Skip to main content
When sales move, a monthly recap tells you what changed. A sales diagnosis asks a narrower question: what patterns in the Hoard data might help explain the change? It is a read-only investigation of your own recorded sales. It does not change listings, pricing rules, or sync settings. First connect your assistant. For gross sales, order counts, fees, refunds, and net sales without a diagnosis, use summarize your TCGplayer sales.

​
Step 1: Ask about a specific month

Name a month when you can, especially if the apparent change is not the most recent one:
“Why were my TCGplayer sales down in [YYYY-MM]? Use my sales diagnostic. Compare the month with the available month-over-month, year-over-year, and trailing baseline. Rank the hypotheses by evidence, show the evidence and confidence for each one, and include every caveat and coverage warning. Read only; do not change anything.”
For the most recent covered month, ask the same question without the month. The assistant uses Hoard’s analytics.salesDiagnostic(month?) read, which performs the comparison on the server instead of trying to infer a reason from raw monthly totals.

​
Step 2: Read the answer as hypotheses, not proof

Expect the selected month’s recorded orders, gross sales, net sales after fees, refunds, and comparisons when enough history exists. Then read the ranked hypotheses with their evidence and confidence. The diagnostic can surface patterns including:
  • A fee-rate difference from the trailing average.
  • An expired sale event that may have made a recent comparison period unusually strong.
  • A seasonal pattern: down versus last month but steady or better versus the same month last year.
  • A refund-rate spike.
  • A large repricing run during the month.
  • No meaningful change, no specific Hoard-visible signal, or insufficient history.
These are leads to investigate, not a single proven cause. A large repricing run gives you context; it does not prove repricing changed sales. An expired sale can make the earlier month a tougher comparison, but it cannot explain every difference by itself. Read refund-rate evidence carefully. Refunds are counted by refund date, while orders are counted by order date. A refund this month may belong to an older order. If the two windows do not line up, the diagnostic flags the mismatch and lowers its confidence; do not repeat the displayed rate as a clean percentage of this month’s orders.

​
Step 3: Check the coverage before drawing a conclusion

Ask the assistant to report the target month, number of history months available, any plan-history warning, and whether a month-over-month, year-over-year, or trailing comparison was unavailable. A missing sales-report month is a coverage gap, not proof of zero sales. Short history can prevent a year-over-year comparison, and your plan’s history window can limit the baseline. If the requested month has no recorded report, the diagnostic may return no result for that month even when older months exist. Check the covered-month list before describing the whole account as empty. If the newest data looks incomplete, check sync health. That confirms what Hoard last observed; it does not fill a missing month.

​
What the diagnosis cannot tell you

Hoard does not see TCGplayer traffic, search placement, Buy Box share, competitor undercutting, or category-wide demand. Do not treat a sales diagnosis as evidence for any of those explanations. Similarly, a fee-rate change in your recorded sales is not proof that TCGplayer changed its policy. It describes the effect in your store data. For the separate policy question, see asking about TCGplayer fee changes.

​
Keep investigating in the right order

  1. Start with the diagnostic and its caveats.
  2. Use the sales summary if you need the underlying monthly totals or a longer recap.
  3. Check sync health when recent coverage is missing or stale.
  4. Treat a high-confidence hypothesis as a focused follow-up to investigate, not permission to change pricing automatically.