Given clean, reconciled ecommerce data, current analysis models are competent. They will compute a margin bridge, rank SKUs by contribution, and describe a cash cycle without difficulty. What they will not do is decide what is worth asking, and most of the disappointment in this category comes from vague questions producing vague answers that get mistaken for insight.
"How is the business doing" returns a summary you already knew. "Which SKUs on which channels lost money after fees, returns, and advertising in the last 90 days, and how much did each lose" returns a list you can act on this week. Same tool, same data, different value.
The gap is expensive. CFO.com reported in July 2026 on Intuit midmarket CFO research finding that 57 percent of finance leaders said delayed data had caused them to miss opportunities. Speed only helps if the question was worth answering.
The default profitability view stops at gross margin. For a marketplace seller, gross margin is a long way from the truth. Referral fees, fulfillment, storage, returns processing, refunds, promotional rebates, and advertising all sit between gross margin and reality.
Ask for contribution after all of it, at SKU and channel level, ranked from worst. Every seller who runs this for the first time finds at least one SKU they were promoting into a loss.
A margin percentage that moved is a headline, not an answer. The useful version decomposes the move into price, mix, landed cost, fee changes, promotional depth, and returns.
Payment terms with suppliers, time in transit, days of inventory on hand, and marketplace payout timing differ enough by channel that a blended cash conversion cycle hides the thing you need to know. A channel that pays out every two weeks and one that holds a rolling reserve are different financing situations for the same product.
Ask for units by age band, by location, with the storage fees attributable to each band. Marketplace long-term storage assessments are the visible cost. The invisible one is capital sitting in goods that will eventually be marked down.
Fee structures differ. Fulfillment differs. Return rates differ, sometimes dramatically. The unit that contributes well on one marketplace can contribute nothing on another, and the answer changes the allocation of both inventory and advertising. eBay's fee and payment structure produces a different contribution profile on identical goods than an FBA channel does.
Freight per unit up 30 percent. Referral rate up one point. Return rate up two points. Ask which SKUs cross from positive to negative contribution under each, and how many units of current stock that represents. This is the question that turns analysis into a decision.
Gross margin fell from 41.8 percent in the first quarter to 37.7 percent in the second, on revenue that rose from $2,840,000 to $3,105,000. Revenue up, margin down. The summary answer, "margin declined 4.1 points on higher volume," is true and useless.
Decomposed, the 4.1 points break out like this:
Four separate stories, four separate owners, four different responses. The freight number is a supplier and logistics conversation. The mix number may be entirely acceptable if the volume growth was profitable in dollars. The promotional depth was a choice someone made and can unmake. The fee change is a repricing question. The returns improvement should be replicated across other listings.
None of that is visible in "margin fell 4.1 points," and all of it is computable from data you already have if the ledger carries fee types, landed cost, and channel identity. That is what SKU and channel profit reporting is for, and it is what makes the decomposition a report rather than a two-week project.
"Am I doing well?" No definition, no comparison basis, no action attached. You will get a paragraph of hedged commentary.
"What should I do?" The model does not know your capital position, your supplier relationships, your appetite for risk, or the wholesale account you are about to sign. Ask it for options and constraints, and make the call yourself.
"Predict next quarter's revenue." Extrapolation from your own history is worth having as one input. Presented as a forecast without a stated method and a range, it is a number with false precision.
Anything requiring data you have not connected. If your competitors' pricing is not in your system, questions about competitive position will get answered from general knowledge rather than from your business. That answer will sound specific and will not be.
Four habits separate useful queries from decorative ones.
Specify the period and the basis. "Last 90 days, on shipped orders, excluding sales tax collected." Ambiguity about the basis is the single most common source of numbers that do not tie.
Specify the cost stack. Name what should be deducted: COGS at landed cost, referral, fulfillment, storage, returns, refunds, advertising. Otherwise you are trusting an unstated default.
Ask for the arithmetic, not just the conclusion. A margin bridge should show its components summing to the total change. If they do not sum, something was estimated.
Ask what is missing. A good analysis tool will tell you which inputs were absent and what it assumed. If it never says this, it is filling gaps silently, and you will find out which gaps at the worst moment.
Every question above is answerable only if four things are true in the ledger: settlements decomposed into transaction and fee types, inventory tracked by location at landed cost with a consistent costing method, channel and SKU identity carried through to posting, and consistent period cutoff rules.
ConnectBooks exists to make those four true across Amazon, Shopify, Walmart, eBay, and TikTok Shop, syncing into QuickBooks or Xero. The analysis is downstream. Sellers who try to skip to the analysis get answers computed on a revenue figure that equals their bank deposits, which understates sales by every fee the marketplace withheld.
ConnectBooks has announced Crunch, an AI CFO designed to answer questions against a seller's own reconciled settlement, inventory, and channel data. It is not open to customers yet. The Crunch page describes what it will do and has a waitlist form for sellers who want to be notified when it opens up.
In the meantime, the six questions above are worth running manually, or with your accountant, on last quarter. Most of them take an afternoon with a properly structured ledger and are impossible without one. That is the real test of whether your books are ready for anything in this category: not whether the tool is good, but whether your data can answer a specific question about a specific SKU on a specific channel without someone rebuilding it in a spreadsheet first.
Running an e-commerce business comes with plenty of challenges, but ConnectBooks is here to make your life easier. With real-time insights, seamless integrations, and detailed tracking of your profitability and inventory, you can stay ahead of the game. Whether you’re selling on Amazon, Shopify, Walmart, TikTok or eBay, ConnectBooks helps you manage your finances with 100% accuracy and confidence, so you can focus on growing your business.
Ready to level up? Start making smarter, data-driven decisions every step of the way. Try ConnectBooks Free Today or Schedule a Demo
