Data enrichment part VIII: implementation tips
This is part 8—and the last—of our blog series on data enrichment. If you missed the first few, you can catch up starting with Introducing the Data Enrichment 101 Blog Series.
Now that you’ve have picked the best data providers for your needs, you’re ready to put those services into production. Here are some tips and things to consider when deploying an enrichment data service into production.
Batch or Continuous Process
You can implement enrichment either as a batch process or a continuous process. Batch process is quicker to implement since it’s usually a manual process. This typically involves simply extracting a file of data to send to your data provider for enrichment. The hard part is the manual process of incorporating the results back into your database. Most people just dump the data into custom data fields, which we don’t recommend. More on this topic later.
A continuous process means either an automated standalone enrichment process, or having the enrichment steps incorporated into other automated processes like list loading, lead routing, or segmentation. This means utilizing data orchestration technologies like Openprise. A combination of these may be the most practical approach. Integrating enrichment steps into other automated processes ensures the best data is available to support those processes when needed, thus optimizing the performance of those processes. Having a standalone process to maintain otherwise inactive records maybe a good practice to ensure no leads go stale for a long period of time.
If you chose to do the batch process, make sure you run the batch process in short enough frequencies to balance data quality vs. available budget vs. resource availability. Remember that the longer you go between batches, the bigger the reconciliation job gets.
Pre-Clean the Data
It’s highly recommended that you pre-clean the data before sending it out for enrichment. Pre-cleaning the data can drastically improve the match rate, thus maximizing the effectiveness and ROI of your data enrichment investment. Pre-cleaning involves:
- Correcting obvious mistakes like email syntax error and invalid domain suffix.
- Filling in blanks that can be inferred from other pieces of data. F, for example, filling in city and state if ZIP code data is available.
- Transforming to a data standard that your data provider prefers to maximize match rate.
- Cleaning up formatting like phone number, ZIP code, and address.
Some data providers offer professional services to clean the data first before enrichment. This is a viable option for batch process. For continuous processing, the pre-cleaning needs to be automated with technology.
Data provider’s ability to use dirty data as-is varies, so it’s best to not leave this to chance and do your own pre-cleaning.
Reconcile New Data Immediately
Perhaps the biggest mistake when it comes to data enrichment is not reconciling the new data into your primary data fields immediately after data acquisition. Most people simply keep the new data in custom data fields. This is a waste of money because:
- Your sales team won’t be looking at this data unless the primary data fields aren’t working.
- Your marketing automation platform or other marketing technologies won’t be able to leverage these custom data fields.
- You’ will end up with multiple sets of information that may be different and become impossible to reconcile later once the context is lost and data of different vintage is mixed together.
Simply put, if you don’t not reconcile your new data immediately, you’re are just throwing your money away.
Come up with the appropriate reconciliation logic about which data fields to keep, which ones to throw away, and when to overwrite. Reconcile the new data into your primary data fields. No reconciliation logic will ever be right 100% of the time, but having one that is 85% right is much better than not reconciling at all.
Once again, for batch processing, this reconciliation task can be done manually. For a continuous process, this needs to be automated.
Normalize and Segment New Data
Depending on your data provider’s flexibility in delivery options, the data you get back may require further processing to meet your data standards. Common examples of these tasks include:
- The data you get back may not be in the right format, so you’ll have to transform it. For example, new data says country is “USA” but your standard is “United States”.
- Instead of receiving the data in your preferred standard, you may just get all the different versions, such as “US”, “USA”, “United States”, so you’ll need to pick the version you want.
- Data fields like industry, number of employees, and annual revenue are generically set by the data provider, which may not be the same way you want to segment your data. For example, the “number of employees” data from the data provider may be “50-100”, but your segmentations are “1-25”, ”26-200”, “201-1000”, etc. You need to map these generic segments to your specific segments.
- Data fields like job function and job level are also generic segments that may not match your needs. You may need to do your own segmentation based on other data fields like job title.
The Role of Technology
Automation technology like the Openprise Data Orchestration Platform should be part of your data enrichment strategy, especially if your database is large, or if you enrich frequently or continuously. Automation technologies improve the outcome of your enrichment process by:
- Pre-cleaning data before it’s sent for enrichment.
- Processing the new data, post enrichment, to ensure compliance with to your data standards and segmentation.
- Reconciling new data into primary data fields.
- Integrating new data into your CRM, marketing automation, support, or other databases.
- Integrating data enrichment in-line within processes such as list loading and lead routing.
Separate your data provider from your automation technology. Changing data providers is easy, but changing automation technologies is hard. Don’t end up being stuck with a data provider you no longer feel is best fit, but can’t swap out because it’ is coupled with your automation technology.
We hope you found this blog series helpful. Data enrichment, when done right, can add tremendous value to your account and contact database. Here are some of the key points worth reiterating:
- It’s rare that one data provider can meet all your data needs. Consider a multi-sourcing enrichment strategy.
- Understanding your own needs and requirements is the key to selecting the right data providers for you.
- Enrichment should be a continuous process and it should be automated. The extra time and effort involved in automating the process will easily pay for itself within a year.
- Reconcile the new data immediately. Data isn’t like red wine. It doesn’t get better with age. It’s more like ripe bananas that start to go bad the moment you acquire them.
- There’s a wide variety of data providers out there and affordable automation technologies are readily available, so there really isn’t any reason to compromise, whatever your budget is.