# Sales volume increased; the notes do not establish a conversion improvement

The operating report shows a flat 2.5% orders-to-sessions ratio despite 20% more orders. The support article gives a useful hypothesis about policy clarity, but its four selected conversations cannot establish a causal conversion improvement. Clarify confusing policy information, then measure the outcome with a defined funnel or experiment. Both sources are fictional research materials supplied for this case.

## Source A: HarborMug weekly operating report

Author: Mia Lane. Organization: HarborMug. Document date: November 3, 2026. Type: internal technical report. Question: what changed between the two trading weeks?

Method and sample: two aggregate weekly rows, 100 orders/4,000 sessions then 120/4,800; no customer-level tracking or controlled experiment (A §2). Gross merchandise revenue increased from $2,400 to $2,880, +20%; refunds increased from $120 to $144; ad spend increased from $300 to $360. Gross excludes shipping revenue and tax.

Recalculation: 100/4,000 and 120/4,800 both equal 2.5%. Gross AOV remains $24. Net revenue after the supplied refund dollars is $2,280 then $2,736, +20%. Refund dollars/gross revenue remains 5%; this is not the percentage of orders returned. No realized profit can be established without the remaining costs and revenue definitions (A §§2–4).

Exact quotation: “Volume grew; conversion did not improve.” (A §3.) Read “conversion” here as the report's aggregate order/session ratio, not a deduplicated shopper purchase probability.

## Source B: HarborMug support observations

Author: Alex Reed. Date: November 2, 2026. Type: supplied article text; no public URL. Question: could clearer delivery and return terms help shoppers?

Method and sample: four selected support conversations mentioning timing uncertainty, with no control group and no measured before/after conversion (B §2). The article's belief that conversion improved is unsupported by its own evidence and conflicts with Source A's aggregate ratio (B §1 versus A §3).

The policy facts are operationally useful: two-business-day dispatch, another three-to-five-business-day delivery estimate, $5 shipping below $50, free shipping at $50, US-only delivery and unused returns within 30 days requiring merchant refund approval (B §3). They do not prove that clarity caused any sales change.

Exact quotation: “Four conversations are a prompt to investigate, not a conversion study.” (B §2.) Convenience selection may overrepresent dissatisfied shoppers. Funding is not stated, the author is merchant-affiliated and opposing explanations were not tested (B §4).

## Comparison and evidence quality

| Dimension | Source A | Source B |
|---|---|---|
| Main claim | Volume grew; order/session ratio flat | Policy clarity may help; improvement is only a belief |
| Method | Two weekly aggregate rows | Four selected support conversations |
| Numeric scope | 220 orders, 8,800 sessions across both weeks | Four conversations, no population denominator |
| Strength | Arithmetic can be reproduced | Identifies a concrete customer question to investigate |
| Credibility | Medium: internal, not peer-reviewed | Low: supplied opinion/article, no independent review |
| Evidence | Medium for descriptive arithmetic; low for causality | Low for conversion effects |
| Recency | High within the fictional exercise dates | High within the fictional exercise dates |
| Objectivity | Medium: merchant-funded, limitations disclosed | Low: merchant-affiliated convenience sample |

Neither source is independent market evidence. The original extractor labels both citations primary; the operating table is primary descriptive data within the fixture, while the support article mixes primary observations with interpretation. Its interpretation should not acquire stronger status merely because a regex classified the citation.

## Supported decisions and next research

First, keep dispatch and transit wording separate and show shipping thresholds near the buying decision. This follows the specific uncertainty in B §2 and the policy constraints in B §3. Second, retain the unchanged conversion finding in the weekly review rather than crediting the page for volume growth. Third, instrument unique shopper exposure and paid purchases, then define a comparison window or randomized test for policy presentation.

Before testing, inspect whether acquisition mix, stock or promotions changed; neither source covers those alternatives. Collect a consecutively sampled set of support questions, preserving anonymous record IDs and the denominator. Ask shoppers to explain delivery cost and timing without coaching. Use these observations to refine the intervention, not to manufacture a percentage uplift.

## Bibliography and traceability

1. Lane, M. (2026, November 3). *HarborMug weekly operating report* [Synthetic internal report]. HarborMug. Source a.txt, §§1–5. The extractor's supplied reference lists only the year; the full date above comes from the document header.
2. Reed, A. (2026, November 2). *HarborMug support observations* [Synthetic supplied article]. HarborMug. Source b.txt, §§1–4. Public URL: not supplied.

brief.txt concatenates A and B and is not a third independent study. The original extraction tool found two citations, matching these two bibliography entries. Both quotations were checked verbatim against their source sections; the calculation file independently reproduces the numerical findings. No external research discovery or genuine merchant outcome is claimed.