Predictive analytics offers a data-enabled view into what is likely to happen and is key to driving organizational change. In this article appearing in InformationWeek.com, Kris Hutton, Director of Product Management for ACL, tells us how we can get started with predictive analytics, and how we can avoid some pitfalls along the way.
Starting off with Predictive Analytics
It’s best to start small with predictive analytics, before expanding to other areas. First, start a proof-of-concept project – for this, you will initially need to target a highly visible business area. These business areas will likely be:
- Marketing Campaigns: Predict which customer segment will positively respond to a campaign or messaging.
- Sales: Predict customer lifetime value, understand what a buyer’s next best offer may be, and identify suggestive products.
- Customer Lifetime: Identify markers that indicate which customers will drop your product or churn altogether.
- Fraud: Predict which recipients of benefits or employees will commit fraud.
- Operations: Predict which employees will quit, allowing for proactively managed churn.
- Supply Chain or Vendors: Identifying less than ideal links in supply chain or model vendors prone to fraud.
Next, identify the data you need, taking care to ensure that it’s clean and plentiful. After this, build a model designed to predict your future target. At first, this model will require some trial-error but can be quickly altered to reach the necessary objectives.
Data quality is the biggest obstacle to getting value from predictive analytics. As such, cleaning and harmonizing of data is crucial to achieving any meaningful insight. Therefore, a good data management program to help manage access, cleansing, harmonizing and analysis should be employed.
There are legal and regulatory obligations surrounding data which need to be considered – these include PCI and industry-specific rules, and other common global standards.
When the analytics program reaches a higher level of maturity, having the general counsel and compliance officers consulting as stakeholders will ensure compliance.
Treating Data Right
Despite the big data craze and low cost of computing and storage, many organizations still fail to leverage their data, instead of treating it as “garbage”. Companies should, therefore, leverage data effectively with the use of predictive analytics to facilitate change and monetize data.
Click on the following link to view the original article: https://www.informationweek.com/big-data/big-data-analytics/predictive-analytics-101-getting-started-and-avoiding-mishaps–/a/d-id/1332891