Ten Reasons Why Building Customer Profiles On Your Own Is a Bad Idea

May 20, 2015

Customer Data Profile

Retailers and brands alike have stated that a top priority for 2015 is to build a “centralized customer data warehouse.” In our January survey of retail executives, 56% said creating a unified customer view across online and offline was a priority for 2015. There is a lot more to building customer profiles than meets the eye. This blog will give ten reasons why it may be better to turn to third party vendors to help you with this effort. If you are not careful, you end up replicating the efforts of a large engineering team working for ten years (as we have at AgilOne).

  1. Data integration is harder than it seems

    Your customer transaction and interaction data is probably sitting in many different silos, such as an ecommerce silo, store transaction silo, web activity silo, and so forth, throughout your organization. Bringing this customer data together in one place will require intimate knowledge of each of these source systems and their APIs. You will also need to know how to access, extract, normalize and transform data from each system.
  2. Integration is just the start: you need data cleansing too

    Data integration is only the start of your project. After getting the data in one place, you will have to normalize the data, deduplicate it and otherwise prepare it for analysis. If you are dealing with customer data, you have to worry about misspelled names, and you need to validate email and physical addresses. If people have moved, you will need to know about this and correct their address before spending the money on any direct mail. If you don’t cleanse your data, your segments will be off, your analysis will be wrong and you will waste money on personalization that will turn more customers off than please.
  3. Customer data needs to be accessible by other systems

    You will have to make the normalized customer profile accessible to outside systems such as customer relationship management software, email systems, store clienteling and the website. Maintaining a reliable API to the customer data profiles will add significantly to the work.
  4. You need an easy-to-use interface for business users

    People will want to look at customer data too. Customer-facing personnel may in fact be accessing customer profiles on a daily basis and marketers may look at segments and insights daily to configure campaigns. Unless you want to spend your days writing SQL queries for business users, you need to develop an easy-to-use graphical interface to give users in your organization direct access to the data.
  5. Building it once is not enough: you need near real-time updates

    Now that business users and marketing systems access customer data all the time, the data needs to be up to date at all times. That means that all steps of data integration, cleansing, and enhancement need to happen every day. So your integration project is not a one-time effort. Instead, you need to invest in writing the code or scripts to ensure that integration, deduplication, cleansing and enhancement happen on a daily basis at the very least.
  6. APIs tend to change over time

    Any data APIs you use to integrate data, either inbound or outbound, will change over time. Vendors of various database and marketing technologies change their software frequently. That means that you will have to rewrite any code you used to import or export that data each time a vendor makes a change!
  7. You may overlook security considerations

    When it comes to handling your customers’ personally identifiable information, you have to be very careful. You do not want that information to accidentally get lost or stolen. Even if you have the in-house expertise to properly protect this data, you will need to make significant investments to build data protection into your infrastructure.
  8. Building in error-detection and reliability is a lot of work

    There are many reasons why a specific data pull may fail on a given day. Troubleshooting failed data extractions can take a lot of time. You will need the infrastructure to give you an indication of what might be wrong. Therefore error detection and handling should be part of your data warehouse design.
  9. Vendors may have pre-existing integrations ready to go

    Perhaps the most important argument why NOT to build a customer data warehouse in-house is that there is no need for you to reinvent the wheel. Vendors who specialize in customer data warehousing, as a stand-alone solution or as part of a broader marketing cloud, will already have ready-to-go prebuilt connectors to many of your systems which can significantly speed up time to deployment and reduce costs.
  10. You may not be able to keep up with the industry

    Lastly, by working with many companies in your market, an outside vendor may have a lot of knowledge to add – about data integration specifically but also about the data itself. A vendor may bring industry benchmarks derived from customer data that will help your company find new business opportunities. And of course, the industry is not stagnant; an outside solution may be more future-proof as it will keep up with evolving requirements for future sources of data and future types of analysis that may emerge in the industry.

Overall, the main reason to consider using off-the-shelf software is the need to be agile, reaping the benefits of data integration more quickly while being able to focus on their key value proposition and strengths. Most companies will not derive any competitive advantage from building a scalable, reliable, accessible customer data warehouse from scratch.

If you are ready to consider alternatives to do-it-yourself watch an AgilOne demo or download our guide “Questions To Ask When Selecting Your Customer Data Platform.”