Last updated: November 4, 2026
Go back to Market Research
StockIQ is supply chain planning software for distributors and manufacturers. It connects to ERP systems to forecast demand, set inventory policies, generate purchase and transfer orders, plan manufacturing replenishment, monitor suppliers, and support sales, inventory, and operations planning. Its Edison assistant adds natural-language questions and actions within the application. (1, 8)
Overall supply chain score: 5.8/10
| Dimension | Score |
|---|---|
| Supply chain depth | 4.8/10 |
| Decision and optimization substance | 5.4/10 |
| Product and architecture integrity | 6.4/10 |
| Technical transparency | 6.8/10 |
| Vendor seriousness | 5.6/10 |
StockIQ’s governing process is conventional inventory planning: select forecasts, calculate stock protections, apply replenishment policies, and adjust orders to supplier and transport constraints. Its manuals expose enough detail to examine these operations. They also expose the economic limitation: matching coverage targets or satisfying an assigned service level does not establish that the resulting purchases make the best use of scarce cash, capacity, and inventory. (2, 4, 6, 22)
There is substantial operational substance behind the product. Scheduled orders can reach ERP without individual approval, and a named customer reports a large reduction in ordering work. However, the public methods do not establish a calibrated joint model of demand and supply uncertainty or a common economic objective across the decisions. Edison adds a new interface while preserving human review; its effect on labor beyond the established automation remains unmeasured. (5, 7, 8, 11)
StockIQ versus Lokad
StockIQ sells a configurable planning application, onboarding, and ongoing customer support. Planners select policies and maintain the assumptions that determine the application’s output. Lokad’s managed inventory optimization service assigns Supply Chain Scientists to develop and maintain the numerical decision process, including data interpretation, economic objectives, and production logic. The difference concerns who engineers the decisions and what the software optimizes, not simply who supplies the interface. (10, 22, 23)
StockIQ’s documented ordering methods make this distinction concrete. Equalizing coverage across products can synchronize replenishment, but a day of coverage has different economic value for a high-margin replacement part, a readily substitutable accessory, and a seasonal product approaching obsolescence. Lokad’s inventory optimization approach ranks feasible actions by their expected economic consequences under uncertainty. The comparison turns on whether the decision process can express those differences and trade them against shared constraints. (6)
StockIQ has delivered labor savings through ERP integration and order preparation. Lokad must demonstrate an additional gain against that existing automation: better decisions, less recurring review, or cheaper adaptation when business conditions change. Supply Chain Scientists’ maintenance time also belongs in that comparison. Neither a configurable application nor a managed service escapes the cost of correcting a faulty economic model. (7, 11)
Forecasting, uncertainty, and purchasing decisions
A disclosed forecasting process
StockIQ’s proprietary forecast combines several approaches and borrows seasonality from higher levels of the product hierarchy. The manual describes regression-based trends, seasonal coefficients, volatility settings, and bias detection. It makes a useful observation: more reactive settings can reduce historical forecast error while making purchasing worse. That warning connects statistical behavior to the decision it drives. (2)
The tournament runs candidate algorithms and parameter combinations against a model backdated three months, selects the lowest-error result, then refits on current data. It currently applies to recurring demand; sporadic, slow, dead, and new items use the StockIQ algorithm directly. This is an inspectable holdout procedure, although the documentation does not provide a published dataset, complete search space, or performance by replenishment horizon. A single selection window can favor a temporarily successful model; repeated selection needs an evaluation period that remains independent of that tuning. (3)
The safety-stock screen exposes the forecast-error input, including its source and sample length, and identifies wMAPE as total units missed divided by total units sold. It distinguishes statistical stock from active overrides, offers several calculation methods, and includes cumulative lead times for bills of materials and hub-and-spoke relationships. Those details help trace an inventory recommendation. They do not reveal a full predictive distribution or establish that the assigned service target is economically justified. (4)
Lead-time measurement does not complete the uncertainty model
StockIQ measures administrative, manufacturing, shipping, and putaway delays. For several item, supplier, and site groupings, it computes means and standard deviations, removes observations outside configured tolerances, recalculates, and adds a planning allowance. It then chooses among increasingly broad fallback groups and caps changes relative to ERP lead times. The resulting components form planning durations. (5)
This handles unreliable ERP lead-time master data. It also creates important modeling choices. A late shipment can be a bad timestamp or a genuine disruption; filtering both the same way removes the very tail that determines the value of protection. A capped update may stabilize ordering while delaying the response to a deteriorating supplier. The public method does not show how demand surges, supplier delays, and shared transport disruptions are combined into the distribution of future inventory outcomes. Those dependencies matter when many products become exposed at once.
