Review of Silvon Software, a Supply Chain BI Vendor

Last updated: December, 2025

Go back to Market Research

Silvon Software is a long-established vendor focused on analytics for manufacturers and distributors, commercialized primarily through its Stratum Analytics Platform: a data-hub + reporting stack that consolidates operational data (typically ERP and related systems) into a curated analytical layer, then exposes that layer through web dashboards and Excel-centric workflows, with optional planning and write-back. Public product documentation indicates an architecture centered on Microsoft SQL Server and SQL Server Analysis Services (SSAS) cubes, with companion “Viewer” web UI components, a “Connector for Viewer” layer that provisions and refreshes SSAS models, and a “Server” component tied to Stratum’s storage database; legacy support for IBM i / DB2 on iSeries also appears in deployment requirements. Silvon positions this platform as supporting sales, finance, and supply chain decision-making (inventory, demand/supply visibility), but available technical materials emphasize BI/OLAP mechanics and governance rather than novel optimization methods; where forecasting or “predictive” outcomes are mentioned, the strongest public evidence points to planner-driven workflows and cube-based measures rather than clearly specified machine learning pipelines.

Overview

Silvon’s supply-chain-relevant product line is best understood as BI + data hub + OLAP, with a planning add-on that enables write-back to analysis models rather than a dedicated “APS-style” optimization engine.

Stratum’s public deployment requirements and help documentation describe a multi-component system:

• Stratum.Viewer (web front-end) paired with a SQL Server database for metadata.
• Stratum.Connector for Viewer, with both a SQL Server metadata database and an SSAS database (cube) it maintains.
• Stratum.Server and a Stratum storage database, which can be hosted on Windows/SQL Server or on IBM i / DB2 (per the documented configurations).¹

This documentation pattern is consistent with an implementation approach where Silvon (or partners) deliver a pre-shaped analytical model (dimensions/measures, often industry-specific), integrate customer data into that model, and provide web/Excel consumption plus optional “planning” entry points.

Silvon Software vs Lokad

Silvon and Lokad address “supply chain” from fundamentally different technical starting points.

Silvon’s Stratum materials describe a Microsoft BI-centered architecture (SQL Server + SSAS + web viewer) where the core deliverable is an analytical layer and curated metrics, with optional planning workflows that write values back into the analysis model (i.e., planning as OLAP write-back + governance + reporting).¹²³ On that basis, the “engine” is primarily the cube/data model and the reporting layer; automation is typically about scheduled refreshes, managed data pipelines, and standard KPIs.

Lokad, by contrast, explicitly frames its core deliverable as decision optimization under uncertainty, presenting a technology roadmap built around probabilistic forecasting (2016) and later optimization paradigms (e.g., stochastic discrete descent, latent optimization).⁴ Lokad’s technical documentation emphasizes a programmatic “white-box” approach (Envision) where forecasting/optimization logic is expressed as code and executed as part of the platform workflow, rather than embedded in a fixed cube schema.⁵ In practical comparison terms: Silvon’s publicly documented mechanism is closer to enterprise BI/OLAP with planning extensions, while Lokad’s documented mechanism is closer to model-driven predictive optimization (forecast distributions feeding decision computation).⁴⁶⁵

Company history, ownership, and acquisition signals

Silvon presents itself as a long-running independent software company. Its leadership material identifies the company as founded in 1987, with Michael Hennel cited as CEO and co-founder (with Frank Bunker).⁷

The most clearly documented corporate transaction located in public third-party sources is from 1998: MKS acquired Silvon’s Software Distribution Management (SDM) unit, described at the time as a business unit rather than an acquisition of Silvon as a whole.⁸ A subsequent third-party trade note in 1999 references Silvon’s DataTracker 3.0 release positioned around performance management / measurement improvements.⁹ Beyond this SDM divestiture, no high-confidence public record (in widely accessible sources) was found indicating Silvon being acquired or executing major acquisitions; given Silvon’s private status, absence of evidence is not evidence of absence, but the discoverable footprint is limited.

