What Your Wedding Guest List Can Teach You About Master Data Management.

Your wedding guest list may have more in common with a company database than you think.

Imagine you are getting married. You and your partner need to send invitations, so you start making a guest list.

You have your list.

Your partner has another.

Then your mother sends you a much longer list from her side of the family.

Now you have three lists and one simple task:

Create one correct list with everyone’s name and latest address.

Sounds easy, right?

It is not.

This is very close to a problem that companies deal with every day: Master Data Management, or MDM.

Step 1: Put the Lists Together

Your first move is probably to open Excel and put all three lists into one file.

Done.

Except it is not done.

You have only put all the messy data into one place.

One list says:

Rajesh Singh

Another says:

Uncle Raj

Your mother’s list may say:

Raj & Family

Before you can decide if these are the same person, you need to make the information easier to compare.

For example, one list may say:

21 Park St.

Another may say:

21 Park Street

One phone number may have dashes. Another may have no spaces or punctuation.

Putting information into the same format is called standardization.

This step makes it easier to find duplicate records.

You can also add a column that tells you where each record came from: your list, your partner’s list, or your mother’s list.

That small step can be useful later.

Step 2: Who Is the Real Uncle Raj?

Now comes the tricky part.

Your list has Uncle Raj.

Your mother’s list has Rajesh Singh.

Are they the same person?

Or are they two different people?

This is where match-and-merge logic comes in.

You can compare details such as:

• Date of birth
• Address
• Phone number
• Other details that identify the person

If the details match, you can decide that Uncle Raj and Rajesh Singh are the same person.

Instead of keeping two records, you keep one.

Your final guest list may say:

Rajesh Singh — 21 Park Street

You have now done a simple form of deduplication and entity resolution.

In simple terms, you found two records that belong to the same person and brought them together.

Step 3: Which Information Should Stay?

Now suppose Aunt Priya has two different addresses in your lists.

One is her old address.

The other is the address she gave you last week.

Which one should you use?

You choose the newer address.

In data management, this is called a survivorship rule.

A company may have several values for the same piece of information. It needs rules to decide which value should be used.

In this case, you used a simple recency rule:

The most recently given address wins.

That address becomes the one you use for the invitation.

Companies use similar rules when they need to choose the best value from several records.

Step 4: Sometimes, You Have to Call Mom

Now imagine your lists contain two people named Meera.

They have different last names.

They live in different cities.

Their contact details are also different.

No rule can tell you if they are the same person.

So what do you do?

You call your mom.

That phone call is a good example of what a data steward does.

Rules and software can solve many data problems. But when the information is unclear, a person may need to check the details and make the final decision.

Sometimes, the person who knows the family best is still the best person to ask.

What Happens When You Get It Wrong?

Now imagine you do not fix these problems.

You send Aunt Priya’s invitation to her old address.

She never gets it.

For a company, this could mean a customer does not receive an important message. A bill could go to an old address. A recall notice could fail to reach the right person.

You also keep two records for Uncle Raj.

At your wedding, you might reserve two seats for the same person.

For a company, this could mean sending two marketing messages to the same customer or keeping their loyalty points in two different accounts.

But the problem can become much more serious.

Imagine that the two Meeras are wrongly merged.

Information about one person could end up on the other person’s record.

In healthcare, this could mean a doctor sees the wrong information for a patient, such as the wrong allergy.

A small mistake in a guest list may seem harmless.

A similar mistake in a business system can have serious results.

So, What Is Master Data Management?

This brings us back to the wedding guest list.

Your three guest lists are similar to the different systems a company uses.

A business may have customer information coming from:

• Sales
• Its website
• Billing
• Customer support
• Other internal systems

Instead of a few hundred wedding guests, a company may have millions of customers, patients, products, or other records.

Master Data Management, or MDM, is the set of tools, rules, processes, and people used to bring this information together, clean it up, decide which records belong to the same person or thing, and create one trusted version of the data.

That trusted version is called a golden record.

In our wedding example, the golden record is the final guest list—the list you trust enough to send the invitations.

The Guest List Is Never Really Finished

There is one more thing your wedding guest list has in common with business data:

It keeps changing.

People move.

Phone numbers change.

People get married.

Customers change their addresses.

New records are added.

Old information becomes outdated.

And sometimes, another Uncle Raj appears.

That is why Master Data Management is not about making one clean list and forgetting about it.

It is an ongoing process of keeping important information accurate, consistent, and useful.

So the next time a company has your new address before you remember giving it to them, or sends you two copies of the same catalogue, think about your wedding guest list.

Somewhere behind the scenes, someone is making sure that the real Uncle Raj gets exactly one invitation, at the right address.

About the Author

Aditya Shauryavardhan Singh
BS in Computer Science, AUM, USA
Follow him on Linkedin