An April 2026 article recognizes that model accuracy is insufficient for inventory decisions. It proposes support, stability, and reliability measures, and explains why AUC is not a percentage of correct decisions. This is a better question than an accuracy contest. The article does not disclose how its combined confidence score is calibrated or translated into the monetary consequences of ordering, waiting, or discontinuing a product. A stable recommendation can still be consistently expensive. (14)
Constraint handling is concrete; the objective remains fragmented
The order-policy catalogue includes order point with EOQ, lot-for-lot, min/max, kanban, and order-driven replenishment. Its sporadic-sales policy disregards the independent forecast and attempts to maintain safety stock. These policies create executable behavior, but partitioning demand into categories also partitions the decision logic. An economic comparison between stocking a slow spare, expediting it, and accepting downtime needs more than a category assignment. (22)
The Order Wizard adjusts quantities for weight, volume, pallets, container capacity, minimums, and pack sizes. Its disclosed methods include adding multiples in rotation and equalizing order coverage, projected coverage, or cycle stock. It can bring additional products into an order as coverage expands. A price-break option compares discounted unit costs with holding-cost penalties. These are practical purchasing mechanisms, with a narrower economic calculation for discounts. (6)
Consider two items competing for the last pallet position in a container. Equalizing their coverage does not price the lost margin from excluding either item, the chance of a late next shipment, substitution, or the risk of remaining unsold. A feasible, full container can still contain the wrong inventory. The public documentation does not establish a joint decision rule that evaluates those alternatives against the same economic objective.
Productivity and configuration maintenance
StockIQ’s auto-ordering settings support supplier/site schedules, triggers after the first daily refresh or at a specified time, and direct approved release or routing for review. Notifications identify orders placed, skipped, or referred for review. This provides a concrete way to eliminate routine order handling. The remaining questions are the share of orders actually released automatically, the work required to keep the policies sound, and the cost of exceptions. (7)
Edison, launched in October 2026, can answer questions, explain recommendations, create purchase orders, change planning parameters, and update priorities. Its current FAQ says the team reviews every action before Edison executes it. The product therefore retains a human decision queue. It may reduce the effort of finding and entering information, but no published before-and-after measurement establishes the additional productivity of Edison over StockIQ’s existing ordering automation. (8, 25)
The key operational measure is total work per reliable purchasing decision: exception review, parameter changes, data corrections, integration support, and recovery from mistakes. A quicker conversation about each SKU does not remove the dependence on planners processing those SKUs. Human approval also needs a workable rate and sufficient context; otherwise the approval becomes a habitual click.
Configuration creates another labor requirement. Forecast responsiveness, classification thresholds, service targets, lead-time filters, fallback rules, order policies, and overrides interact. A supplier disruption can require changes in several places. The documentation explains individual controls, but does not show a general mechanism for versioning the complete customer policy, replaying it against prior operating conditions, and safely promoting a revised policy. (2, 4, 5, 22)
A stockout investigation should therefore reach beyond increasing a buffer. Perhaps sales history concealed censored demand, the substitute relationship was wrong, or the replenishment options omitted an economically attractive expedite. Fixing that kind of failure requires changing the formulation and retesting its consequences. A growing collection of item-level exceptions can conceal the original problem and make the next revision more expensive.
