7 tips for data segmentation
- Cross-reference your business and consumer data. A recent promotion or job change can influence buying behavior on the consumer side.
- Make sure your segments are unique, useful and not overlapping. But, don’t define data to a level where it doesn’t make sense.
- Avoid segmenting to the nth degree. You could end up with segments that include just three customers. You could never justify marketing to them.
- Test, Test and Test some more. You should test a segment against a control group with every effort.
- Data should be easy to obtain and unambiguous. This is why the Recency, Frequency, Monetary (RFM) models are the most useful.
- You need to be able to Identify, Quantify, Locate and Target. Be sure you’re able to put any consumer into one of the segments, know how many exist in your files, find them in your market and can target them through their preferred channel.
- Don’t do any segmentation without good data. Your data quality is the key. This is the only way your segments are actionable.
Failure of CRM attributed to bad data
Is dirty data hindering your lead generation and CRM initiatives? Many organizations are using prospect data that is 20% to 40% wrong according to a recent ZoomInfo report. The Gartner Group estimates that more than 50% of CRM implementations are viewed as failures due to data integrity. The United States Post Office reports that 5.4 billion pieces of mail have address errors each year. Over 2 billion of these mailpieces go straight to the recycling bin. The wasted cost of printing, postage and production for this is huge, but the bigger cost is the steady erosion of an organization’s credibility among its customers. Data quality is crucial to your CRM initiative.
Always rent odd list quantities
Most list companies require a 5,000 or 10,000 name minimum when you order a list. Those are common test quantities. Some list managers will knock out a 5,000 name segment from a house file, and use it to fulfill all test orders for that quarter. That means those names get used a lot and suffer a higher percentage of ‘list fatigue’. So, when you place your list order, order an odd number of names, like 5,200 or 10,200. That way, you will be assured of a fresh ‘list pull’ with names that have not been oversold.