Trust and Validation

Trust

This is the most important thing that plays in Informatica MDM. That plays a vital role in MDM, like trust. So, different sources provide data for the same entity. Each system different trust score. There are multiple data systems. So the multiple data systems have multiple trust values, like in this case. In this picture. As we can say, see that the source system a orders and the source system. B marketing. So the name of the customer, that is, the dog McDoggle group, and the dog.

And see the double group. So this is the both names are the same, but the representation of that name is different, like here is, know the word using, and here the d word is used in the name So, the phone number is also different for both systems.

So, as we are now, we can identify data. So we are identifying data through the trust score. So the name in this system is 62, and the phone number trust value is 56.

And another system B, in this system, the name trust value is 71, and the phone trust value is 37. So what, Madams? Do it will just. It will compare the trust value from both systems, System A and System B, and just take the value that has the highest trust score.

So final consolidated record, like the golden data. So we are taking the name from System B, that have the highest trust score than System A,and we are taking the phone from the system, a sorry from the system. A. Because of the system, we have less trust value. Okay? So I think, understand how the MDM works in the back end, and how the golden data will come into the picture..

What is validation?

So validation rules, check data quality, and reduce the trust score. If the data does not meet a specific condition. So in validation, we are just applying some conditions and rules to our trust record. So when the data violated that rule. So it will degrade the trust record according to the data. Okay.

How does validation work?

So, how does validation work? If data fails validation? So it will downgrade our trust.

And we have multiple options. Like we are here, we are providing the downgrade percentage, like what percentage it will be downgraded if the data fails, and the validation rule.

And yeah, the other option is the reserve minimum trust. So we can also reserve minimum trust at that time, if the data fails the validation. So it will take the minimum trust.

Okay, so this is for all for the validation. And now we can move to the

Exact Match Vs Fuzzy Match

It very important thing in MDM, so we are matching our data. And now and then, we are merging our data to get our golden

Exact match Basically

It is used to identify duplicate cards and is later Consolidated. And this happened on the base object when we determine whether the record should be merged based on the exact match or Fuzzy match.

So exact match record will be identical. Same like same to same, then it will merge.

Fuzzy Match It will allow some variation in data like some misspelling, miss, spelling, mistake, and all it will allow to merge.

And when exact match when we do an exact match, there will be no tokenization.

Match tokenization will happen, and merging will happen through the tokenization.

So, like exact match, referring to it should be an exact match. So it is only a match when the data is looking like exact.

So it will not accept any tone tolerance for variation and work best for structured and standardized data, faster processing, because it should not generate a tokenization. So it is a faster process.

In the Fuzzy match.

We can, as you said, we can accept some minor variation in that data so

or determine that data, we apply some technique, some algorithm to match that Fuzzy data.

Okay? So here also are some characteristics for the Fuzzy match, like we can tolerate some spelling, abbreviations, and phonetics.

So, as this picture shows that when we are getting data from the register may job and then what should we do? What do we want to do, like Fuji or exempt? So we can check. If in Fuji we are generating team, then we are then we are matching that candidate according to the generated keys so generating, generating keys, is referring here the tokenization process, so it will generate token for the different different candidates, and when the tokens are similar, so it will match that candidate. In the Fuzzy match, compare the Record to match against the match candidate. It will compare, and then it will populate into the match table and we our process our end match.

Our Fuzzy match will end here, and exactly what it will do. It will go to the record And just comparing that record

from the base object table, and if the data should match, then it will populate the match table.

Good. So I think we observed. And I concluded some major topics in the debt presentation. I hope you guys understand this thing and enjoy the presentation. So thank you guys for, yes, sir, we can.

✅ Conclusion

In summary, trust and validation are the backbone of reliable master data in Informatica MDM. Trust scores help determine which source system’s data is most reliable, enabling the creation of a golden record that represents the most accurate and trusted version of an entity. Validation rules further enhance data quality by enforcing business conditions and penalizing invalid data through trust score downgrades.

The distinction between Exact Match and Fuzzy Match is also critical. While Exact Match ensures precision and works best with clean, standardized data, Fuzzy Match allows for minor variations and uses intelligent algorithms to identify and merge similar but non-identical records.

Together, these mechanisms ensure your MDM system produces high-quality, trusted, and unified data that supports better business decisions and operational efficiency

If you want to accelerate your career in the field of data, this is the right time to start. Join our Informatica MDM on Premises Training

Ready to take your security to the next level? Let’s talk!