Review of SKU Science, Supply Chain Forecasting Software Vendor

Last updated: December, 2025

Go back to Market Research

SKU Science is a software editor delivering a cloud SaaS application for sales/demand forecasting and operational performance tracking, positioned around S&OP-style business review: users load historical sales (and related dimensions), obtain automatically generated baseline forecasts (the vendor claims selection from “644 statistical combinations”), review accuracy and bias metrics (including monetary valuation of forecast errors), and use prebuilt dashboards and custom reports to monitor performance vs budget and prior year, with exports intended for downstream systems (ERP/BI). The public product narrative centers on speed-to-start, forecast editability across hierarchies (item/customer/store/territory/product family), and KPI instrumentation (including “forecast value added” comparisons between the baseline and user-modified or externally uploaded forecasts). The company states that customer data is encrypted and hosted on AWS, and it lists multiple security/compliance labels on its website; however, publicly accessible third-party attestations (e.g., audit firm, report availability, certification numbers) are not clearly provided in the surfaced materials, so such claims should be treated as self-declarations unless independently validated. SKU Science also publishes named customer stories (e.g., Bridgestone, Ocean Spray) but, outside of its own materials and a small number of trade-press mentions, independently verifiable technical detail on its forecasting/ML internals, optimization components, and engineering stack is sparse. ¹²³⁴⁵

SKU Science overview

SKU Science’s publicly described deliverable is a forecasting-and-review layer: (1) ingest historical data; (2) compute a baseline forecast and associated KPIs; (3) support forecast overrides at multiple aggregation levels; (4) track accuracy, bias, and “value added” of overrides; (5) provide dashboards for operational review; and (6) export reports for reuse in other IT systems. The strongest public evidence supports “forecasting + performance tracking,” not end-to-end supply chain decision optimization (e.g., automated replenishment, allocation, scheduling). ¹²

SKU Science vs Lokad

SKU Science and Lokad sit in different parts of the supply-chain software spectrum, even though both touch “forecasting”:

• Primary output: SKU Science’s product pages emphasize forecasts, KPIs, dashboards, and editable forecast hierarchies (plus reporting/export). Its public-facing descriptions do not clearly commit to producing prescriptive decisions (e.g., order quantities or allocation plans) as the core artifact. ² Lokad describes its platform as delivering predictive optimization apps for supply chains, i.e., decision-oriented outputs (prioritized decisions) as the primary deliverable of a “Quantitative Supply Chain” initiative. ⁶⁷
• Mechanism and extensibility: SKU Science frames forecasting as automatic selection from a finite family of statistical configurations (“644 statistical combinations”), with user overrides on top. ² Lokad’s approach is explicitly programmatic: it centers on a domain-specific language (Envision) and a platform designed to encode bespoke constraints/objectives, with documentation positioning Envision as a DSL for “predictive optimization.” ⁸⁷
• Uncertainty modeling: SKU Science’s public materials highlight trend/seasonality detection and accuracy tracking, but do not clearly document probabilistic forecast distributions as a first-class output. ² Lokad’s public material explicitly foregrounds probabilistic forecasting as a core paradigm (including technical documentation on probabilistic demand forecasting and explanatory pages on probabilistic forecasts). ⁹¹⁰
• Project/rollout model: SKU Science markets an “ultra quick start” and being “in production in a few days,” suggesting a relatively standardized, productized onboarding. ¹ Lokad explicitly documents a project-based delivery model (“initiative,” phases, deliverables) where code artifacts and dashboards are produced as part of an implementation. ⁶⁷
• Evidence posture: SKU Science provides concrete feature statements and named customer stories, but limited externally checkable technical depth on algorithms, architecture, and certifications in the surfaced sources. ²³⁴⁵ Lokad publishes substantial technical documentation and long-form explanations of its forecasting paradigm and platform mechanics (while still being vendor-authored). ⁷⁸⁹

Company identity, history, and corporate footprint

SKU Science appears to operate as a French-registered entity (“SKU SCIENCE”) with a SIREN/SIRET footprint visible in French company registries and aggregators; these sources support basic corporate facts (registration identifiers, legal form, dates, addresses) but do not, by themselves, validate product capabilities. ¹¹¹²

Public-facing “Company” messaging positions SKU Science as a forecasting-focused vendor; however, beyond self-published pages and limited trade press, evidence about funding rounds, material corporate milestones, or M&A activity is not prominent in the sources reviewed here. This absence should not be read as proof of absence; it should be read as “not evidenced in the consulted public set.” ¹³¹¹¹²

