Evaluating the Quality of Your Data Provider: It's Not Just About Data Accuracy

Forrester Research recently recognized Lattice as a leader of the Predictive Marketing Analytics for B2B Marketers space. In its comprehensive report, it pointed out

“Data accuracy, quality, and capability claims are challenging to validate. Understanding the differences between Random Forest, nearest neighbor, Jaccard similarity, and k-Means algorithms is child’s play, compared with understanding the differences between vendors’ data relationships. Determining from whom suppliers buy data, who’s got better access to relevant data as a result, and who’s curating a database with the right volume and coverage for a marketer’s needs is confusing at best and a mystery for most. And validating vendors’ claims about data quality and value is a challenge for 64% of our reference customers.” -  The Forrester Wave™: Predictive Marketing Analytics for B2B Marketers, Q2 2017

If you think the quality of a data provider is based solely on data accuracy think again. What most customers also find challenging is matching their own lead and account data with that supplied by third party providers. For instance, you may have lots of leads in your database that have provided you with a public domain email address (e.g. Gmail, Yahoo, etc.). This means that if you want to enrich that lead with third party data, you need to match on company name and other relevant fields. This is why as we’ve developed our Lattice Data Cloud™, we focus not just on “data accuracy” (e.g. do we have the right location for a company, right number of employees, accurate picture of what technologies they’re using) but also on  “match accuracy” – so we can ensure that you are able to match (and thus enrich) as many leads and accounts in your database with the information and insights contained within our Data Cloud.

For example, when marketers have thousands of inbound leads, how would they identify if we even matched to the right account? No matter how impressive the data accuracy numbers are, matching trumps everything. We solve this challenge with our intelligent Lattice Decision Graph™ that finds the best match. At Lattice, we can measure both match and data accuracy, from the start, by matching our data with customers’ CRM.

Now, the million-dollar question most customers have is – How do we know if we matched correctly?

It starts with what we compare against. We all know there is no source of truth for B2B data, so, we created it. We have manually verified business datasets of a substantial number of businesses to make sure we have statistical significance on all the numbers we see. We call them the Golden Datasets. We have curated these datasets across large, mid-market and SMBs. We cover large (Forbes Global 2000) and mid-market companies globally in our Golden Datasets. However, 95% of what we use in SMBs are US based businesses.

We take all these datasets, and run them through our platform making sure we A/B test using multiple input scenarios that our customers use for matching. For example – some customers may want to match on domain and some may want to match on name and location of the account. In location, we have a varying level of granularity from country, state, city, right down to the postal codes. All these scenarios are tested so that we are consistent in all forms of integration with our customers’ CRM. This whole exercise is automated so Lattice Data Cloud™,  always has up-to-date information.


In Large businesses segment, we see our match accuracy is >95%. This is considering we use just the domain or Company Name with Country at the most minimum level to match. As we move down the company size, we see the match accuracies decrease.

Data management was another issue Forrester Research highlighted.

Dedicated processes for managing customer data. Sure, you don’t have to have perfectly clean, perfectly integrated data to get started with PMA. However, predictions are more accurate and more revealing when your data is more complete. Survey respondents agree: 54% said that poor data quality and accessibility substantially limits their ability to demonstrate results quickly and effectively. And first-rate data management and governance processes will improve or eliminate any data deficits that you have.”  -  The Forrester Wave™: Predictive Marketing Analytics for B2B Marketers, Q2 2017

Lattice fully understands the importance of data management best practices and everyone in our company plays a part in improving the quality, not just engineering. We have people from our sales, customer success and marketing teams working closely with our engineering teams to brainstorm how we can improve the data accuracy and match accuracy for our customers. We do not stop with the numbers, but take samples out of the false positives and try to identify root causes. We try to solve them algorithmically for long term values and not just have some quick data hacks.

Data accuracy is a tough nut to crack, and is an evolving effort within the organization. When marketers out there are auditing various vendors, we urge them to do this simple match test and then move on to data accuracy. Although the match accuracies hugely rely on the quality of CRM data, we do our best to give the right match.

If you are a Lattice customer, we would love to have you try this through our Company Profile feature. If you are a prospect, please reach out to us for a demo.

Written by

Afroz Mohammad
June 27, 2017

Comments