A short diagnosis first.
Then only the work worth doing.
No six-month roadmaps. We start with the one inventory problem costing you most, show you the evidence, and build only what has a clear economic case. If the answer is a setting you already own or a process change, that’s the finding.
See the four steps arrow_downwardHow an engagement moves
Each step is a gate. You decide whether to continue, and you control scope and spend throughout.
A 30-minute fit conversation
Tell us where inventory keeps going wrong: the recurring miss, the systems involved, the job someone does by hand. Nobody can name a root cause in thirty minutes, so we won’t. We’ll say whether a diagnostic is worth it. If it isn’t us, we say so.
The inventory diagnostic
Three weeks, one costly problem, your own data. Real cases traced across your systems, the cost estimated from evidence, every cause named: a setting, a process, a vendor, or something genuinely missing. One problem done properly, not your whole company at once.
Fix what the diagnostic justified
Configure what you own. Connect what you run. Recommend a proven tool where one fits (no referral fees). Build only the piece that’s missing, inside the tools you already use.
Keep it running
Only after a build, and only if you want it. We watch the checks, handle what they raise, and re-tune the rules as your suppliers, channels and tools change.
Where the money usually hides
Different systems don’t have to show the same number. We reconstruct what the state should be, then find the unexplained differences and the expensive manual decisions filling the gaps. Off-the-shelf tools handle standard workflows well. The cost lives in the business-specific seams they don’t understand. We start with four inventory seams; the rest follow with the same method.
The weekly reorder decision
Nothing is broken. Someone just spends half of every Monday deciding what to reorder, when, and how much, from the planner, six spreadsheets and gut feel, on a lead time the supplier promised rather than the one you get.
Where the units should go
A business decision on top of a forecast: channel margins, wholesale commitments, Amazon capacity, a reserve at the 3PL. Every brand splits it differently, and the rule usually lives in one person’s head.
Inbound stock vs sellable stock
Between delivered and available to sell sit receiving, counting, putaway, a sync and a channel update. At your warehouse, then again inside Amazon. Days, sometimes weeks, where you’ve paid for the units and still lose the sale.
Available-to-sell vs what’s actually on hand
Owned, on hand, reserved, in transit, available to promise: each system measures a different one. Some differences are legitimate. The expensive ones are a bundle mapped wrong, a return never restocked, the same unit promised twice.
3PL charges vs actual activity
Duplicate pick-and-pack fees, storage creep, and dimensional-weight overbilling.
Returns vs the ledger
Refunds issued for products that never came back; restocked items written off as lost.
Amazon settlement vs expected settlement
Fees, reserves and chargebacks buried in the deposit; credits you’re owed never claimed.
A clear, plain-language findings report
Not a list of problems you already knew about. A quantified map of where your operation leaks, why, and what to do first.
Exactly where it leaks
Where and how often your systems disagree, down to specific orders, shipments and payouts.
What it costs you
Each problem in hours and dollars, from your own transaction history, with a confidence grade on every number. A lost sale can be estimated, not always measured.
Why it happens, in plain English
A missing setting, a broken handoff, a process gap, or a real integration bug. Named, not guessed.
What to fix — and what to leave
A ranked list: worth fixing now, worth fixing later, not worth fixing at all.
The right kind of fix
Whether it needs custom automation, a process change, or a setting in software you already pay for.
The highest-value next step
The one fix worth making first: automation, configuration, process, or something custom. If it’s an automation, the expected return comes with it.
What a finished audit looks like
An excerpt from a sample diagnostic on a fictional home-and-kitchen brand, Cedar & Field. This one covers returns, a smaller surface than inventory that shows the method clearly. Every number is invented to show the shape of the deliverable. We walk through the full seven-page sample on a call.
What they told us
Returns felt like a money pit. Four systems showed four different numbers. The ops lead lost half of every Monday reconciling them by hand, and had once floated a cost of about $10,000 a month that nobody trusted.
What the audit found
About a quarter of that guess, and shaped differently. Nine in ten of the disagreements between systems were timing or policy, not loss. The hard dollars we could defend were about $12,000 a year. The larger cost was capacity: roughly 675 hours a year of reconciliation that no system was doing for them.
What we recommended
Three must-dos. One setting they already owned in their returns platform. One named owner for a bin nobody was clearing. And one build: a nightly reconciliation across their returns platform, their warehouse’s report and their storefront, delivering a short, dollar-ranked Monday list of only the returns that broke a rule.
Why four systems disagreed: a year of Monday mismatches, classified
| What the spreadsheet flagged | Rows / yr | Leak? |
|---|---|---|
| Normal timing or policy: early refunds, keep-it refunds, scan lag, the nightly sync, re-grades | ~4,280 | No |
| Refunded, but the unit never came back | ~301 | Yes |
| Restocked as sellable, never reached the storefront | ~134 | Yes |
| Total | ~4,715 | 9% leak |
A tool that flags every mismatch would flag all 4,715. That is why no number felt trustworthy: the real 435 were buried under 4,280 that were fine.
