1. IT and Marketing don’t talk enough
As we discussed in our blog “Ten Ways Marketing and Technology Teams can Collaborate to Harness Customer Data”, in many companies the marketing and technology teams don’t talk and collaborate enough. There is a lot of complementary skills and these two groups should talk early and often, especially when it comes to customer data projects.
2. Try to build everything in-house
There is a strong temptation to build a customer data warehouse in-house. We discussed ten reasons why building a customer data warehouse in-house may be a bad idea in a previous blog you can read here [link]. In summary, getting the data together one time may look easy, but it is very hard to build an infrastructure that allows for daily, automated updates of the data, also cleanses and de-duplicates the data and gives business users and marketing applications alike access through an easy to use user interface and programmatic API. If you are trying to do this all in-house you may be duplicating the work of a forty-person engineering team over ten years, without clear and direct competitive advantage.
3. Not planning for frequent updates
Historically, the requirements for a customer data warehouse were focused on analysis. Management and marketing alike wanted to see customer segments and customer insights quarterly. Today’s requirements are to personalize experiences across channels near real-time based on the data change everything. Now daily or even near real-time updates are required and the whole process of collecting, cleaning and analyzing data has to be more automated. If you personalize an experience based on data that is a quarter old, you are better off not personalizing at all.
4. Not providing self-service functionality to the business users
In a world where only occasional data analysis is needed, you don’t have to worry about a slick user interface that everyday marketers can use. However, in the modern customer centric organization a lot of different business users and customer facing personnel need to access customer data on a daily basis. Store clerks and customer service personnel may need to pull up customer profiles and marketing may want to look at customer reports and segments daily as well. Unless the technology team wants to spend their days building SQL queries for business users, it is important to think about a self-service infrastructure that gives marketers direct, on-demand access to customer data and insights.
5. Forget to provide information in context where and when needed
Today’s requirements to integrate customer data are bi-directional. There is a need to centralize customer data to properly analyze it. However, there is also a need to use customer data in third party applications, such as a customer relationship management application, to trigger personalized email campaigns or to personalize your website. All these applications will need API access to customer data and customers insights.
6. Try to boil the ocean
When it comes to customer data, it is tempting to truly collect ALL customer data in one place all at once, and business users will have a never-ending list of analysis they are interested in. Analysis of customer data is good, but acting on customer data is even better. It is often better to take an iterative approach and collect only SOME customer data initially, try to run some personalized campaigns using that data, and then go look for more data to refine your initial campaigns. For example, first collect online customer transactions. Second, add website browse data and email click data, and in a third phase add your store transactions and loyalty data. By taking an iterative approach you will learn a lot in the process, and make a lot more money, more quickly!