Fast and solid decisions help companies to achieve economic success. SAP Data Warehouse Cloud and SAP Analytics Cloud provide easy access to information and insights to draw the right conclusions. So do cloud solutions represent the exclusive future in SAP’s business intelligence world? In this interview, Fabian Hartje, Head of Product Management for Data Warehousing at SAP, looks ahead and outlines the developments that will shape reporting, analytics and planning in the future.
Fabian Hartje: In my view, the increasing number of data sources and the constantly growing volume of data are the causes of many challenges that companies have to face. Both factors result in an increasing number of analysts in the company. After all, only those who manage to maintain or improve the quality of information will gain a competitive advantage in the long run. The IT department does not usually grow at the same rate, creating an imbalance. It’s all about providing data and necessary information in a timely manner. If this does not succeed, growing shadow IT landscapes are the result, or the actual business logic migrates to the consuming tools and becomes increasingly difficult to manage.
Companies often have several data-consuming tools in use at the same time. These are not only reporting tools, but also, for example, customer-specific apps that also have a need for information. The result: business logic is duplicated and not followed up in all places when changes are necessary.
For planning, the same questions arise: What database do I use? Do I use the same dimensions and planning structures across business units? Are the results of different planning versions and plans comparable at all? I can see the need and the will of many customers to provide a central data platform. However, this platform must meet the challenges mentioned above. In other words, it must strike a balance between freedom, control, security and trust.
Fabian Hartje: To put it simply, data and information must be understandable, comprehensible and usable for all stakeholders in the company. I will only gain valuable insights if I can work freely with the data and reuse models without reinventing the wheel every time.
To achieve free working with data and models, knowledge of the data and its contents is essential. From my point of view, the metadata and thus the semantics or the context play a superior role. Conversely, this means that a company must ensure that all stakeholders can access the knowledge and enrich it with relevant information. To do this, it is imperative to have access to the semantic knowledge in the respective source systems in order to use data in the right context.
For modeling, it must also be ensured that the semantic knowledge and the enrichment achieved are carried forward, thus achieving consistency. On the data modeling side, it is also important to establish governance to get the most out of the data. It is advisable to establish a layered architecture right from the start and to prevent use case specific aspects from being modeled in the generic layers. This requires a separation of data integration and business logic.
The goal must be the greatest possible openness in accessing data and information – from classic reporting to analytics to data science. In this way, companies ensure that everyone works with the same knowledge and enriches it with the insights gained to draw the right conclusions.
Fabian Hartje: In general, there are three possible solutions: SAP Data Warehouse Cloud, SAP BW/4HANA and SAP HANA for SQL Data Warehousing. Depending on preference – cloud or on-premise – there are certain paths that SAP existing customers, but also new customers, can take. With a cloud preference, companies can choose between SAP BW/4HANA and SAP Data Warehouse Cloud. A hybrid data warehousing scenario complements SAP BW/4HANA with SAP Data Warehouse Cloud. This option has its strengths, for example, in opening up towards the business department, in data integration from non-SAP sources and in self-service scenarios.
SAP Data Warehouse Cloud is a cloud-only product and brings with it the classic advantages of a SaaS solution, for example with regard to scalability and elasticity. As already mentioned, SAP Data Warehouse Cloud targets disparate and heterogeneous system landscapes that need to be integrated in the company and closes the organizational “whitespot” between the responsibilities of business & IT and their tooling. To achieve this, SAP Data Warehouse Cloud provides its own layer for business logic modeling (business layer), tooling for integrating external data and standard content.
Those who prefer an on-premise solution can choose between SAP HANA for SQL Data Warehousing and SAP BW/4HANA. The first option takes an agile, flexible, and database-driven approach to data warehousing and targets companies whose resource skillset is heavily in the SQL environment. SAP BW/4 HANA gives companies the opportunity to establish a fully integrated data warehousing solution. It scores with standard content and data tiering to actively manage the data footprint.
Fabian Hartje: There is no standard answer to this question. Every company must answer this question individually for itself along several dimensions and aspects. For example, the degree of freedom and flexibility I need in my company is relevant. This ranges from administration and expandability to full control over parameterization. In terms of the cloud, I have to decide between software-as-a-service and platform-as-a-service. Another dimension is the issue of security and regulations. Depending on what data I generate and process in my company, I may need to be able to find out where my data is stored at any time. This is where a hybrid approach can be beneficial to ensure compliance with applicable regulations.
From my point of view, one of the most fundamental decision dimensions is whether I even have the necessary experience and skills in the company to build and operate a data warehouse on-premise. Can I ensure security, availability and scalability? Cloud providers deliver these elements directly. This means the company can focus on actual tasks such as data provisioning, modeling, performance, etc.
Also relevant: Can I realistically assess my business needs? Are they fixed or do I need the possibility of flexible growth? If there is uncertainty about future needs, this speaks in favor of a cloud solution. The time period in which the system must be available is also not insignificant. Is there sufficient time to procure hardware and set up the complete system, or must a solution be available as soon as possible? Especially for this question, the use cases and the scope I want to map play an important role. For companies that want to start with individual use cases and business areas, the budget question inevitably arises. With a cloud solution, I can get started with individual use cases and easily adapt my system over time through scalability and elasticity, rather than having to start immediately with the full scope and cost.
Fabian Hartje: In discussions with customers, we are increasingly finding that there is a desire for a central “platform” for all kinds of corporate data. The landscapes of companies are extremely heterogeneous – both vertically, meaning that there are several identical systems, for example several ERP systems, and horizontally, in terms of the multitude of systems. Therefore, a layer is needed that abstracts the accumulated complexity. In contrast to a classic data warehousing approach, which consolidates data from a multitude of sources in a rather static, well-defined layer architecture and makes it available for evaluation purposes, today’s requirements tend to go in the direction of agility, semantic integration and simple connection of external data sources. Behind this is always the goal of solving department- and industry-specific problems and supporting decisions.
Beyond that, I see various developments whose importance will increase in the future: In terms of connectivity, understanding the connected data is absolutely key to successfully turning data into information. Driven by heterogeneous system landscapes and external data from suppliers or data providers, topics such as harmonization, cleansing and the rule-based bringing together of disparate data are central to data interaction. Given the huge volumes of data, centralizing data in one place before modeling work is a utopia, so more flexibility is required. Against this backdrop, data virtualization has a significant role to play. At the same time, companies must also keep an eye on the aspects of response times, system load and egress costs. In addition to pure modeling, making data and information accessible is becoming increasingly important. I like to call the ability to make the business logic anchored in modeling readable and processable for humans and machines the “responsibility to provide”. Specifically, we want to achieve this through a catalog that spans everything from analytics to the data warehouse to the source systems.
Finally, the boundary between data warehousing and analytics will (have to) soften more and more. As a result, the integration between today’s data warehouse and an analytics tool must be as seamless and frictionless as possible – not only in terms of user experience, but also in terms of connectivity, modeling, and administration.