Optimal Chain is the only inventory platform that runs a 17-model AI tournament against every single SKU, stress-tests your purchase orders against price hikes and supplier delays in real time, and tells you exactly how to spend your replenishment budget — down to the penny, across thousands of SKUs, in milliseconds.
Inventory teams are stuck in a doom loop: capital locked in stock you can't sell, while your fastest movers stock out three Fridays a month. Here's why.
Millions of pounds of working capital is locked in inventory you'll never sell at full margin. Meanwhile your fastest movers stock out three Fridays a month — and nobody knows the £ damage until the quarter closes.
Your safety stock is calculated by Z × σ × √L — a textbook formula that assumes demand is a perfect bell curve. Yours isn't. It never was. Christmas products look "erratic". Slow movers get classified as "obsolete". Decisions get made on bad maths.
Every PO cycle, your team rebuilds the same model in Excel. By the time it's signed off, the supplier prices have already changed. Enterprise platforms cost £250K and take 18 months. Existing inventory tools use the same broken formulas you already have.
Optimal Chain fixes all of this. World-class inventory intelligence, deployed in days, not months. Built by people who've sat in the procurement chair — for businesses that have outgrown spreadsheets and refuse to spend a quarter-million on Blue Yonder.
Microsoft Dynamics 365 Business Central via native OAuth, or guided CSV upload. We onboard you personally. White-glove setup. Days, not months. We learn your bespoke field names, warehouse codes, and supplier hierarchies — you don't.
In minutes, every SKU is forecast by an AI tournament, classified, projected day-by-day for the next year, and given a stocking policy. Three policies plus your own custom one are computed simultaneously — switch between them in under a second.
Drag a budget slider. Stress-test a port closure. Negotiate a supplier deal with break-even pricing to four decimal places. Export adjusted POs straight to your ERP — fully audited, ready to sign off before Friday's board meeting.
Every other inventory tool gives you a forecast and a reorder point. We give you an inventory operating system — built by people who've actually sat in a procurement chair.
A tournament. Not a single model.
17 specialist forecasting models plus a global LightGBM AI layer compete against every SKU using rigorous multi-fold cross-validation. Then — and this is the difference — the winning models from each specialist group are blended: trend, seasonality, intermittency, and stability are each contributed by the model that's best at that specific component.
For the moments every supply-chain manager dreads.
A price-increase letter lands on Monday. The port strikes on Tuesday. Oil jumps on Wednesday. Friday is the board meeting. The Disruption Planner replaces a week of spreadsheet panic with a 60-second decision tool. Stress-test against supplier delays, cost surcharges, or both stacked together — at portfolio, supplier, or SKU level.
"I have £750K. What should I order?" — answered in milliseconds, every time.
Type a budget. The engine recalculates the optimal purchase order across every supplier and every SKU in your catalogue, in under 5 milliseconds. Drag the slider — watch projected service level move live. Drag the service-level slider — watch the budget cost recalculate. Bidirectional, pre-computed at 80 points, zero server round-trips.
The maths underneath is a world-class MOQ-quantized discrete optimisation engine with integer programming. It scores every SKU by Expected Stockout Cost (revenue × exposure × criticality), pre-reserves cliff items where MOQ forces an all-or-nothing decision, runs a binary search to find the optimal fill rate, then sweeps remaining budget in MOQ increments to the highest-marginal-value gaps. The result: every funded SKU has a physically orderable quantity — no sub-MOQ phantom orders, no spreadsheet-allocator cliff bugs, no starvation, and budget utilisation maximised to the penny.
This is the single most operationally valuable feature in the platform. Procurement teams stop rebuilding the same spreadsheet model every PO cycle and start making informed £-anchored decisions in seconds.
Throw away the textbook formula. Use your actual data.
We don't use Z × σ × √L. Instead, we bootstrap your actual forecast residuals 5,000 times to build the empirical distribution of total error over the lead time, then pick the quantile that matches your service-level target. Works for any error distribution — normal, zero-inflated, lumpy, intermittent, anything. Distribution-free. Self-bounding. The same methodology Amazon uses.
"They're offering 8% off if we take 5,000 units. Should we?"
The Unified Deal Optimiser models up to 8 pricing tiers simultaneously, computes break-even prices to 4 decimal places, runs an optimal-quantity search per tier, and shows you a visual savings waterfall. Includes the Queue Effect — the carrying cost of stock waiting in line behind your existing pile, which most businesses miss entirely and which can entirely erase a 10% discount.
