Retailers today are responsible for a significant amount of data. As a customer, I expect that the data I provide to a retailer is handled properly following the Payment Card Industry Data Security Standard (PCI DSS), General Data Protection Regulation (GDPR) and other compliance guidelines.
I also expect that the data I give to a retailer is being leveraged in a way that improves my shopping experience. For example, I want better recommendations given my purchase history, my “likes,” and my browsing. I expect the retailer to know my email and payment information for a single click checkout. These are small and trivial frictions that can be eliminated, as long as the data is handled properly with the right permissions. And the capture and analysis of big data (general information aggregated from all customers) is an opportunity for the retailer to fine-tune the overall customer experience.
However much of the data collected goes unused. This occurs because the infrastructure within an organization is unable to make the data accessible or searchable. It can’t be used to improve decision-making across the retail value chain. The unused data comes from many sources: mobile devices, digital and physical store shopping, and IoT. This means vast amounts of data points are ripe for collection. But the investment and modernization required to effectively manage and leverage the data hasn’t caught up. Today’s barriers to gaining value from data are:
- Data silos
- Incongruent data types
- Performance constraints
- Rising costs
A healthy data environment is one that enables value to be generated by moving the organization from “what happened?” to predictive and action-oriented insights: “what will happen”? This then leads the question “what should I do?” in which predictions can be automated, moving from insight to action, and generating new value.
Modern practices allow data to be more efficiently and effectively leveraged in a flexible, extensible and secure way. Deriving value from your data also means that your organization has achieved:
- Data integration
- Support for diverse data models
- Unlimited scale
- Fully-managed infrastructure
- Lower total cost of ownership
Chart: Ideal characteristics of a data ecosystem
Barriers to gaining value from data
Deriving value from your data
|Data silos||Data integration|
|Incongruent data types||Support for diverse data models|
|Rising costs||Fully-managed infrastructure|
|Lower total cost of ownership (TCO)|
In conjunction with leveraging new technology capabilities such as the cloud and AI, organizations must put effort into effective data management. In a way, it should be a pre-requisite for modernization and/or transformation efforts.
A data management model consists of these states: ingest, prepare, store, analyze and action. Data can be categorized into three types of data sources: purchased, public and proprietary data. In legacy environments, data is often siloed and constrained by a variety of factors such as channel, system or by utility. A major goal should be to remove these traditional constraints so that the right data can be leveraged across an organization at scale, securely. And nurturing a healthy data environment is a competitive advantage, especially as an organization’s data science capabilities grow.
Great data management provides a significant strategic advantage and enables brand differentiation when serving customers. In retail, it doesn’t matter if you are data rich or data poor. If you can’t action and operationalize the insights to power and automate decision making, your data management strategy needs to be reexamined.
Recommended next steps
- For more in-depth considerations regarding handling and processing data read Data Management in Retail.
- To apply data management in context, read Inventory optimization through SKU Assortment + machine learning use case that applies this information to a real retail challenge that is solved by data and machine learning capabilities in the cloud.
Source: Azure Blog Feed