Speaker 1: Most people meet SAP DataSphere and assume it's just another data warehouse. That assumption is the first thing to let go of because DataSphere is built around a different idea. SAP calls it a business data fabric, which means it connects scattered systems and keeps the business meaning attached to the data. It started live as SAP Data Warehouse Cloud back in 2019 and took the DataSphere name in March 2023. It runs on SAP's cloud platform and feeds Microsoft tools like Power BI and Fabric. Without you hand-building every pipeline, stick with me and you'll walk away able to explain Datasphere in plain words, limits included. So what is it really? Datasphere is SAP's cloud service for pulling business data together and keeping its meaning intact. SAP describes it as a business data fabric, and that phrase carries most of the weight here. A fabric connects things, so Datasphere connects the systems that hold your data instead of asking you to pile everything into one place before you can start. One thing to be clear about is what Datasphere is not. It isn't your ERP system and it isn't your reporting tool. It sits between them, taking data from the systems that run the business and preparing it for the tools that explain the business. Keeping that middle position in mind makes the rest of this much easier to follow. The name is new, but the product isn't. It began in 2019 as SAP Data Warehouse Cloud, and in March 2023, SAP relaunched it under the Datasphere name, renaming the existing tenants that same day. That history explains a lot of the confusion you'll run into, because older articles, older job posts, and older project notes all use the first name. Underneath, Datasphere runs on SAP Business Technology Platform, which is SAP's cloud platform. You don't need to know that platform in detail to use DataSphere, but it helps to understand that this isn't software you install on a server. It's a subscription service that lives in the cloud, which is why the release cycle and the pricing model look different from the on-premise SAP systems you might be used to. It also helps to name the people who use it. A data engineer builds the connections and the flows, a modeler shapes the views and the analytic models, and a business analyst consumes the result in a report. Datasphere is designed so those three roles work in one place instead of throwing files over a wall. That alone explains a big part of why SAP built it. Start with the problem it's aimed at. In most companies, the data that answers a simple question is spread across several systems. A sales order sits in SAP S4 HANA, a support ticket sits in a separate service tool, and a forecast sits in a spreadsheet that one person guards like a secret. So when a manager asks how a product is really performing, the answer takes a week and three meetings. Worse, two teams often arrive with two different numbers, and the argument becomes about whose report is right instead of what to do next. That's not a people problem, it's a data problem, and it costs real money in slow decisions. Datasphere exists to close that gap. It connects the sources and harmonizes the data. That means it settles on one definition for things like revenue, customer, or product, then applies that definition everywhere the data is used. When everyone works from the same meaning, the meeting gets shorter. and the number on the screen stops changing depending on who built the report. The old alternative was worse. Companies wired systems together one pair at a time. Every new source meant a new custom pipeline, and every pipeline was its own small project with its own owner and its own quirks. Datasphere replaces that tangle with a shared layer where the connections and the definitions live side by side. This matters even more if you already live in Microsoft Tools. Plenty of teams run Power BI dashboards and Azure workloads right next to their SAP systems. The daily question is how to get SAP data into those tools without losing the business context on the way. That question is exactly where Datasphere earns its keep and we'll come back to it with specifics. The reason SAP built a fabric instead of just tuning an older product comes down to what a warehouse alone can't do. A warehouse is very good at holding data you've already copied in, It's much weaker at connecting data. You'd rather leave where it lives. Hold that thought, because it's the difference between a warehouse and a fabric. Let's clear up the biggest misconception first. A traditional data warehouse stores data. You copy information in, you organize it, and you report on it from inside the warehouse. Datasphere can do that, but storage is only part of the job. A fabric adds two things on top. First, it connects sources so you can leave data where it lives and read it without copying everything first. Second, it carries the business meaning with the data, so a field called revenue still means revenue when it lands in a report in another tool. Those two ideas, connection and meaning, are what the word fabric is really pointing at. That combination gives you a choice a plain warehouse can't offer. You can copy data into Datasphere and keep it there, or you can reach out and read it where it sits. Most teams mix both, copying the data they change often and reading the rest on-demand. Practical benefit is that you're not forced into one giant migration before you can answer your first question. A quick comparison helps here. A data lake holds huge amounts of raw data, often with little structure or meaning attached. A warehouse holds structured data you've already organized. A fabric leans on both ideas and adds the connections and shared definitions on top. The same business term then means the same thing, whether the data sits in a lake, a warehouse, or the source system itself. A concrete example makes this clearer. Suppose