Evaluate an ecommerce accounting vendor on seven things: what level of detail it posts to your ledger, how it handles cost of goods sold, how far back it can pull history, which channels and accounting systems it actually supports, what the price does as you grow, what happens to your data if you leave, and what it explicitly cannot do. That last one is the question that separates a useful demo from a sales call, and almost nobody asks it.
Here is how to run the evaluation, with the specific questions that produce useful answers.
1. Summary or detail: what lands in the ledger
This is the fundamental architectural split in the category, and vendors on both sides will describe their approach as the correct one.
Summary tools post one journal entry per settlement period, breaking the payout into revenue, fee, and tax accounts. Your ledger stays small and fast, and reconciliation to the bank deposit is clean. Detail tools push data down to the individual SKU, so inventory and per-product cost live inside the accounting system itself.
Summary is usually right if your accountant drives the relationship and you run analytics elsewhere. Detail is usually right if you need QuickBooks itself to hold inventory quantities and per-item cost. A ledger with per-SKU detail on a large catalog also gets heavy, which is a real operational cost rather than a talking point.
Ask: show me the exact journal entry this produces for one settlement period, on a catalog my size.
2. Cost of goods sold mechanics
Ask three questions and do not accept a yes to the first alone. Does it post COGS automatically? Which valuation method does it use, and can you change it? What happens to already-posted entries when you correct a historical unit cost?
That third question matters most and gets the vaguest answers. Some systems restate history, which is clean but can move a closed period. Others adjust forward only, which preserves closed books but leaves the old number wrong. Both are defensible. Not knowing which one your vendor does is not. The IRS discusses inventory valuation and accounting method consistency in Publication 538, and changing methods is not a casual toggle.
Also confirm whether COGS is included in the tier you are pricing. A2X, for instance, lists Amazon plans from $29 per month on its site as of September 2026, but cost of goods sold and FBA inventory locations are excluded from that entry tier and begin at the $59 plan. That is a fair and clearly disclosed structure. It is also exactly the kind of detail that makes a headline price misleading if you stop reading at the number.
3. How far back it can reach
History limits determine whether you can rebuild a prior year or only go forward.
Limits vary meaningfully. A2X offers maximum available history on its premium tiers, with Amazon Pay data capped at 24 months. ConnectBooks states it can download transactions up to 18 months back, with the first six weeks before signup included free and historical data beyond that quoted separately. Those are different propositions, and if you are switching mid-year or cleaning up a messy prior period, this single constraint may decide the vendor for you.
Ask: what is the maximum history, is it the same on my tier, and what does backfill cost?
4. Channel and ledger coverage, checked against your actual stack
Every vendor lists integrations. Check the intersection with what you run, and what you plan to run in eighteen months.
Coverage genuinely differs. A2X supports Amazon, Shopify, eBay, Etsy, Walmart, and PayPal, and posts into QuickBooks, Xero, and NetSuite. ConnectBooks covers Amazon, Shopify, Walmart, TikTok Shop, and eBay into QuickBooks Online, QuickBooks Desktop Enterprise, and Xero. If you sell on Etsy, take PayPal volume, or run NetSuite, A2X covers ground ConnectBooks does not. If TikTok Shop is a growth channel or you are on QuickBooks Desktop Enterprise, the comparison runs the other way.
Neither is the better product in the abstract. They are built for overlapping but distinct stacks, and the only version of this question that matters is the one scoped to yours.
5. What the price does as you grow
Most vendors in this category price on monthly order volume, which means your bill rises with the thing you are trying to increase. Model it at two and three times current volume before signing.
Look for whether the price steps or scales smoothly, whether it falls again in a slow month, and whether marketplace count is a separate multiplier. ConnectBooks publishes pricing tiers that adjust with monthly order volume, with Gold listed from $149 and Diamond and Platinum from $199 per month as of September 2026, and states that billing moves with actual order count in both directions. A2X’s published Amazon ladder runs from $29 at 200 orders to $1,499 at 250,000, with custom pricing above.
Ask for the number at your projected volume in writing, not the starting price.
6. Exit terms
Ask before you sign, because the answer after you sign is whatever the vendor decides.
Can you export your full historical data, in what format, and does the export include the mappings and cost history or only transactions? What happens to already-posted journal entries in QuickBooks if you cancel? Is there a contract term, or is it month to month?
Entries already written to your ledger normally stay, since they are yours. The configuration that produced them usually does not travel, which means switching vendors involves rebuilding mappings and re-entering cost history. Budget for that rather than discovering it.
7. What it cannot do
Ask the question directly: what do customers most often ask for that this product does not do?
A vendor that answers honestly is telling you something useful about the next two years. ConnectBooks, for example, states on its own site that it tracks stock at the warehouse level but not by bin or zone, that purchase orders can be downloaded as PDFs but not emailed to suppliers from inside the platform, and that it does not currently offer an open API for external use. None of those are defects. They are scope decisions, and they are only problems if they collide with how you operate.
Published limitations are a credibility signal. A vendor whose answer to this question is that nothing is missing has either not been asked often enough or is not answering.
Run a real trial
Demos use clean data. Connect a trial to your actual account, let one full settlement period post, and reconcile it by hand against the marketplace report and the bank deposit. If those three agree, the tool works on your data. If they do not, you have learned the most important thing available before signing anything.
Bring your accountant into the trial rather than the demo. They will ask about the chart of accounts mapping, which is where most of these implementations actually succeed or fail. If you do not have one who knows ecommerce, the AICPA and marketplace advisor directories are reasonable places to start.
