Data-Driven Analysis: Beyond BI
Procurement digitalization requires data-driven decision-making on sourcing, demand management, and price compliance. Data-driven analysis also underpins pre- and post-merger integration, risk analysis, contract analysis, and sell-side/organizational analysis including balance of trade and headcount.
Unfortunately, the state of the art for procurement data analytics is to deliver pre-cooked data inside a BI tool like Tableau, Power BI, or Qlik. Visualization tools are useful, but they severely constrain analysis:
- A BI tool’s fixed schema prevents further segmentation of the data, such as breakdowns by internal reporting structures, preferred vendor designations, FTE count, and so on.
- Existing segmentations can’t be altered because there is no ability to map dimensions.
- Custom analyses are discouraged because the next refresh of the schema will wipe out user changes such as new datasets and new dimensions.
- Models requiring complex dependencies must be built by extracting data to external tools like spreadsheets, leading to unmaintainable and uncertain results.
Transcending these limitations requires an approach and an architecture that’s fundamentally different from the database-derived ideas underpinning the leading BI tools. Spendata’s architecture was designed from the ground up to support change, exploration, and analysis, so that users can maximize the value of their data. Here are its key attributes:
Spendata’s architecture was designed from the ground up to support change, exploration, and analysis
- It supports real-time changes. Users can create new datasets, dimensions, dashboards, views, measures, and mappings in real time, usually measured in seconds. Unlike a fussy and fragile BI tool, Spendata’s underlying schema is built and maintained automatically and invisibly.
- It enables dynamic modeling. The power of a spreadsheet is its ability to create and maintain dependencies between cells, which in turn enables the creation of dynamic models. Spendata creates and maintains dependencies between dimensions, measures, mappings, views, datasets, and inputs. This allows users to build dynamic models at database scale.
- It preserves custom analyses. Centralized databases and BI tools discourage custom analysis, because when the data or the schema changes, custom work must be redone. With Spendata, users can make profound alterations to the cube – adding new datasets, dimensions, dashboards, views, measures, mappings, and so on – with no worries about preserving those custom analyses. When the shared workspace is updated, the centralized changes are merged automatically with no loss of work.
- It is powerfully extensible. Dimension derivations and measures are scriptable without limitation. Derivations can depend on other derivations and on data structures assembled by “script-only” derivations, enabling computed and rolled-up data to factor into derivation and mapping decisions.
- It can be fully automated. Unlike restrictive BI tool APIs, Spendata’s API is not constrained to a subset of functions. Anything a user can do directly can be done through the API. Users can load new data, replace data, remove data, refresh cubes, map dimensions, apply filters, run reports, and extract results. A RESTful interface empowers your RPA tool of choice – or any ordinary scripting language like Python, Perl, bash, Powershell, etc. – to access Spendata’s full range of capabilities.
- It loads data intelligently. When input data formats change, Spendata automatically aligns new and old fields, and extends old data formats with new default fields as necessary. When more advanced transforms are required, scripting support enables complex and auditable transforms of arbitrary complexity.
- It is secure by design. Because user data never leaves the user’s machine, Spendata complies with every data privacy and security regulation including GDPR, HIPAA, CCPA, and CPRA. And, since Spendata never has possession of the data, your data cannot be turned over to third parties or governments.
Some managed services providers argue that their clients don’t need these and other related capabilities, but we disagree. Even if part or all of a procurement transformation effort is outsourced, access to powerful data transformation and data analysis tools is essential for building data-driven competency within an organization.