your sales figures live in S slash Vanhanna and your product master data lives in a separate system. With a warehouse-only mindset, you'd copy both, reconcile them by hand, and hope the definitions match. With a fabric, you connect both, agree on what a product is once, and let that single definition travel with the data into every report that uses it. Now, an honest note. Fabric is a marketing word, and it sounds grander than what happens underneath. Under the surface, you're still working with tables, views, and models, the same building blocks as any data platform. The fabric idea is about connecting systems and preserving meaning, not about inventing new physics. That's the concept. Next, let's meet the parts you'd actually touch in your first week. The place to start inside Datasphere is the space. A space is a separate workspace, usually for one team or one project, with its own storage and its own people. Think of it as a room with a lock on the door. so the finance team and the sales team can both work without stepping on each other. Spaces matter because they carry governance. A space administrator can create data access controls which apply row-level security to the objects inside that space. In plain terms, a regional manager can open a report and see only their own region while the numbers for everyone else stay out of view. The rule travels with the data, so it holds no matter which report someone opens. That sounds like an admin detail, but it answers a question beginners ask all the time. When you're unsure where a new dataset should live, the real question is usually which team owns it and who should be allowed to see it. Pick the space that answers both, and the rest tends to fall into place. Take a concrete picture. A company sets up a finance space and a sales space. Both draw on the same source system, but each keeps separate models, separate access rules, and a separate audience. One shared source feeds two governed use of it. Nobody has to duplicate the raw data, and no one has to trust a spreadsheet that slowly drifted out of date. Spaces also hold storage, and that's where cost enters the picture. The data you keep in a space consumes capacity, and capacity is what you pay for. So the storage decision is never only technical, Every table you persist is a small ongoing commitment, which is one more reason to copy only what you actually need. Keep one thing in mind as you go. The interface and the cost both have rough edges, and we'll get to those later. For now, just hold onto the idea that spaces are how Datasphere keeps work organized and safe. Inside a space, the next step is figuring out how data gets in and how it gets shaped. It starts with connections, Datasphere links to source systems, and those sources can be SAP systems or systems from other vendors. That range matters because no company keeps all of its data in one place, and a fabric that only talks to SAP would miss half the picture. Once a connection exists, you decide how to treat the data. You can point at it where it sits and read it directly, which SAP calls a remote table. That's useful when you want a quick look and don't want to copy anything. When you want speed or a local copy you can reshape freely, you bring the data across instead. For the copy and shape work, three names come up and the differences are worth learning early. The first is a replication flow which brings data across and can keep up with ongoing changes so the copy stays fresh. The second is a transformation flow which moves and reshapes data in a single step with more options for cleaning and joining. The third is a data flow the older graphical style that still works and still shows up in existing projects. SAP now points teams toward replication flows and transformation flows as the preferred way to load and write data, so if you're starting fresh, learn those two first. Here's how it looks in practice. You use a replication flow to pull open sales orders out of S slash 4HANA, then a transformation flow to clean them, join them to customer data, and land a tidy table. Two steps, one pipeline. and the result is something a model can pick up. Then comes the part beginners forget. A flow has to run, and a person has to watch it run. You schedule the loads, you check the logs when a source system hiccups, and you handle the rows that fail. None of that is glamorous, but it's what separates a demo from a setup that holds up in production. That's the movement. Now let's see what happens once the data has landed. Data in a table is still raw material. Nobody opens a report and reads rows of raw numbers, so the next layer turns tables into something a business user can actually use. Views and analytic models do that work. A view is a shaped version of the data, and an analytic model is a multidimensional model built for analysis. That's the structure that lets a report slice numbers by region, by quarter, and by product without the report author rebuilding the math each time. SAP describes analytic models as the foundation for consumption in SAP Analytics Cloud. So this is the layer where your modeling work turns into someone's dashboard. From there, people consume the data, Inside SAP, SAP Analytics Cloud is the natural home. Outside SAP, DataSphere exposes data through an OData API and through ODBC, which is the route many teams use for Power BI. That detail has a big consequence because the same model data can feed SAP reports and Microsoft reports without being built twice. A concrete example ties this together. Picture a model built for sales performance. It takes the cleaned orders from your transformation flow joins them to product and customer data, and defines measures like total revenue and average order value. A dashboard then slices those measures by region and quarter. The model is where the business rules live and the dashboard