When I first met the acronyms DLO and DMO, they looked like the kind of jargon that exists to keep beginners out. They are not. Once you see what they really are, they become two of the clearest ideas in all of Data Cloud. So let me walk you through them slowly, the way I would with a student sitting beside me.
You have already brought data in through data streams. Now we need to make that data usable. This is the HARMONIZE step, where data from many sources learns to speak one shared language.
First, the raw data: DLOs
When a data stream ingests data, it lands in a Data Lake Object, or DLO. A DLO holds the data more or less as it arrived. If your e-commerce system sent a field called cust_email and Salesforce sent one called Email, both come in untouched, each in its own DLO.
Think of a DLO as a faithful copy of the source. It does not judge or reshape anything. It simply preserves what came through the pipe. This is good, because the raw data stays honest and traceable. You can always look back and see exactly what a system sent you.
But raw data has a problem. Every system named things its own way. One calls it cust_email, another EmailAddress, a third email_1. To Data Cloud, these look like three unrelated fields, even though they all mean the same thing: how to reach a person.
Then, the shared structure: DMOs
A Data Model Object, or DMO, is the harmonized, standardized version of your data. Salesforce provides a standard model with familiar shapes like Individual, Contact Point Email, Order, and Product. A DMO says, in effect, “this is what an email address looks like in our shared language, no matter which system it came from.”
So a DLO is the raw arrival, and a DMO is the cleaned-up, agreed-upon structure everyone downstream relies on. Segments, calculated insights, and identity resolution all work against DMOs, not the messy raw DLOs. That is why the data model matters so much. It is the common ground.
A DLO is your data as it arrived. A DMO is your data as everyone agrees to understand it. Mapping is the bridge between the two.
The bridge: mapping
So how does a field in a DLO become part of a DMO? Through mapping.
Mapping is simply telling Data Cloud, “this raw field means that standard field.” You connect cust_email from your e-commerce DLO and Email from your Salesforce DLO both to the same Contact Point Email field in the DMO. Now Data Cloud understands that these two different-looking fields are actually the same idea.
You repeat this for the fields that matter: names, addresses, phone numbers, order amounts, product identifiers. Field by field, you are translating many local dialects into one shared language.
This is the part of teaching I always loved most. A bağlama and a piano are tuned and built completely differently, yet they can play the very same melody once you agree on the notes. Mapping is that agreement. The instruments stay different; the song becomes one.
Why this two-layer design is a gift
It might seem simpler to just clean the data on the way in and skip the raw layer. But keeping DLOs and DMOs separate gives you two real advantages.
- The raw truth is preserved. Because DLOs keep the original data, you can re-map, fix mistakes, or bring in new fields later without re-ingesting everything.
- Sources stay independent. When a new system arrives next year, you do not rebuild your model. You ingest it into its own DLO and map it to the same DMOs you already trust. The shared language simply grows another speaker.
This is what lets Data Cloud scale gracefully. New data does not break the structure; it joins it.
Where this sits in the journey
Let me place it in the bigger arc so it stays clear.
- Connect — data streams ingest raw data into DLOs.
- Harmonize — you map DLOs to DMOs so everything shares one language. (You are here.)
- Unify — identity resolution stitches records into one profile per person.
- Activate — you build segments and put that unified data to work.
Harmonizing is the quiet middle step that makes unification possible. You cannot match “the same person” across systems if every system describes a person differently. The DMO gives identity resolution a consistent shape to work with.
So when you hear DLO and DMO, hear it plainly: raw arrival, and shared understanding, joined by mapping. That is the whole idea. Everything else in Data Cloud builds on this calm foundation.
Your next step
Continue along the data journey from raw ingestion toward a single customer profile:
- Data Streams: Getting Data Into Data Cloud — how the raw data arrives in the first place.
- Identity Resolution for Beginners — how harmonized data becomes one profile per person.
- Browse more in the Data Cloud category.