Not basic tables. Two interactive command centres.
The Customer Hub is a 10-section command centre with BCG-style strategic quadrants (Stars / Emerging / Anchor / Review), 12-month lifecycle heatmaps, AI-powered customer stockout risk scoring, and forward revenue projections. The Supplier Hub is a 12-section analytical powerhouse with 5-axis radar scorecards, HHI concentration analysis, 90-day stockout-risk timelines, and a £-quantified action board where every recommendation has an exact pound value attached.
3 ready-to-go policies — Cost Saver, Balanced, High Availability — computed simultaneously in a single analysis run. Plus a fully custom 4th policy where you set the service-level target for every one of the 9 ABC/XYZ segments, persisted between sessions. All four run through the exact same mathematical pipeline. Switch between tabs in under a second. Plus per-SKU service-level overrides for the handful of SKUs you can never afford to stock out on, regardless of policy.
Forecast-error CV (not historical). Christmas products correctly classified as predictable, not erratic.
Day-by-day forward simulation with thresholds that change with the seasonal forecast — not flat averages.
Identifies where minimum order quantities are forcing 12+ months of stock. Hands you the negotiation data.
From any SKU's deep-dive modal, change lead time, unit cost, MOQ, stock quantity, or service level (50%–99.9%) — the entire pipeline re-runs live. Safety stock, reorder point, ROQ, target level, days of cover and order-now quantity all recalculate in front of you. Always simulated from the original pre-optimisation snapshot — no compounding errors. Apply permanently to write the change through to all four policy tabs and the financial aggregates, or revert to original with one click. Customised SKUs are flagged across every view.
Forecasting kernels written in Rust. 30×–200× faster than pure Python. 10,000-SKU catalogue runs in minutes.
Spreadsheets break the moment your catalogue grows. Off-the-shelf tools use textbook formulas from the 1980s. Enterprise platforms cost £250K+ and take 18 months. We built Optimal Chain because none of those options actually solve the problem an inventory team faces every Monday morning.
| Capability | Spreadsheets | Netstock / EazyStock / Slim4 | Blue Yonder / o9 | Optimal Chain |
|---|---|---|---|---|
| Forecast models per SKU | 1 | 2–4 proprietary | 5–10 | 17 + AI, blended |
| Safety stock method | Textbook formula | Textbook formula | Mixed | Empirical bootstrap |
| XYZ classification | Historical CV | Historical CV | Historical CV | Forecast-error CV |
| Disruption stress-testing | None | Basic | Yes | Cost + delay, 3-tier, live counter |
| PO Budget Allocator | Manual guesswork | Basic ranking | Yes | MOQ-quantized IP, <5ms |
| Deal analysis | Manual | Manual | Yes | 8-tier optimal-quantity + break-even |
| Catalogue size | 200–500 | 5K–25K | Unlimited | Any catalogue size |
| Time to deploy | n/a | 6–12 weeks | 12–18 months | Days — we set you up |
| Annual cost | Free + hidden | £15K–£40K | £250K+ | Transparent — see pricing |
Built for the people who know what RMSSE is. If that's not you, hand this section to your data team — they'll get it.
Each SKU is evaluated against ETS variants, Theta, Croston SBA, TSB, ADIDA, iMAPA, Seasonal Naive, SES, plus a global LightGBM AI layer. RMSSE-scored. M5-style multi-fold expanding-window cross-validation.
The winner of each specialist group contributes its best component (level, trend, seasonal shape, demand probability) to a blended bespoke forecast. Most platforms pick one winner. We blend four.
5,000 bootstrap draws of your actual forecast residuals build the empirical error distribution. The quantile at your service-level target is your safety stock. No assumption of normality. No textbook caps.
Forecasting kernels written in Rust. 30×–200× faster than pure Python. A 10,000-SKU catalogue completes in minutes — and the budget allocator runs entirely client-side at <5ms.
Any size. Real warehouses. Real suppliers. Real cash-flow pressure. If your inventory is a board-level conversation, you're our customer.
5,000+ SKUs, hundreds of suppliers, MOQ headaches, port-strike exposure, customer-mix complexity. We turn your replenishment cycle from a three-day spreadsheet exercise into a five-minute decision.
