Conducting an Entity Resolution Proof of Concept (PoC)
One of the best ways to determine whether to buy or build an entity resolution system is to assess the accuracy of the options you’re considering. Watch this video as Jeff Jonas talks about the steps he recommends you take in order to conduct an entity resolution proof of concept (PoC) for your organization.
At Senzing®, we have a one-day PoC program that allows you to test the accuracy of our entity resolution in about six hours! Whether you are assessing Senzing entity resolution, your own homegrown system, or any other entity resolution software, you should do the following:
- Perform an accuracy audit on your own data.
- Explore the results using tools that allow you to gain confidence in your results and the entity resolution engine’s accuracy overall.
- Compare scalability and cost differences of the options you’re considering by loading more data and seeing how each system performs. (Make sure you clearly understand how much data your entity resolution production system will need to ingest and how it will perform under those conditions.)
Read more in our Should You Buy Entity Resolution or Build It In-House blog.
Video Transcript
Timestamps
0:00 Intro
0:16 One-Day Entity Resolution Proof of Concept (PoC)
0:32 Download on Your Premises
0:55 Exploratory Data Analysis (EDA) Audit Tool
1:29 Scalability & Cost
1:43 Demonstrate Entity Resolution Accuracy and Scalability with Low Total Cost of Ownership (TCO)
When you’re getting ready to make your next entity resolution engine decision – buy [or] build – one of the important things to do is assess the accuracy of the different options.
0:16 One-Day Entity Resolution Proof of Concept (PoC)
What we have at Senzing to help speed this process up is the one-day PoC. That’s right, one day, because there’s a lot of other ways to do entity resolution PoCs with other methods that are going to take you weeks or months. We’ve collapsed that to a day.
0:32 Download on Your Premises
You’re going to download our software and we’ll watch over your shoulders by the way, just in case you want a little help. You’re going to download the software to your premises or to your cloud. You’re going to take your real data and you’re going to use a vertical slice, maybe everybody with the last name [starting with] A or maybe everybody from Washington, and you’re going to run it on Senzing. And then Senzing is going to produce a result of matches, possible matches and relationships.
0:55 Exploratory Data Analysis (EDA) Audit Tool
We then have in our open-source community, which you can find on Github, a suite of tools called Exploratory Data Analysis (EDA), and one of these tools is the audit tool, and you’re going to take Senzing’s results, and then you’re going to add results from your homegrown or other competing engine. Then you’re going to be able to get a report about all the things that are different between the two entity resolution methods, and you can audit those with these Exploratory Data Analysis tools. You can dig in and say why did that match? Why didn’t it match? And this is how organizations like yours can gain a lot of confidence in the accuracy of the engine.
1:29 Scalability & Cost
Now, after you’re done with an accuracy PoC, you’ll want to think about things like scalability, which is a different scale of record sets. Because, in an accuracy PoC, you’re going to have a million, or ten million records. Another thing you’ll want to be thinking about is cost.
1:43 Demonstrate Entity Resolution Accuracy and Scalability with Low Total Cost of Ownership (TCO)
And so, net net, we’re going to help you prove Senzing is accurate with the one-day PoC. You’ll also find that we scale to billions of records and thousands of transactions a second. And, we have a total cost of ownership footprint that is an order of magnitude less than anything you’re going to find on the market.
Come check us out at Senzing.com.