is just a window onto them. This layer is also where beginners tend to get stuck and the reason is worth spelling out. Modeling is the moment where data stops being data and starts being business meaning. You have to decide what a customer is, what counts as an active order and how returns should be treated. Those are business decisions wearing a technical code and no tool makes them for you. Get a business owner in the room early and this layer gets much easier. One more note before we leave this layer. SAP used to offer a separate area called business builder aimed at business users and that area is now deprecated. If you find an older tutorial that leans on it, treat it as history and build on the current modeling path instead. Now we get to the part I promised earlier, the Microsoft side. If you work with Power BI and Microsoft Fabric, Datasphere gives you a cleaner path to bring SAP data into that world, and this is where the two platforms actually meet. Start with Power BI. Power BI can read Datasphere data through ODBC or through the OData API, so a report can sit on top of a modeled dataset and refresh on a schedule. For a lot of teams, that alone solves the daily problem of getting trustworthy SAP numbers into a familiar dashboard. Then there's Microsoft Fabric, which goes a step further. Fabric can mirror SAP data using SAP Data Sphere, and Microsoft describes that as a no ETL, near real-time integration. In plain language, instead of building every pipeline by hand, you let Fabric keep a near real-time copy of the SAP data in one lake, and your Fabric workloads read from there. Microsoft also points to shortcuts and mirroring as ways to pull SAP data in, which keeps the setup lighter than a traditional extract. That's why this pairing makes sense. SAP holds the business meaning, the definitions, and the rules that make the data trustworthy. Microsoft holds the analytics, the reporting, and the broader data platform. DataSphere sits in the middle and lets each side do what it's good at, instead of forcing one platform to do both jobs badly. That division of labor is the whole point of the connection One thing to keep straight is where security lives. Access to the data is governed in SAP through Spaces and the data access controls we covered. Access to the fabric side is governed in Microsoft. The two systems pass data between them, but they don't merge their permission models, so you plan identity on each side rather than assuming one covers the other. One more thing about where this is heading. SAP Business Data Cloud. Launched in February 2025 with Databricks, groups Datasphere together with SAP Analytics Cloud and SAP Databricks, so the product now lives inside a larger data platform story, and for you that means the skills you build here keep their value as the story grows. This is the honest part, and it's the part that saves projects. Start with the mistake I opened with. People treat Datasphere as a renamed data warehouse, build it like one, and then wonder why the fabric idea never pays off. If you only copy data in and report on it, you've used half the product and left the better half on the table. The second mistake is skipping governance. Spaces and access controls can feel like setup chores, so teams rush past them. Then, six months later, everyone can see everything, and a small problem has grown into a real one. Set the space boundaries and the row-level rules early while it's still easy because retrofitting security onto a busy system is far harder than starting with it. The third mistake is ignoring cost. Datasphere is consumption-based, priced on capacity units, and third-party sources put the range at roughly a third of a dollar to a dollar per capacity unit per hour. Treat that as a rough guide rather than a quote and size the workload before you commit to a number in a budget meeting. there are real limits too. Community posts report that the interface can feel slow and that large data volumes need careful planning. So don't assume you can throw a massive historical load at it on day one. Not every fabric feature for SAP is generally available everywhere, so check the current status before you promise it in a plan. And that older business builder area is deprecated, so don't build a new approach around it. One more honest point. Datasphere doesn't replace SAP BW overnight. For most companies, the two run side by side for a while and moving off BW is a project with a plan, not a switch you flip. Knowing that upfront keeps expectations reasonable on each side of the table and it stops a solid platform from getting blamed for an unrealistic timeline. None of this means Datasphere is the wrong choice. It means the choice has conditions. If your data lives across SAP and non-SAP systems, if you care about keeping business meaning consistent and if you run Microsoft Analytics on top then it fits well. If you just need a small reporting database and nothing else, a simpler tool will do the job for less. Let's bring it back to one picture. DataSphere takes the data your business already runs on, brings it together, and keeps its meaning attached as more tools use it. Spaces organize the work and guard who sees what. Connections and flows move and shape the data. Models turn it into something people can use. And on top, Power BI and Fabric can read that data so your SAP numbers show up where your team already works. If you take one thing away, let it be this. Datasphere is less a new database and more a way to keep SAP data meaningful while more tools use it. That reframe is what the word fabric keeps pointing at. So here's a question for you. What's the first SAP data source you'd want to bring into a tool like Power BI, and what's stopping you today? If this made the topic feel less intimidating, subscribe to M365FM to master the tools and the mindset shaping the future of work.