Shipper Intelligence
Your network, already modeled
A living digital twin of your supply chain, built from public data before you send us a single file, trued up against your actuals, and able to answer 'what if' in plain English. The $300k network study, replaced by a fixed-fee diagnostic that leaves a living model behind.
The old way: a six-month network study
A network study today means network design software run by specialist modelers, or a consulting engagement: months of data wrangling, a fee in the quarter-million-dollar range, and a model that is stale the day it lands. So most teams only re-run it when leadership panics. Day-to-day questions, like what closing a distribution center actually does to cost and service, wait for an analyst with a model that no longer matches reality. Meanwhile tariffs change twice a quarter and the board gets answered from a spreadsheet.
Every one of those tools demands your data before it shows you anything. That onboarding cost is why the mid-market goes unserved: the integration bill is larger than the deal. The diagnostic flips the order.
| Traditional network study | Hylios Network Diagnostic | |
|---|---|---|
| Time to first answer | Months of data integration first | Hours: the first meeting opens on your twin |
| Duration | Roughly two quarters | 4 to 8 weeks, fixed scope |
| Fee | Roughly $250,000 to $300,000 | $50,000 to $100,000, fixed |
| What you keep | A static deck, stale on delivery | A living twin you can keep interrogating |
| Data required to start | 12+ months of your shipment data up front | None: we start from public data, then true up |
How the diagnostic works
The engagement is fixed scope and runs in four moves. Nothing here is a dashboard subscription. It is a bounded piece of work with a concrete deliverable.
1. Zero-onboarding twin
We build a digital twin of your network from just your company name: facilities, suppliers, products, demand by geography, capacities derived from real building footprints, and cost-to-serve from a multi-echelon flow solver on real road distances. No files from you yet.
2. Live workshop
We walk your team through the twin across four chapters (your network, your true cost-to-serve, where the model deviates from plan, and how the pieces connect), running the what-if scenarios that matter to you live in the room.
3. True-up against your actuals
You send us your files. We overlay them and show you, per metric family, where our outside-in estimate differed from your actuals and by how much. This is the trust mechanism, described below.
4. Findings report
You get a branded PDF with an opportunity register of dollarized findings, each traced to a named scenario run and labeled with its data source and confidence, delivered through a tracked link.
The trust mechanism: we show you where we were wrong
The buyer can check every number we produce against their own books. So instead of hiding the gap between our estimate and reality, we lead with it. After we ingest your files, the report opens each metric family with our outside-in estimate, your actuals, the delta, and what changed in the twin once we trued it up. The sample below shows the format on the Home Depot reference twin.
Sample deliverable, Home Depot reference twin on public data, not a real engagement
| Metric family | Our outside-in estimate | Your actuals | What we do with the delta |
|---|---|---|---|
| Demand by region Sample | Estimated from public store footprint and trade areas | From your shipment history | Re-weight demand, re-run the flow solver |
| Lane volumes Sample | Modeled from DC-to-store assignment | From your freight invoices | Correct lane assignment and mode mix |
| Facility throughput Sample | Derived from building footprints | From your DC throughput data | Recalibrate capacity and handling cost |
| Freight cost-to-serve Sample | Modeled from public rate structures | From your freight invoices | Reconcile transport cost per lane |
In a real engagement the estimate and actuals columns carry your numbers and the honest delta between them. The point is not that the outside-in estimate is perfect. It is that you can see exactly where it was off before you trust anything downstream.
A sample opportunity register
The centerpiece of the findings report is the opportunity register: the dollarized findings, each from a named scenario run against your own twin. First we decompose your cost-to-serve so every finding points at a real dollar bucket. The decomposition below is the Home Depot reference twin's own modeled output.
Sample deliverable, Home Depot reference twin on public data, not a real engagement
Modeled annual cost-to-serve basis: 4,368,104 units per year at a landed cost of $8.72 per unit, about $38.1M per year. Modeled
| Cost bucket | Share of cost-to-serve | Modeled annual amount | Source |
|---|---|---|---|
| Fixed facility allocation | 39.2% | ~$14.9M | Modeled |
| Variable inventory holding | 28.1% | ~$10.7M | Modeled |
| Customs duty | 25.1% | ~$9.6M | Modeled |
| Freight | 4.0% | ~$1.5M | Modeled |
| Handling | 3.6% | ~$1.4M | Modeled |
Each register finding names the cost bucket it targets and the scenario that quantifies it. On a real engagement the delta column carries the dollar result of that scenario run against your trued-up twin. On this sample it names what would be quantified, so we are not putting an invented savings number in front of you.