Product and architecture

Core architectural pattern: SQL Server + SSAS + web/Excel front ends

A key piece of non-marketing evidence is Silvon’s own Stratum.Viewer/Connector requirements documentation (v6.2), which lays out multiple server topologies (single-server, split app/storage, multi-app + storage) and explicitly names:

• SQL Server databases for Viewer and Connector metadata
• an SSAS database for the Connector
• Stratum storage database
• Stratum.Server as a required component in the overall system¹

This matters because it constrains what the system is likely doing technically: the “analytics brain” is predominantly the cube schema (dimensions/measures) and the ETL/refresh process that keeps it synchronized with operational sources. In the same requirements document, Silvon also documents scenarios where Stratum storage resides on IBM i / DB2 and lists client-side providers (IBM i Access for Windows, Microsoft OLE DB Provider for DB2) required for those deployments, indicating a footprint in IBM i-centric manufacturing/distribution IT environments.¹

Planning module: write-back into the SSAS model (not a demonstrated optimization solver)

Silvon’s Stratum.Viewer Planning Module materials describe the planning feature as an add-on to the Viewer environment.² In addition, Silvon’s help documentation for the Viewer Planning module describes operational behavior consistent with cube write-back: enabling write-back on a cube partition, processing dimensions, and managing the write-back table as part of the workflow.¹⁰ This strongly suggests “planning” is implemented as controlled data entry and governance on top of the OLAP model (often valuable), but it is not, on its face, evidence of algorithmic optimization.

Deployment and rollout methodology

Silvon publishes an “8-step” BI deployment approach as a standalone PDF. It describes a structured project method (phased delivery) rather than a self-serve product onboarding model.¹¹ In the same direction, Silvon’s Stratum “Data Import – Installation Steps” documentation (revised as recently as 2024) indicates ongoing maintenance of installation/runbook-style guidance and supports the interpretation that deployments are integration-heavy and operationally prescriptive (data mapping, refresh scheduling, environment setup).³

Taken together, these artifacts support a rollout model where value realization depends heavily on:

• shaping and validating the analytical model against customer data,
• standing up SQL Server/SSAS infrastructure (or compatible equivalents),
• building repeatable refreshes and data quality controls,
• and training users on Viewer/Excel workflows.

Machine learning, AI, and optimization: what is (and is not) evidenced

Silvon’s marketing and thought leadership frequently references forecasting and supply-chain outcomes, but public technical documentation reviewed for Stratum (requirements, planning module, and operational help) is largely about data modeling, OLAP, and governed write-back rather than about trained predictive models or reproducible optimization algorithms.

Concretely:

• The strongest “how it works” evidence available for planning points to SSAS write-back mechanics.¹⁰
• The strongest “how it’s deployed” evidence points to traditional enterprise BI implementation mechanics (servers, databases, SSAS, structured deployment steps).¹³¹¹

This does not imply Silvon cannot deliver predictive value in practice (e.g., via customer-defined measures, statistical forecasting embedded in upstream tools, or partner add-ons), but it does mean that publicly verifiable, technical substantiation for state-of-the-art ML/optimization is limited relative to vendors who publish model classes, evaluation regimes, or solver architectures.

Commercial maturity and market presence

Silvon’s own “About” materials claim a scale of operations (including a stated professional staff size) and position the product as established in manufacturing/distribution contexts.¹² However, independently verifiable funding-round history was not found in the accessible public sources reviewed for this page; Silvon appears to operate as a private company without a conspicuous venture-funding footprint in readily accessible press or filings.

Public customer references

Silvon maintains a public “Company We Keep” customer page with named logos/references.¹³ From an evidence-quality perspective, this is vendor-authored and should be treated as a claim unless corroborated. One notable externally corroborated historical customer reference is HarperCollins, appearing in a Microsoft case study that describes HarperCollins using Silvon DataTracker (an earlier Silvon product line) on Microsoft SQL Server for decision-support reporting.¹⁴ Beyond that, additional customer corroboration would require either customer-authored references, partner-authored case studies, or independent reporting.

Conclusion

Silvon Software’s supply-chain-relevant offering, as evidenced by its publicly available technical documentation, is best characterized as an enterprise BI/analytics platform tailored to manufacturers and distributors: it consolidates data into SQL Server-backed stores and SSAS cubes, exposes KPIs via web and Excel, and optionally enables controlled planning through cube write-back. The documentation is relatively concrete about infrastructure, components, and operational steps, which supports credibility for the BI/governance aspects. By contrast, public evidence for state-of-the-art ML forecasting or optimization is weak: the most explicit “planning” mechanisms documented align with OLAP write-back and workflow control rather than with a demonstrated solver or modern probabilistic modeling stack. Commercially, Silvon presents as an established niche vendor with long operating history, but with limited publicly accessible third-party disclosure regarding financing and corporate transactions beyond a documented divestiture of an SDM unit in the late 1990s.

Sources