You have the question. You have the data. Nobody has the week.
Not everything needs a system. Sometimes there's one specific thing the business needs to know, the data to answer it already exists in exports, and the only missing ingredient is somebody with an uninterrupted week and no other job that month.
One question, answered in a week: $1,800. $900 for the first three.
Two lines by email · no obligation · answered by the person who would do the work
The questions that come up most
Where margin actually goes
Contribution by channel, by SKU family, after shipping, returns, discounts and fulfilment cost — not the version that stops at gross.
Why returns moved
Which products, which reasons, which cohort, and whether the change is a product problem, a sizing problem or a photography problem.
Which SKUs stopped paying
Slow movers that still occupy pallet positions and working capital, ranked by what they cost to keep.
What peak actually costs
Surcharges, overtime, split shipments, expedites — assembled from the invoices and the shipment log after the season, while it's still arguable.
Whether a 3PL is performing
On-time, accuracy, cost per order over time against what was promised, in a form you can put in front of them.
Is the discount working
Incrementality on a promotion, done honestly: what would likely have sold anyway, and what the discount bought.
“Which channel is actually profitable?” — what the week looks like
The most common question we get, and a good illustration of why the answer takes a week rather than an afternoon: the question contains four decisions before any arithmetic happens.
Decide what “profitable” includes
Product cost and shipping, obviously. Then: fulfilment per order, payment fees at 1.5–3.5% depending on channel, returns at whatever your rate actually is, discounting, marketplace commission at 8–15%, and advertising — which is where most channel comparisons quietly break, because platform-reported attribution flatters the channel that reports it.
Reconcile before analysing
Channel revenue in the store rarely equals channel revenue in accounting: settlement timing, refunds, gift cards, tax treatment. If the two disagree by more than a percent or two, the analysis waits until we know why — otherwise the conclusion inherits the error.
Attribute the shared costs honestly
Warehouse minimums, storage, software, the person who handles wholesale paperwork. Allocated per order, per unit or per revenue dollar — the choice changes which channel looks worst, so it is made explicitly and stated in the answer rather than buried.
Test whether the answer survives
Re-run on a different period and a different allocation basis. A conclusion that flips when storage is allocated per unit instead of per order is a conclusion with a caveat attached, and you get told that instead of a clean chart.
Typical outcome: one channel is 5–15 percentage points worse than the reported view, usually wholesale once the paperwork and compliance costs land, or marketplace once commission, returns and storage are counted properly. Sometimes the answer is “they are within noise of each other and the real question is something else” — that is a legitimate result and you still get the working.
A week, start to finish
Day one — agree the question
Thirty minutes to make it precise enough to answer. Most vague questions have a sharp one hiding inside, and finding it is half the work.
Days two and three — the data
You send exports. We check whether they can actually answer the question, and say so early if they can't rather than at the end.
Days four and five — the work
The analysis, plus the checks that make it trustworthy: does the total reconcile, does the sample match, does the conclusion survive a different cut.
End of the week — the answer
A short written answer with the number, the working, the caveats, and the file so somebody can re-run it next quarter.
Fixed, per question
- Written answer with the working shown
- The data file, so it's repeatable
- Caveats stated, not buried
- If the data can't answer it, we stop and refund the balance
Two or three questions in a row usually means you want reporting instead — the same money spent once answers them repeatedly. We'll say so rather than selling four analyses.
What "with the working" means
Analysis without visible working is just an opinion with charts. Every answer we send includes four things, because without them you can't tell a good answer from a confident one.
The ones that come up
Our data is messy. Is that a problem?
It's the normal starting condition. Messy is workable; ambiguous is the real risk — when there's no way to tell which of two records is the truth. We flag that in the first two days rather than discovering it at the end.
Can you sign an NDA?
Yes, mutual, before anything is sent. And the smaller protection matters more: strip customer names and addresses from the exports. We almost never need them.
What if we don't like the answer?
You still get it. We don't tune conclusions to what the buyer hoped for — an analysis that always agrees with you is worth nothing, and you'd find out the expensive way.
Where to go next
Two lines by email. No form, no calendar link.
Which two systems, roughly what size, what you have already tried. You get a straight answer the same working day — including “this is not something we would help with”, when that is the honest one.
Two lines by email · no obligation · answered by the person who would do the work