Sign In

Martin Kihn

SVP Strategy @ Salesforce | 4x Bestselling Author | Keynote Speaker | Podcast Host (PaleoAdTech) | ex-Gartner, Publicis, MTV | AI Data Scientist

Martin Kihn is the Senior Vice President of Strategy at Salesforce Marketing Cloud, a key figure shaping the industry's top customer data platform, AI innovations, and marketing automation solutions since 2018. With over 15 years of expertise in digital marketing, data science, and AI, he excels at integrating marketing technology, advertising technology, and AI to drive business success.

Before his tenure at Salesforce, Martin held the position of VP of Research at Gartner, overseeing research on programmatic advertising, digital marketing analytics, and the application of data science, AI, and GenAI in marketing. A respected speaker and thought leader, he has shared valuable insights with Gartner's global clientele and contributed extensively to industry conferences and publications.

Martin's professional journey includes strategy, analytics, and AI roles at premier digital marketing agencies, collaborating with renowned global brands across diverse sectors. His educational background features an MBA from Columbia Business School and a Bachelor's Degree from Yale University.

A renowned author with four bestsellers on digital marketing and advertising, Martin is a sought-after keynote speaker at major conferences and hosts the #PaleoAdTech podcast, delving into the future of ad-tech. Beyond work, he is an enthusiastic reader, runner, and traveler.

Highlights

Sep 10 · linkedin

💥 💥 Bright and early next Tuesday morning 9/17 at 8:30amPT --> it's your chance to find out "What's New & What's Next for Data Cloud" at #Dreamforce24 --> and you'd have to be crazy to miss it (link in comments)

🤠 🤠 Join me and two very special guest stars Joyce Avila and Andrea Lin as we talk about the revolution in customer data management happening at Salesforce and the magic of Data Cloud as the engine of #AI and #GenAI in the enterprise

🍿 🍿 It's your chance to hear from Joyce about her new book "Hands on with Data Cloud" (out now) and from Andrea about her & my new book "Customer 360: How Data, AI and Trust Change Everything" (out next month)

It's worth rising early Tuesday and starting #Dreamforce24 right! See you there!

#datacloud #cdp #digitalmarketing

Jul 31 · linkedin

Lately there has been a trove of lightly-researched POVs on Salesforce Data Cloud's "Zero-Copy" capabilities coming from competitors.

Here is how it works:

First:

  • There is no copy - that is what we mean by "zero"
  • It is the same mechanism used by composable CDPs

Zero-copy methods are one of the more exciting recent developments in data management. Not just for marketing; for everyone. They point to a future when large data sets can simply stay where they are and yet be available to other systems for analysis and use.

In the context of Data Cloud, zero-copy lets you keep all your operational or product data - say - in Snowflake or Databricks and yet have it act as though it were stored in our data lake. And vise versa.

How does this happen?

The actual mechanism differs depending on the data warehouse and the direction (in or out) in which the information flows. (Data warehouse here refers to newer cloud data warehouses, i.e., Snowflake, Databricks, and those offered by Google Cloud, Amazon Web Services (AWS) and Microsoft Azure.)

This is how it works for us:

  • From Data Cloud (CDP) to Data Warehouse

We can call this data sharing, accessing data sitting in the CDP from the warehouse.

In this case, the user identifies the CDP data to share (called data lake, data model or calculated insights objects). These objects are linked to the target warehouse, where the user can do queries on warehouse and Data Cloud data at the same time.

But how does this magic happen without copying?

Behind the scenes, the process creates "virtual tables" that describe the structure of the CDP data but not the data itself. It's a blueprint or pointer.

  • From Data Warehouse to the CDP

This process is often called data federation.

Here Data Cloud mounts tables from the warehouse as external data lake objects. Mounting is a virtual-blueprint-creation process similar to the one described above. Data Cloud then does what it does with ID management, analysis, segmentation, etc.

Voila the CDP can access data from the warehouse by performing federated (or combined) queries that include data in Data Cloud and objects provided by the warehouse, which are treated in the same way.

There are a lot of good reasons to use zero copy. An ETL process itself (moving and changing data) can introduce errors - and be complicated, expensive and slow. Storage costs are also real.

And there's a logical elegance to separating different types of data. It's what the economist David Ricardo called comparative advantage.

Focus the CDP's data lake on customer data useful for its tasks. Other data that's relevant from time to time can be accessed using zero-copy.

#zerocopy #datawarehouse #cdp #customerdata

Martin Kihn
Add to my network

Location

New York, New York, United States