Blog
Contact Us
.png)
Anyone who has ever owned a home knows the feeling of moving into a clean, empty space. Initially, everything is in order, and everything has a place. Over time, however, things begin to change… clutter starts to collect, carpets begin to wear, small things start to break, paint starts to peel, etc. Left its own devices, a home that is not maintained will eventually collapse and become part of the dirt as it is weathered and worn.
Home ownership has a number of striking parallels to CRM custodianship. Initially, we architect things exactly the way we like them and we load our clean, fresh data into the system so we can begin using it. Over time, despite our best intentions, the quality of the data starts to diminish…duplicates emerge, incomplete records are entered, new attributes are added without backfilling into existing records, other functional areas make changes to the system that create inconsistency. As the pattern continues, the usefulness of the system begins to decline because the outputs are unreliable. Eventually, user adoption is negatively impacted, collaboration is lost, and “the system” is blamed as a failure.
There has been much written on the topic of how much bad data can cost an organization. I recently did a Google search on the “Cost of bad CRM data” and received almost 2.5 million results. I have not read them all, but I know that this is a significant issue, but one that can be managed. The cost can be illustrated by asking questions like:
- How much time is spent following up on unqualified leads?
- How much marketing effort is spent marketing to stale contacts?
- What is the bounce-back rate for campaigns?
- How much time is spent trying to segment data for marketing?
- How do you reconcile duplicate records?
- Can you accurately score your leads?
- Is there really a single view of the customer?
- How much “automation” are you able to get from your sales automation tool as result of bad data?
One rule of thumb that we often use to illustrate the cost of bad CRM data is the 1:10:100 rule. This rule proposes that for every dollar spent on proactive data maintenance would equate to $10 spent to reactively clean up. To ignore the problem and continuously work around dirty data bumps that number up another factor of ten to $100. In other words, prevention is key and procrastination is very costly. Furthermore, we know that, on average, up to 25% of personal information (phone numbers, email addresses, etc) go stale on an annual basis. Without continual updates, the cost of doing nothing is high.
The good news is that there is a way to keep your cloud based CRM system out of the dirt through a well planned and managed data maintenance strategy. This critical step can make the difference between long-term CRM success and short-term fanfare.
Governance
The first step in a data management process is governance. The role of governance is to define for your organization the standards that will be required of your data. This includes the creation of business rules and definitions that will be used to cleanse and monitor results. For example, the governance group shall define:
- How will data quality be defined?
- What criteria will be used to define a duplicate contact?
- What distinguishes a lead from a contact?
- What level of activity (or inactivity) will be used to define whether a record is stale?
- What is the strategy to deal with stale records?
- What defines a complete record?
- What type of data enrichment processes or technologies will be used?
- What is an acceptable bounce back rate?
- How will leads be scored?
- What are the acceptable thresholds of duplicates as a % of the overall database?
- How will data be segmented?
- What is an acceptable rate of incomplete records?
- What is the threshold for timeliness?
These sample questions provide some direction as to the type of governance policies that need to be in place to align the data strategy with the business goals, set benchmarks for performance, and create standardization.
Analysis/Profiling
Prior to building a house, an architect creates a detailed blueprint of the structure. Unfortunately, most data quality projects don’t take this same approach to building better data. Data profiling helps you determine the current state of your data and enables the discovery of issues. These issues might be related to structure, completeness, redundancy, relationships, etc. By understanding the strengths and weaknesses of your data, particularly as they relate to the metrics that have been established by the governing body, profiling gives you a starting point for making substantial improvements.
As part of profiling, it is helpful to blueprint your CRM system to identify trends and high level inconsistencies. The longer the system has been in use, the more likely it is that there are data discrepancies that have gone unnoticed over time. These issues may range from major architectural problems to invalid pick-list entries or orphaned data element.
The ability to profile data and conduct a through analysis is as much related to the experience and ability of the analyst as it is the tools that are used. Each situation and strategy requires careful consideration so be cautious of any silver bullets in this area.
Cleansing
Once patterns have been identified, the next step is to prioritize and address the highest impact issues. On an initial pass, it may be necessary to assess the dependencies that exist between fixes and quantify the value behind addressing individual problems before you dive in. Every decision has an opportunity cost associated with it, so it makes sense to make sure you are getting the highest value for your efforts.
Data cleansing is a general term merging duplicate records, purging/archiving outdated or invalid data, and generating exception files for those records that require some manual inspection. Overall, this process is akin to housecleaning in that, the more often you do it, the less painful it is. If you are the type that procrastinates all year until it is time for Spring cleaning, chances are it takes significant time and energy to get through the entire house. However, if you develop a disciplined pattern of reviewing and cleansing on an ongoing basis, the process becomes much more manageable (and less disruptive to end users).
I alluded to the fact that some cleanup is manual. This is a fact that is most commonly seen for de-duplication of data. Two contacts exist with the same name, address and phone number; which is the “master”? Sometimes these cases require manual inspection and such an effort should be built into your plan and governance policy.
Integration
One of the best ways to maintain clean data after an initial clean up is a sound integration and automation strategy. Integration to a back end system helps to define and maintain integrity, uniqueness, and completeness of data. The fact that some data records are often maintained in multiple systems is useful in establishing checks and balances, but be sure to clearly identify the system of record for all data components.
Common examples of data integration include product master integration or integration with an external marketing automation tool. Both of these prove to enhance the end-user experience by establishing a single access point for all customer data.
Enrichment
Data enrichment extends data quality and integration by supplementing existing data with additional information from external sources. The power of web services makes enrichment easy and establishes a synergy between your own organizations data and that of a third party. For example, D&B, Hoovers and Jigsaw all offer data solutions that intelligently synch with your CRM solution to provide users with IP that they don’t have to source!
Data enrichment represents the best of both worlds when it comes to CRM and web services. However, too much of a good thing can have a price. It is the role of the governance team to provide direction to the organization as to which external data sources will be used for enrichment, how they will be used, and who is authorized to use them. A sound strategy will help establish meaningful customer interactions and prevent abuse of your customers’ information.
Monitoring
Data quality issues are not solved once the system is set up. Just like owning a home, it is necessary to continually observe changes, identify needs, and prioritize projects. Maintaining high quality data requires diligence to prevent the atrophy of your solution and your end-user confidence.
The process of monitoring data should be supported by a series of key dashboards that support the data quality metrics of the organization. In addition to these know metrics, the work to establish new standards should persist as changes to the CRM strategy require. The house you move into today may not support your needs a few years from now, so it makes sense to keep your eyes open for changes that will impact your strategy, impact your data, and impact your approach to ensuring that your organization can successfully transform that data into information and knowledge.