Long lead times, raw-material price shocks, multi-tier suppliers, bill-of-materials ripple effects. Stress-test the next price-increase letter before it lands. Know the £ damage of every supplier delay before the meeting.
Customer-concentration risk, seasonal lifecycle products, fast-moving SKU mix, multi-warehouse complexity. Forecast-error-driven XYZ classification finally gets your seasonal range right.
The single working-capital number you give your CFO. Net of required investment. Auditable line by line.
Order Now Quantity, ROP, ROQ, Target Level, projected stockout date — for every SKU, on every policy.
£-impact-ranked negotiation board. Know exactly who's costing you the most before the next contract review.
Cost Saver, Balanced, High Availability — computed simultaneously — plus a fully custom 4th policy where you set every segment's service level. Sub-second tab switching. Per-SKU overrides for the strategic ones.
Adjusted PO CSVs ready for D365 writeback. Boardroom-ready PDF reports your MD will actually read. Premium multi-sheet Excel deliverables.
Per-SKU Forecast DNA explaining exactly why the AI made every decision. Tournament leaderboard. ML transparency built in.
Optimal Chain is a concierge product. We connect your ERP personally, learn your bespoke data variables, configure the platform for your specific business, and validate the first blueprint with you in person. Once you're set up, you simply refresh your data export whenever you need a fresh analysis.
"You don't configure software. You drink coffee while we configure it for you."
Annual contracts are quoted by email — pricing scales with SKU count and ERP complexity. No surprises, no per-seat fees.
For most customers, we go from "kick-off call" to "first validated blueprint" in 5–10 working days. The pace depends on your ERP and the cleanliness of your data. We do the heavy lifting — schema discovery, field mapping, edge-case handling — so your team can stay focused on the day job.
No. D365 Business Central is our deepest integration (native OAuth, schema discovery, PO writeback), but we work just as well with guided CSV exports from any ERP — SAP, NetSuite, Sage, custom systems, anything. We help you build the export the first time. After that it's a button click.
Honestly: any inventory-holding business. We deliberately built Optimal Chain so the maths scales — a 200-SKU independent retailer gets the same 17-model forecast tournament and empirical safety stock as a 50,000-SKU national distributor. There is no SKU cap, no minimum, no "you're too small" gate.
That said, customers see the biggest £ ROI when inventory is already a board-level conversation — typically when working capital tied up in stock, MOQ pressure, or supplier risk is one of the top three things the management team worries about. If a 10–25% working-capital release would meaningfully change your year, this platform pays for itself many times over.
Yes. Data is encrypted in transit (TLS 1.3) and at rest. UK-region hosting on Microsoft Azure. Your data is never used to train cross-customer models — your blueprint is yours. GDPR compliant. Full audit logs available on request. We can sign your DPA on day one.
Honestly: we like those tools, but they were architected a decade ago. They use 2–4 forecasting models against every SKU; we run a 17-model tournament. They use the textbook Z × σ × √L formula for safety stock; we use empirical bootstrap quantiles — the same method Amazon uses. They classify XYZ on historical demand variability; we classify on forecast error, which means seasonal products finally get classified correctly. And we've added a Disruption Planner and an MOQ-quantized PO Budget Allocator that no competitor offers at all. See the comparison table above for the full breakdown.
It always is. That's why we onboard you personally. We have a Data Quality Dashboard that grades your data on coverage, missing fields, outliers, and date continuity — with plain-English impact statements ("47 SKUs have no sales history; these will use a fallback method with low confidence"). We tell you exactly what to fix and what we can work around.
Email us at contact@optimalchain.co.uk with a quick description of your business — SKU count, ERP, what's keeping you up at night — and we'll come back with a live blueprint of a similar-sized business so you can see exactly what your one will look like. If we both think it's a fit, we'll do a paid pilot on a slice of your catalogue before any annual commitment. We don't do self-service free trials because the platform is too sophisticated to evaluate without proper onboarding — we'd rather earn your trust on a real dataset.
Email contact@optimalchain.co.uk — every enquiry is read and responded to personally, normally within one working day. Tell us about your business, your ERP, and the inventory problems you're trying to solve, and we'll come back with a live blueprint on a similar business's data. No salespeople. No PowerPoint. Just a real conversation.
Get in touch. Tell us about your business and your inventory headaches, and we'll come back — normally within one working day — with a live blueprint on a similar business's data.
Contact Us