The implementation scope puts prices on part of this work
The standard scope covers one established ERP, one nightly refresh, validation of ten required feeds, core configuration, an order writer, twenty hours of training, and five weeks of post-launch consultation. Additional integrations, custom outputs, retraining, structural changes, and phased rollouts fall under services priced at $250 per hour. Extended adoption support is listed at $1,500 per week, usually for two or three one-hour meetings. (10)
Those boundaries make the initial commitment unusually inspectable. They also show why subscription price cannot represent ownership cost. Acquisitions, new business units, ERP changes, and employee turnover can create separately billed work while consuming customer specialists’ time. StockIQ’s pricing page offers unlimited users and describes continuing customer-success support, but supplies no public subscription tariff or measured annual configuration workload. (23)
There is also a concrete limit to the launch’s real-time forecasting language: the standard implementation imports ERP data nightly. Continuously reconsidering an unchanged snapshot cannot reveal transactions that have not arrived. A more frequent integration may be possible, but its refresh schedule and operating cost must be part of the deployment specification. (10, 25)
Customer results and economic accountability
Jofran’s case study dates implementation to 2018 and describes a custom SharePoint purchase-order integration. Supply Chain Manager Deniece Fredericks says ordering takes about one-tenth as long; the document reports a 70% reduction in stockouts. President and CFO Kristen Connell attributes at least 10 times first-year ROI on inventory savings to StockIQ. The case also describes vendor changes made to close identified gaps and a later ERP transition. (11)
These are named accounts of operating improvements and corrective engineering. The financial statement leaves an important accounting question unanswered: does inventory savings mean cash released from stock, avoided carrying costs, write-offs prevented, or a combination? A one-time release of working capital and a recurring profit increase are different benefits. The case does not expose the investment denominator, service costs, or a counterfactual purchasing policy.
Auveco’s story reports a 60% reduction in excess inventory, service levels above 95%, stockouts below 7%, and payback within six months. It dates the Dynamics GP implementation to 2020 and describes the later move to NetSuite. Operations executive Josh Wood supplies the customer attribution. Excess inventory is a subset of inventory, however, and its definition can change with the target policy. The page does not provide the starting values or the calculation tying those operational percentages to payback. (12)
The homepage also displays a testimonial about forecast error falling from 20% to 5%. Without the error formula, forecast horizon, SKU/location resolution, weighting, and treatment of zero demand, this cannot support a claim of generally attainable 95% forecast accuracy. Aggregating demand can reduce measured error while hiding the item-level errors that drive orders. The reported operational improvements must stand on their own evidence. (1)
StockIQ’s service-level manual helps expose another measurement choice. It prefers comparing actual shipments with requested orders, but can estimate service from daily demand and opening stock when shipment detail is unavailable. Non-stock items can be excluded. These alternatives answer different questions: estimated availability cannot establish actual on-time, in-full delivery. Customer case comparisons need consistent definitions across their before-and-after periods. (13)
The public implementation scope assigns customer staffing and validation responsibilities. The public EULA points to a signed master agreement obtainable from StockIQ; it does not publish the agreement. Consequently, the public record establishes an implementation deliverable and customer-attributed benefits, but not a defined economic failure threshold or a commitment to remediate disappointing returns. Attribution should follow the evidence: faulty vendor rules, customer overrides, execution failures, and missing input data require different corrections. (10, 24)
Technology history and delivery
StockIQ was founded in 2015. Its April 2019 inventory-classification article recommends ABC/XYZ segmentation and differentiated stocking strategies, including attention to irregular demand. The current policy documentation retains this structure. The continuity is visible: the application has developed around configurable inventory-planning rules rather than replacing them with a unified economic decision model. (15, 16, 22)
The April 28, 2025 Serent Capital partnership announcement promised investment in AI forecasting and price-elasticity modeling. Huron Peak’s October 2025 release then documented elasticity curves, price-change simulations, and variant-planning bills. It also recorded the migration from .NET Framework to .NET Core, SSO changes, and numerous ordering and calculation fixes. This is identifiable engineering follow-through. The elasticity feature uses supplied or customer-edited curves; publishing a simulation interface does not establish that those curves are causally estimated or reliable outside the observations used to fit them. (16, 17)
The AI language deserves a separate assessment. The tournament manual calls model selection and combination an AI/ML ensemble technique. StockIQ’s AI + IQ material also groups unusual-sales detection, end-of-life signals, and customer-ordering alerts under AI. These descriptions identify functions, but do not publish the incremental benefit of a new learning method over the preceding implementation. The technical value lies in the method and its decision consequences, not the expanded label. (3, 26)
Edison’s October 2026 launch adds a language interface and workflow actions. Its broad assertion that it understands inventory logic lacks a published failure analysis or evaluation protocol. At the review date, this is a recent release with announced capabilities; the longstanding customer savings do not measure its contribution. StockIQ’s roadmap shows delivered application and infrastructure work, alongside a much less substantiated step from conventional planning to AI-led improvements. (25)
Software bill of materials
Deployment instructions and an employer-authored engineering vacancy reveal two stages: the established product’s runtime upgrade and a newer platform-modernization effort.
