How SAP Wields Its Data Lake

Emerging tech helps merge data sets more efficiently.

An interview with:
Mika Yamamoto
chief digital marketing officer
SAP

Marketers constantly seek more efficient ways of storing and utilizing their data. As the amount of data that marketers rely on ballooned, a new piece of jargon arose called data lakes, which simultaneously store many types of data in their native format and allow people to merge disparate data sets more easily. Mika Yamamoto, chief digital marketing officer at SAP, spoke to eMarketer’s Ross Benes about how the enterprise software firm uses its data lake.

eMarketer:

How do you approach data management for marketing purposes?

Mika Yamamoto:

We're looking a lot at making sure our data is most accessible. We're placing our data in a data lake just to drive accessibility to access all the data we need to be able to make sure we can drive that personalized experience.

eMarketer:

How long have you been using a data lake to store your data?

Mika Yamamoto:

It's actually really new. From a marketing standpoint, we've started to explore the notion of a data lake this year, which then opens up our ability to drive things like multitouch attribution to see what interactions actually do drive the best the customer outcomes.

eMarketer:

Do you use a vendor for this?

Mika Yamamoto:

A lot of companies do use a vendor. We happen to be a database company as well, so we’re using our own product.

eMarketer:

Before you began using a data lake, how did you manage your data?

Mika Yamamoto:

Different types of data were housed in their own platforms. We had to pull a lot of information manually and it took time to relate two different sources of information.

eMarketer:

What did moving to a data lake accomplish?

Mika Yamamoto:

There used to be a time lag between when we needed data and when we could access it. The other effect is that we can now take different data sources and link them together in one instance for a specific report.

eMarketer:

What all goes into building a data lake?

Mika Yamamoto:

It's identifying all the sources of information. It's being able to link all the sources of data to make sure that it's usable and accessible. And then it's understanding how do we make sure we're protecting the privacy and preferences of our customers as we access information about our customers.

eMarketer:

Did you have to dedicate a lot of resources to this?

Mika Yamamoto:

Absolutely. It takes dozens of people across multiple functions to make this happen.

eMarketer:

Which parts of the company got pulled into building the data lake?

Mika Yamamoto:

There's legal, in terms of what we're able to do and not do with our customer data. There's our enterprise analytics folks who look at the data and essentially architect what our databases look like. There's the business side of things, which would include me, that examines how do we want to use the data and how do we want to make sure we access the data in a way that helps us report on how well we're doing in the market and then helps us drive the best customer experiences. And then there’s the reporting side and there’s IT too.

eMarketer:

How long did this take?

Mika Yamamoto:

It takes multiple months. I mean, it’s taken multiple quarters of time to plan and execute this type of work. And we’re continually learning and exploring.

Interview conducted on July 5, 2018

"Behind the Numbers" Podcast