Product scope and what it delivers

Forecast generation and selection

SKU Science states that it “automatically select[s] the best forecast from 644 statistical combinations,” with trend/seasonality detection and forecasting at multiple aggregation levels (item, item/customer, and other combinations). ² The public materials do not enumerate the 644 combinations (e.g., exact model families, parameter grids, intermittent-demand handling, event/promo modeling, cross-series pooling), so the claim is best interpreted as: an internal model-selection process across a predefined set of statistical configurations, rather than an openly specified ML architecture. ²

Forecast editing, value-added tracking, and KPIs

SKU Science states that forecasts can be modified “at any level” (item, customer, store, territory, product family, etc.). ² It also claims to compute accuracy/bias and track “forecast value added” by comparing user-modified or externally uploaded forecasts (e.g., sales input) against the baseline, and it states that errors are “financially valued” to focus attention. ² These are operationally meaningful features for an S&OP cadence, but they remain in the category of forecast governance and performance measurement rather than optimization of decisions. ²

Dashboards, business review, and reporting/export

SKU Science positions dashboards as “off-the-shelf” business review tools, including comparisons vs fiscal-year budget and last year, and backlog vs forecast/budget visualizations. ² It also states that reports can be exported and reused in other IT systems (ERP/BI). ² The public wording does not specify whether exports are file-based, API-based, or both. ²

Deployment and roll-out signals (public evidence)

SKU Science’s homepage messaging emphasizes rapid onboarding (“sign up… be in production in a few days”) and includes at least one testimonial suggesting a file-based workflow (“sent all the data files… it was all ready” by Monday). ¹ This supports the hypothesis that batch data uploads / file transfers are a primary ingestion path, but the public documentation surfaced here does not provide a rigorous integration specification (connectors, schemas, APIs, CDC patterns, etc.). ¹

Security and compliance claims (how substantiated?)

SKU Science states on its homepage that “all data is encrypted and hosted on AWS servers,” and its security page lists labels such as “SOC 2 (Type II),” “ISO 27001,” “HIPAA,” and “Privacy Shield.” ¹³ In the reviewed materials, these appear as vendor claims without clearly exposed third-party attestations (e.g., auditor name, certificate number, report availability). Therefore, a skeptical reading is:

• Strongly supported: “Vendor intends AWS hosting and encryption” (self-attested). ¹
• Not independently verified here: Specific certifications/attestations as listed (requires external validation beyond the surfaced pages). ³

Machine learning / AI / optimization: what is actually evidenced?

SKU Science uses “statistical combinations,” automated selection, and trend/seasonality detection language for forecasting. ² This is consistent with classical forecasting/model-selection workflows; it does not, by itself, evidence modern ML architectures (deep learning, probabilistic programming, differentiable optimization) or prescriptive optimization engines.

No public, technical artifact was identified here that would allow a third party to reproduce or scrutinize the internal modeling choices (e.g., whitepaper with model definitions, benchmark methodology, open-source library, patent filings, or academic collaboration explicitly tied to the product). As a result, claims that imply “AI beyond forecasting” should be treated as unproven unless SKU Science provides additional technical disclosure. ²¹

Publicly named clients and case material

SKU Science publishes named customer stories including Bridgestone and Ocean Spray. ⁴⁵ These are useful as named references, but they remain vendor-authored; independent confirmation (e.g., customer press releases, conference talks, procurement references) was not established in the materials surfaced here. Therefore:

• Named, vendor-published references exist: Bridgestone; Ocean Spray. ⁴⁵
• Independently verified references (in this corpus): not established.

Commercial maturity (evidence-based)

Registry presence and an operational SaaS site support that SKU Science is an active commercial entity and product. ¹¹¹¹² The product appears relatively focused (forecasting + S&OP review analytics) rather than a broad APS suite. ² However, public signals about scale (revenue, headcount, large enterprise penetration, long client lists, extensive third-party case studies) are limited in the sources reviewed here. This pattern is more consistent with a commercially active but relatively small vendor than with a large established enterprise software publisher—while acknowledging that private companies can under-disclose. ²¹¹¹²

Conclusion

Publicly available evidence supports that SKU Science delivers a SaaS product centered on demand/sales forecasting, forecast governance (edits, value-added tracking), and S&OP-oriented dashboards and KPIs, including monetary valuation of forecast error. The vendor claims automated selection among “644 statistical combinations” and rapid onboarding, which—absent deeper disclosure—most plausibly indicates a bounded model-selection framework over classical statistical forecasting configurations rather than a clearly documented modern ML stack. Security/compliance labels are presented, but the surfaced pages do not clearly expose third-party attestations, so those should be verified independently for high-stakes procurement. Compared with Lokad, SKU Science looks more like a forecasting-and-review layer, while Lokad positions itself as a programmable predictive-optimization platform with explicit probabilistic forecasting documentation and a project-based delivery model. ^23679814

Sources