We run a tight practice on purpose. Every engagement we take falls into one of four lanes — and each one has a defined entry point, defined output, and defined economic payoff. If your problem doesn't fit cleanly, we'll tell you that before you write a check.
Most CPG planning teams are running statistical baselines from 2012, layered with manual overrides, layered with PowerPoint apologies. We don't tear it out — we put intelligence on top of it.
Who this is for: Operators running SAP IBP, Oracle, Blue Yonder, Kinaxis, or homegrown planning, with $200M+ in finished-goods inventory and forecast accuracy somewhere between "unacceptable" and "actively embarrassing."
Typical engagement: 12–20 weeks, $250K–$900K depending on scope. Outcome is measured in days inventory outstanding, fill rate, and working capital released — not adoption metrics.
Layer ML on your existing planning stack.
Promo lift, weather sensitivity, channel-mix shift, and lifecycle modeling — added on top of your IBP investment, not in place of it. We pick the SKUs where ML actually helps and leave the rest alone.
Right-size safety stock across plants, DCs, and retailer hubs.
Move stock from where it's idle to where it's earning. Typical outcome: 12–22% working capital release without service-level erosion.
Make planning actually drive decisions.
If your monthly S&OP produces a deck nobody acts on, the process is broken — not the people. We rebuild cadence, decision rights, and the meeting itself.
Find the 20% earning 80%.
Rationalize the long tail without losing margin or shelf space. We pair quantitative analysis with category and retailer politics — because both are real.
Every operator we talk to is somewhere on the same map: pausing imports, qualifying suppliers in Vietnam / Mexico / India, rebuilding landed-cost models, redrawing the distribution network. The work is enormous and the timeline is now.
Who this is for: Brands with meaningful import exposure to high-tariff jurisdictions, mid-scale separation events (carve-outs, spin-offs), or distribution networks designed for a world that no longer exists.
Typical engagement: 16–40 weeks, $400K–$2.5M. Outcomes are measured in dollars of margin protected, days of qualification cycle compressed, and risk concentration reduced.
Get out of one country and into three — without breaking the line.
Country-of-origin analysis, supplier discovery and qualification, dual-sourcing roadmaps, and qualification-cycle compression. We use generative AI for supplier research and RFI synthesis — same depth as a big-firm engagement, fraction of the burn.
What if Section 301 goes to 60%? What if it goes to zero?
SKU-level landed cost models that flex on tariff regime, FX, freight, and supplier mix. Built for finance and ops to share — same model, same numbers, no more parallel spreadsheets.
When your ports change, your network changes.
If your DC footprint was optimized for Long Beach and your goods now land in Houston, your network is wrong. Greenfield modeling, consolidation analysis, and rebid support.
When a business unit separates, inventory becomes the messy middle.
TSA design, separated planning systems, supplier contract bifurcation, shared-SKU governance. We've run this for both the parent side and the spinco side.
We are aggressively unimpressed by AI strategy decks. The companies actually capturing value are the ones with three or four use cases in production, generating real cycle-time reduction. That's the bar.
Who this is for: Operators who've heard the AI pitch, kicked the tires on a pilot, and want a roadmap that pays back in months — not a vision deck.
Typical engagement: 4–16 weeks, $75K–$600K. The Opportunity Assessment is a 4-week fixed-fee entry point.
3–5 use cases, sized in dollars, sequenced by ROI.
A structured sprint to identify, size, and prioritize 15–25 AI use cases across supply chain, finance, customer service, and commercial ops — then pick the ones to actually start. Output is one document your CFO can read in 20 minutes.
Request Assessment →Where repetitive judgment meets repetitive data.
Procurement (PO generation, supplier comms, expediting), customer service (retailer chargebacks, deduction management), back-office (master data, vendor onboarding). We pick the targets and build the agents.
Your contracts already contain the answers.
Vendor agreements, retailer trade terms, compliance documents — extract, normalize, and monitor at scale. Often a 90-day payback on its own.
Production AI needs tests, not vibes.
Eval frameworks, policy and acceptable-use design, vendor selection (Anthropic, OpenAI, Azure), prompt engineering training for internal teams. The discipline that separates "we have AI" from "AI works."
The unglamorous practice. If your item master has 30,000 SKUs and three definitions of "active," no AI initiative will survive. We do the work most consultancies won't touch — and we do it without trying to upsell you onto a 24-month digital transformation.
Who this is for: Operators whose downstream analytics, forecasting, and AI keep stalling on data quality issues — and who are tired of being told "it's a data problem" without anyone fixing the data.
Typical engagement: 12–32 weeks, $300K–$1.5M. Outcomes are measured in data quality scores, downstream tool adoption, and how many "the numbers don't match" emails the CFO stops receiving.
Item, supplier, location — one source of truth.
Governance design, data steward operating model, deduplication, taxonomy. The work that makes every downstream initiative cheaper.
Real-time visibility, plant to retailer shelf.
Cross-system dashboards with DSI / weeks-of-supply by customer, exception alerting, and predictive stockout signals. Built to be operated, not admired.
Pull sell-through into sell-in.
Walmart Retail Link, Amazon Vendor Central, Target POL, Kroger 84.51° — feed downstream demand signals into your forecasting engine. Most CPGs leave this on the table.
Power BI / Tableau / Looker — shipped, not styled.
If your dashboards are slow, contradictory, or unloved, the issue is upstream. We rebuild the semantic layer and rationalize the report sprawl.
Four weeks. Fixed fee. We tell you which practice (or which combination) actually pays back — and which to deprioritize. Most engagements start here.
Request AI Opportunity Assessment