| Component | Attributable role and scope |
|---|---|
| .NET Framework → .NET Core / .NET 8 | Huron release and installation instructions identify the application migration. Hosted customers with a local sync agent also have runtime prerequisites. (17, 18) |
| React, Next.js, TypeScript | Frontend of the platform described in the current full-stack vacancy. (9) |
| C# REST APIs, Dapper, PgBouncer | Application and database-access layers in that modernization work. (9) |
| TimescaleDB / PostgreSQL | A schema-per-tenant database deployment is explicitly described. (9) |
| SQS, OpenTelemetry, Grafana Loki | Event handling, distributed tracing, and log collection in the same engineering remit. (9) |
The vacancy describes a small engineering team in the early stages of modernization. It does not provide a migration timetable or say which existing customers use the new stack. The unresolved operating cost is maintaining and moving between product generations: data conversion, connector behavior, authentication, and customer-specific configuration all need testing. (9)
The components have identifiable roles, but their combined cost includes database tuning, queue recovery, telemetry retention, application patching, and tenant migrations. A schema per tenant also requires a clear answer to who can cross schema boundaries and how those privileges are enforced. The public materials do not quantify infrastructure cost per planning workload or the effort of upgrading configured customer instances.
Security and integration boundaries
StockIQ publishes concrete authentication mechanisms. The Huron migration moves SSO to OIDC authorization flow with PKCE and documents the identity-provider changes needed to preserve login. Its public API guide describes role and business-data scopes, one-way storage of token hashes, and exchange for an hour-long JWT. These controls can limit credential exposure and the data available to an integration. (19, 20)
The API guide nevertheless gives inconsistent expiry instructions: the token-creation form offers an expiry date, while the authentication section says the underlying secret does not expire. An operator needs to know which rule the running version enforces. The documentation also leaves the production evidence for cross-tenant enforcement, restore testing, and Edison’s access to model providers unresolved. Human approval of an order does not address disclosure of the data used to propose it. (20, 8)
Integration is more inspectable than the internal forecasting engine. The API publishes endpoint documentation in HTML and JSON, including filters and pagination. An order-file interface documents headers and lines for purchase, transfer, and work orders, with a database-output alternative. It describes test orders and retaining enough internal information to reproduce exports. This supports independent integration work and tracing the handoff to ERP. (20, 21)
The boundary still matters for recovery: an ERP write must not become a duplicate purchase when a transfer retries, and a regenerated order must not overwrite a later commercial commitment. The public order-file instructions do not specify an end-to-end idempotency and reconciliation protocol. Exporting records also does not make StockIQ’s forecasting and policy implementation portable to another engine.
Supply chain score
Supply chain depth
| Criterion | Score |
|---|---|
| Economic framing | 4/10 |
| Decision end-state | 7/10 |
| Conceptual sharpness on supply chain | 6/10 |
| Freedom from obsolete doctrinal centerpieces | 3/10 |
| Robustness against KPI theater | 4/10 |
Average: 4.8/10.
Economic framing — 4/10. Discount and carrying-cost comparisons address part of purchasing economics. The governing coverage and service targets do not establish a common objective for competing products, scarce capacity, and uncertain supply. (4, 6)
Decision end-state — 7/10. Orders have a documented execution path, including automatic release. The coverage and service policies still leave the economic choice among feasible actions incompletely specified. (7, 21)
Conceptual sharpness on supply chain — 6/10. The manuals distinguish demand patterns, supplier-delay components, and the consequences of reactive forecasts. These distinctions support implementation; their aggregation into policy rules loses some of the dependencies between products and uncertain events. (2, 5, 22)
Freedom from obsolete doctrinal centerpieces — 3/10. Classification, EOQ, stock buffers, and coverage remain central to the decision process. Their economic assumptions become increasingly restrictive when substitution, obsolescence, and shared constraints determine the best purchase. (15, 22)
Robustness against KPI theater — 4/10. The reliability article correctly separates statistical discrimination from good decisions. Undefined headline forecast error and configurable service measurements nevertheless weaken comparisons between customer outcomes. (1, 13, 14)
Decision and optimization substance
| Criterion | Score |
|---|---|
| Probabilistic modeling depth | 4/10 |
| Distinctive optimization or ML substance | 4/10 |
| Real-world constraint handling | 7/10 |
| Decision automation and labor productivity | 7/10 |
| Resilience under real operational complexity | 5/10 |
Average: 5.4/10.
