up to $83 for every 100 records in a database.
Meanwhile the relevancy of your data is steadily slipping away. Never-mind being agile, you’ve got another 100,000 tables to go.
Datatron provides a better way. Our AI data cleansing processes engages the five deep learning models represented in the following graphic to address duplication, outlier removal (for more representative metrics), identity resolution, enrichment (rectify incomplete entries), and subsidiaries (recategorize with mother company).
Access to big data does not automatically increase your marketing prowess. Analytics based on accurate, timely, and complete data support the optimization of your marketing plans and the identification of the next best action. Effective and automatic data cleansing technology is the first step in making your data work for you.
Take the example of Talkdesk, a cloud-based software provider for call centers, headquartered in San Francisco. Talkdesk has over 300,000 records in its database, and adds 3,000 records every month.
Upon the discovery that up to 40% of its marketing and sales data was dirty and unusable, Talkdesk engaged Datatron. The company had used an array of different providers over the years, but still had not gained the insight to truly maximize marketing decisions.
Through Datatron’s automated AI data cleansing process and improved analytics, Talkdesk was able to immediately secure new marketing gains:
  • 11.5% increase in sales and marketing leads
  • 15% increase in the size of sales based on improved forecasting
  • 11% increase in deal velocity leading to shorter sales cycles
  • 25% increase in efficiency in the selling time of sales team
  • 27% improvement in overall data quality
According to Mike Leeds, Talkdesk Senior Director of Sales Operations, “The power of using Datatron AI surprised us with how much they made an impact on areas we were not expecting – like cost savings within our SDR team. And the confidence and trust our sales team has in the data going forward is a huge upside – for both our sales and marketing efforts as well as improved morale. Ongoing cleansing of our data 24/7 is how we do business.”
Don’t let your dirty data get you down – harness the true potential of your sales and marketing staff through improving the quality of your data through Datatron’s automated AI data cleansing capabilities.
Filed under Datatron Platform | Tagged: Data Cleansing Deep Learning Predictive Analytics Artificial Intelligence