Seven pages: scope, method, graded findings, the build we would and would not do, and the safety rules a build runs under.
No lock-in. The audit stands on its own.
You own everything we find. Take it to your own team, to another vendor, or let us fix it. No obligation either way. If the right answer is a process change or a setting you already pay for, we say so and build nothing.
Careful by default
An audit means trusting us with real operating data. These are our rules.
Read-only, by default
Exports first. Scoped, read-only access only where a build needs it. Nothing changes in your systems during an audit.
Scoped to one area
Only the sources that area touches, for 30–90 days. Not your whole business.
No customer personal data
Buyer names, emails and addresses are dropped on ingest. The audit works on orders, SKUs and money, not people.
AI, carefully
Where we use AI on your data, it runs through business-tier APIs on de-identified rows. Never consumer chat tools.
Yours, then gone
You own every finding. Raw exports are deleted after the findings session unless you ask us to keep them for a build.
Confidential from the first note
Happy to sign a mutual NDA before any data changes hands.
Fair questions, straight answers
Most brands we talk to already own good software. The real questions are about what’s left.
Our forecast is probably wrong. Is that what you fix?
expand_more
Sometimes, and then we say so. But we don’t sell forecasting. The forecast is one of five things that can go wrong before a best seller runs out. The others: the lead time the planner uses versus the one you actually get, the promotion nobody told the planner about, stock sitting in the wrong channel, and units that arrived but aren’t yet sellable. The diagnostic tests each against your last year of orders, so the fix lands on the one that caused the misses.
We already have an inventory planner. What is left for you to do?
expand_more
Keep it. A good planner (Inventory Planner, Cogsy, Prediko, Flieber and others) does the arithmetic well when the data underneath is coherent. What it can’t do is own its inputs: whether the counts are true, whether lead times reflect reality, whether the reorder it recommends is what actually gets ordered, and how units split across Amazon, your store and wholesale. That layer around the planner is where we work, and where most misses come from.
Do you take referral fees from the tools you recommend?
expand_more
No. When a proven tool fits, we say so and help you adopt it, and take nothing from the vendor. Independence is the point of the recommendation.
Which systems do you work with?
expand_more
The ones you already run. Shopify and other storefronts, Amazon (FBA, AWD, FBM) and other marketplaces, wholesale and retailer orders, your 3PL’s portal or warehouse system, your planner, and NetSuite, Cin7 or QuickBooks. If it has an export or an API, it can be in the diagnostic. If it only has a spreadsheet, that counts too.
Why not just buy an inventory platform like Cin7 or Linnworks?
expand_more
Sometimes you should, and we’ll say so. Those products keep one master count in sync across channels. Our work starts after that: the software is installed and someone still spends every morning reconciling numbers by hand. What survives the purchase is business-specific (your policies, your contracts, your warehouse’s habits), and no off-the-shelf product ships it.
Are you replacing our existing systems?
expand_more
No. We work beside the systems you run, read-only during an audit. A common outcome is fixing the configuration of software you already pay for. If a system genuinely needs replacing, that’s a recommendation you can take to any vendor, not a project we push.
Are you an inventory management platform?
expand_more
No. We’re a small practice, not a product. We diagnose where an operation leaks money or hours, fix what’s worth fixing, and maintain what we build. Nothing to subscribe to.
What happens after the audit?
expand_more
You get a plain-language findings report you own outright, each leak quantified and traced to its cause. Then it’s your call: your own team, another vendor, or us. The audit stands on its own either way.
Do you always end up building custom software?
expand_more
No. An audit that always recommends a build shouldn’t be trusted. The honest order: do nothing where a difference is normal, fix settings in software you own, fix the process, take the evidence to your warehouse or vendor, and only then build the piece that’s genuinely missing. In our worked sample, two of three recommendations aren’t software at all.
What if the problem is actually our 3PL, or how a tool is configured?
expand_more
Then that’s the finding. The audit traces each discrepancy to specific orders, units and timestamps, which turns “our 3PL’s numbers are always wrong” into “here are the exact cases where the handoff failed.” That evidence is useful in a vendor conversation whether or not we build anything.
How is this different from hiring an operations consultant?
expand_more
Two ways. The diagnosis runs on your actual transactions, not interviews, so findings come with specific orders and dollar amounts attached. And the engagement ends in something that runs: a fix built, monitored and maintained, not a slide deck.
Start with a conversation
A 30-minute fit conversation with Ankita and Aikansh. If it isn’t us, we say so.
Book a 30-minute fit conversation calendar_today