Probabilistic modeling depth — 4/10. Forecast error and measured supply variability influence planning. The disclosed methods do not establish calibrated joint demand and lead-time distributions or economic evaluation over their possible outcomes. (4, 5, 14)
Distinctive optimization or ML substance — 4/10. Holdout selection and explicit ordering heuristics make the numerical process inspectable. Published evidence does not establish a distinctive improvement in decision quality against strong conventional alternatives. (3, 6)
Real-world constraint handling — 7/10. The purchasing mechanics cover practical supplier and transport restrictions. Their limitation is the objective used to decide which feasible quantities to choose when those restrictions compete. (6, 22)
Decision automation and labor productivity — 7/10. Scheduled execution and the named Jofran testimony establish meaningful work reduction. Total remaining labor and Edison’s incremental effect have not been measured publicly. (7, 8, 11)
Resilience under real operational complexity — 5/10. The Jofran account describes vendor changes that closed operating gaps. Public controls support adjustment, but a complete process for revising data semantics, testing alternative formulations, and safely deploying those corrections remains unexposed. (11, 17, 22)
Product and architecture integrity
| Criterion | Score |
|---|---|
| Architectural coherence | 6/10 |
| System-boundary clarity | 8/10 |
| Security seriousness | 6/10 |
| Software parsimony and ownership cost | 5/10 |
| Compatibility with programmatic and agent-assisted operations | 7/10 |
Average: 6.4/10.
Architectural coherence — 6/10. The runtime migration and modernization stack have identifiable purposes. Customer migration, coexistence, and retirement of the preceding architecture are not sufficiently documented to establish the long-term operating shape. (9, 17, 18)
System-boundary clarity — 8/10. Feed responsibilities, deployment prerequisites, and order outputs give implementers a concrete boundary. Transaction reconciliation across failed or repeated ERP writes remains insufficiently specified. (10, 18, 21)
Security seriousness — 6/10. Scoped credentials, hashed secrets, and modern SSO provide assessable mechanisms. Contradictory expiry guidance and incomplete disclosure of isolation and assistant-data boundaries leave material controls unverified. (19, 20, 8)
Software parsimony and ownership cost — 5/10. Published service boundaries expose several recurring expenses. The larger uncertainty is the combined cost of customer policy maintenance and platform migration; neither has a measured operating baseline. (9, 10, 23)
Compatibility with programmatic and agent-assisted operations — 7/10. Documented extraction and execution interfaces support external automation. Customer control over the underlying numerical methods remains narrower than control over records and application actions. (20, 21)
Technical transparency
| Criterion | Score |
|---|---|
| Public technical documentation | 8/10 |
| Inspectability without vendor mediation | 7/10 |
| Portability and lock-in visibility | 6/10 |
| Implementation-method transparency | 7/10 |
| Evidence density behind technical claims | 6/10 |
Average: 6.8/10.
Public technical documentation — 8/10. The manuals expose algorithms, policy behavior, interfaces, and release changes in unusual detail. Full model specifications and reproducible performance evaluations remain missing. (2, 3, 17, 20)
Inspectability without vendor mediation — 7/10. A prospective customer can examine material calculation and integration choices before a sales demonstration. Independent reproduction of the core forecasting and stock calculations is still constrained by undisclosed implementation details. (4, 5, 6)
Portability and lock-in visibility — 6/10. Data and order outputs have documented formats. Reproducing the configured decision behavior outside StockIQ would require more than exporting those records, and contractual exit provisions are not public. (20, 21, 24)
Implementation-method transparency — 7/10. The scope specifies deliverables, customer contributions, training, and extra-service prices. Ongoing model ownership and regression testing are less clearly bounded than the initial installation. (10)
Evidence density behind technical claims — 6/10. Operational manuals substantiate many product functions. The broad AI claims have substantially less supporting evaluation detail than the established planning and integration features. (3, 17, 25, 26)
Vendor seriousness
| Criterion | Score |
|---|---|
| Technical seriousness of public communication | 6/10 |
| Technology roadmap integrity | 7/10 |
| Conceptual sharpness | 5/10 |
| Economic accountability | 5/10 |
| Defensibility in an agentic-software world | 5/10 |
Average: 5.6/10.
