Importance of Data Quality & Data Governance for Sales Analytics
Importance of Data Quality & Data Governance for Sales Analytics
The strength of analytics insight is a direct function of the data quality and Sales Analytics is no exception to this. In fact, in sales the dependency on external factors for data collection is so high, whether its client, customers, channel partners, marketing, operations etc., that sometimes it is impossible to produce any valuable insight from sales data.
Many of you reading the above paragraph would conclude something like this – ‘That’s why I keep telling all this analytics isn’t relevant for sales, it’s the customer relationship that matters!”. Wait, if the data or information in the sales system is inaccurate, forget analytics, it impacts business performance and that hurts the organization and you. Organizations spend more resources and effort on sales for lower return on investment. Burdens of poor communication and wrong decision making are bonus. So poor data quality in sales is not an excuse but a reason to pursue sales analytics.
When it comes to data quality, inferior quality and arrangement of data makes it difficult to perform substantial analytics. Other times, data is usable but in functional or process silos. For example: Data of one particular channel is good.
In general data quality can be assessed using the following attributes:
- Accuracy
Consistency - Completeness
We are never going to get 100% good data, as a rule of thumb, to start with, if 66% of your data is error free, some useful insights and inference can be drawn out.
Signs of good data are reflected in the overall mindset and data governance associated with sales. For example, when key data domains have been defined and created central data repositories and further when integrated, accurate, and common data is maintained in a central warehouse with an eye to look for new metrics and data, the quality of data is likely to automatically improve.
Tactically, there is nothing like collecting the data right, first time. Rules & tools for automating the data collection, preventing duplicates, ensuring quality of meta data, assigning accountability, creating data management hierarchies, etc. are some best practices.
However all of this depends on what type of data we are dealing with. Imaginery or audatory data handling are quite different and AI algorithms are available that can be deployed and used for data quality validation or even improvement of quality.
Tools are available for performing the following operations to improve data quality:
- Cleaning remove incorrect or duplicates or modify the values
- Parsing and standardization break the data into components and bring them to a unified format.
- Matching integrate or merge similar data
Profiling to score the quality using patterns, exceptions, and anomalies - Monitoring to control data quality.
- Enrichment to use external data to improve internal data quality
Here are some of the popular Cleaning tools :
OpenRefine (Google Owned), Trifacta Wrangler, Drake & TIBCO Clarity