Sample deliverable, Home Depot reference twin on public data, not a real engagement
| Finding | Targets | Scenario that quantifies it | Delta | Confidence |
|---|---|---|---|---|
| Fixed facility allocation dominates cost-to-serve Illustrative | ~$14.9M fixed allocation | Network consolidation what-if (close or merge weakest DC) | Quantified per engagement from your twin | Low until trued up |
| Inventory holding concentration Illustrative | ~$10.7M variable holding | Safety-stock and replenishment policy what-if | Quantified per engagement from your twin | Low until trued up |
| Import duty exposure Illustrative | ~$9.6M customs duty | Sourcing and FTA what-if, reported as a range | Range only, incidence-caveated (see below) | Low until trued up |
| DC-to-store lane assignment Illustrative | Freight and handling buckets | Lane re-assignment and mode-shift what-if | Quantified per engagement from your twin | Low until trued up |
Scope and price
- What you get: a public-data twin of your network, a live workshop, a true-up of the twin against your actuals, and a branded findings report with the opportunity register, delivered through a tracked link.
- Price: $50,000 to $100,000, fixed fee, fixed scope.
- Timeline: 4 to 8 weeks.
- How it runs: during the diagnostic the engagement runs on Hylios's development environment, stated plainly in the contract. It is a bounded engagement, not an always-on subscription. Your files stay private to your workspace, never enter any shared catalog, and are deleted on written request.
- Availability: we run at most two diagnostics at a time, so the work stays founder-led and the quality bar holds.
See it modeled on your own network
Book a 30-minute working session and we'll build a digital twin of your supply chain and run the scenarios that matter to you.
Or start with the daily supply chain brief
Curated, AI-friendly daily briefing. Free, unsubscribe anytime.
Frequently asked questions
What is a supply chain network diagnostic?
A network diagnostic is a fixed-fee, 4 to 8 week engagement that builds a digital twin of your supply-chain network from public data, walks your team through it in a live workshop, ingests your own files to true the model up against your actuals, and delivers a branded findings report with a dollarized opportunity register. It compresses the kind of analysis that a traditional network study spreads across two quarters into a few weeks, and it leaves a living model behind instead of a static deck.
How is this different from network design software or a consulting network study?
Every incumbent asks for your data before showing you anything: months of shipment history, DC throughput, and freight invoices loaded before the first model runs. Hylios starts the other way around. We build your twin from public data first, so the first meeting opens on your network already simulated, then we overlay your files and show you exactly where our outside-in estimate differed from your actuals. You see value before any integration work, and you keep a model that stays current instead of one that is stale the day it lands.
Do you promise a specific savings number?
No. We do not publish or promise savings percentages. The diagnostic produces an opportunity register of dollarized findings that come from your own network's simulated scenarios, each labeled with its data source and a confidence level. The numbers are yours, generated from your twin, not a benchmark borrowed from someone else.
How accurate are the numbers before I send you my data?
The outside-in twin is an estimate built from public data (facility footprints, filings, store locators, public rate structures). It is precise enough to make the trade-offs visible and to structure the workshop, and it gets sharper when we overlay your actuals. The honest part is the true-up: we show you, per metric family, where our estimate was high or low and by how much, before we quantify any opportunity. A vendor that volunteers where it was wrong earns the right to be believed everywhere else.
What do you need from me, and what happens to my files?
For the true-up we ask for up to 12 months of shipment history, DC throughput, freight invoices, and a store and DC list. Your files stay private to your workspace. They never enter any shared catalog, and they are deleted on written request. The engagement runs on Hylios's development environment during the diagnostic, which we state plainly in the contract; it is a bounded engagement, not an always-on subscription.
Can you benchmark what my carriers should charge me?
During a diagnostic we model your transport cost from public rate structures. Carrier-specific rate benchmarking is on the roadmap pending partner agreements, so it is not part of the diagnostic today.
Want to see it on your own network? The workshop opens on your twin, already built. Read how we construct and label every number at /methodology.