Technical seriousness of public communication — 6/10. Method documentation and candid warnings about statistical behavior support serious evaluation. Broad assertions about AI understanding outrun the published evidence for that newer capability. (2, 14, 25)
Technology roadmap integrity — 7/10. The elasticity commitment has visible delivery, and the runtime and identity changes have release documentation. The remaining gap concerns the incremental numerical and economic benefit of the newer AI positioning. (16, 17, 26)
Conceptual sharpness — 5/10. StockIQ recognizes that a good statistical score can produce a poor decision. Its general decision process still relies on proxy targets without a published synthesis of uncertainty, economic value, and shared constraints. (2, 6, 14)
Economic accountability — 5/10. Named customers attribute gains to StockIQ and the implementation obligations are visible. The public evidence does not define net-return evaluation, economic failure, or responsibility for remediation when the expected gains fail to materialize. (10, 11, 12, 24)
Defensibility in an agentic-software world — 5/10. ERP integrations, maintained business logic, and operating customers provide more substance than a conversational wrapper. Durable differentiation still requires better decision economics and lower total work than competitors can obtain from increasingly accessible automation. (7, 11, 20, 25)
Sources
All sources were accessed on November 4, 2026.
StockIQ homepage. Product scope and forecast-error testimonial.
[2] StockIQ Forecast Algorithm
Algorithm manual, updated September 29, 2025. Forecast construction and responsiveness.
[3] Tournament Forecast Algorithm
Tournament manual, updated September 29, 2025. Selection and eligibility.
Safety-stock manual, updated October 30, 2026. Inputs, methods, and overrides.
[5] Shipment/Receipt-based Lead Time Calculations
Lead-time calculation manual. Filtering and assignment logic.
Order Wizard manual. Constraints and quantity-adjustment methods.
Auto-ordering configuration. Release triggers and approval settings.
Edison product page. Actions and review requirement.
[9] Full Stack Engineer vacancy
StockIQ’s employer-authored vacancy, observed in 2026. Platform modernization and component roles.
[10] Standard Implementation Scope
Implementation scope. Deliverables, responsibilities, and additional-service charges.
[11] Jofran Furnishing Case Study
Customer case PDF, undated, hosted under June 2026. Named customer statements.
Auveco case. Operating metrics and implementation history.
Service-level calculation manual, updated August 27, 2025. Measurement alternatives.
[14] How StockIQ Quantifies AI Decision Reliability in Inventory Planning
Jake Latham’s article, April 29, 2026. Confidence measures and decision claims.
[15] Inventory Classification 201
Jake Latham’s article, April 19, 2019. Historical planning doctrine.
[16] StockIQ Teams Up with Serent Capital
Serent Capital’s Business Wire release, syndicated by FinancialContent, April 28, 2025. Company origin and product commitments.
Version 2025.10.30, updated November 7, 2025. Feature and runtime changes.
[18] .NET 8.0 Install Information
Installation instructions. Deployment prerequisites.
[19] StockIQ SSO Changes for Mt Huron
SSO migration instructions. Authentication protocol and upgrade requirements.
Public API guide, updated September 11, 2026. Credentials, scopes, and extraction behavior.
[21] Consume StockIQ Order Files into your ERP
Order-interface instructions. Transaction outputs and testing.
[22] What are the various order policies?
Order-policy catalogue. Replenishment behavior.
Pricing and service description. Commercial sizing and support.
[24] End User License Agreement
Public EULA page. Reference to the signed master agreement.
[25] AI Supply Chain Intelligence Agent
StockIQ’s supplier-authored announcement in Packaging World, October 9, 2026. Edison launch claims.
StockIQ’s AI overview. Feature positioning and dated webinar programme.