Analytics

The Analytics page in the Pinch dashboard is the daily-glance view of how Pinch is performing on your store. It has three tabs at the top — Orders (reseller abuse), Returns (return abuse, the default for most accounts), and Custom — plus a top-right time range selector with Today, 7D, 30D, 3M, 6M, 1Y, and Go Live (since enforcement started).

The Returns tab

The Returns tab is split into five sections, top to bottom.

Top Metrics

Four headline numbers at the top of the page:

Metric	What it measures
Abuse Impact	Total dollar value of refunds Pinch flagged as abuse over the selected window — the savings story.
Return Abuse Rate	Share of all returns that were flagged as abuse. Most stores run between 2% and 8%.
Avg Return Cost / Item	Your fully-loaded cost per returned item, including forward shipping, return shipping, quality inspection, and any restocking discount. Editable inline.
Return Rate	Items returned divided by items ordered, across the selected window.

Hovering the info icon on each metric shows the exact formula Pinch uses.

Key Insights

Auto-generated alerts that compare the last 90 days against expected baselines from your historical data. Each insight has a severity badge — High, Medium, or Low — based on how far the actual value is from expected.

Common insight types:

• VIP Customer Count Increase / Decrease — significant shifts in your highest-LTV segment
• Return Rate Increase / Decrease — overall return-rate trend against baseline
• Abuse Rate Increase / Decrease — abuse rate against baseline
• Egregious Abusers Count Increase / Decrease — number of customers flagged as the highest-risk segment
• Suspected Abusers Count Increase / Decrease — middle-tier risk segment

A High severity insight is worth a 30-second look every morning. Low severity insights are typically just noise in normal weekly fluctuation.

Customer Segmentation

The four-segment breakdown of your full customer base — VIP, Others, Suspect, Egregious — with the share of customers in each segment, the share of orders, the return rate, and the restock rate per segment. The total customer count is shown above the cards.

Below the segment cards, three tabs — Egregious, Suspected, VIP — show the top users in each segment with their order count, return rate, and restock rate, sortable by any column.

Geographic Distribution

A list of states (with an All / US toggle at the top) showing case counts and the percentage change vs. the previous comparable period. A big negative number (e.g., -83%) on Texas means abuse cases dropped by 83% there compared to the prior period. Use this to spot regional shifts that warrant a manual look.

Product Categories

Two halves of the same panel:

• New product detected — when a SKU first enters the return stream, Pinch flags it with a NEW badge so you can watch it during its initial window before the model has enough data to score it confidently.
• Biggest return-rate changes — products sorted by the largest percentage change between the previous period's return rate and the current period's return rate. A 19% return rate that jumped from 1% (+1900%) is usually worth investigating with merchandising — could be a fit issue, photography issue, or a TikTok-driven impulse-buy wave.

Transaction Trends

A time-series chart showing Orders, Returns, and Damage (damaged-on-arrival items) on a log scale across the selected window. A callout at the top of the chart flags the peak transaction day. Patterns to watch for:

• Orders and Returns moving in opposite directions (returns climbing while orders flatten = trouble)
• A sudden spike in Damage relative to Orders (could be a shipping carrier issue)
• A return wave 7–14 days after a promo or holiday peak — normal but worth modeling against

The Orders tab

For merchants with reseller abuse (limited-release SKUs, bot-driven order placement, bulk-buyer circumvention of purchase limits), the Orders tab is where you watch the order-side picture.

Bot Trend

A daily bar chart comparing Total Orders against Bot Orders — orders that Pinch's signals indicate came from automated agents. A sudden spike in bot orders (especially around a drop or restock) is the leading indicator that resellers have found your launch.

Top Resellers

A ranked table of the customers Pinch has identified as most likely to be resellers. Columns:

• # (rank)
• Name
• Identifier value (e.g. Email, Phone, Address — chosen via the Category dropdown in the top-right)

Use the Category dropdown to flip between identifier types — sometimes the email reveals the reseller pattern, sometimes it's the shipping address (many emails shipping to the same forwarder) or the phone number.

Transaction Trend ($)

A multi-line chart showing transaction value over time, broken into three series:

• Transactions — total order value
• Returns — total refund value
• Returns Abused — refund value on returns Pinch flagged as abusive

The relationship between these three lines is the high-signal view of your overall financial exposure to abuse.

Decision Outcomes Trend

A daily stacked bar chart of every decision Pinch produced on incoming orders — Allow (green), Review (yellow), Deny (red). Use this to see how aggressively your workflow is firing day-over-day and to spot policy drift.

The Custom tab

The Custom tab is where you pin widgets you've built yourself via the AI Widget Builder (Widget Studio).

A brand-new account sees: "No Widgets Added Yet. Use the AI Widget Builder to create and pin widgets here."

Once you've created and saved widgets, they appear here as a customizable grid you can resize and rearrange.

AI Widget Builder (Widget Studio)

Click AI Widget Builder in the top right of either Analytics tab to open Widget Studio. This is a natural-language interface for building custom analytics — no SQL.

You type a question in plain English and Pinch generates a chart, table, or KPI tile. Example prompts:

• "How many orders did I get last month?"
• "Analyze return trends over the past 6 months"
• "Top 10 customers by spend this quarter"
• "Show returns by reason code"

Widget Studio has two side panels:

• Chat History — every prior question you've asked, with timestamps and message counts. Click any prior chat to resume.
• Templates — save a chat as a reusable template so you can re-run the same query later with one click. Useful for recurring weekly reports.

Once a widget produces output you like, pin it to the Custom tab to see it on every Analytics visit.

Time range and the "Go Live" selector

Every panel that supports time ranges has the same selector in the top right: Today, 7D, 30D, 3M, 6M, 1Y, and Go Live.

Go Live is a Pinch-specific range: it starts from the day Pinch began enforcing decisions on your account (not just scoring in shadow mode). Use it when you want to measure cumulative impact since launch — most useful for board reports and quarterly reviews.

Exporting data

For the customer segmentation tables and any product or geographic breakdown, hover the panel and click the export icon (top-right of each panel). CSV download is available for most lists. For bigger pulls — e.g., the full customer list with all variables — use the Download button on the Customers page.

What to do with this page day-to-day

Most operators check Analytics once a day, in this order:

1. Glance at Key Insights for any High severity alerts.
2. Look at the Top Metrics for the last 7 days — did Abuse Impact climb or drop noticeably?
3. If anything looks off, drill into Customer Segmentation to see whether one segment is shifting.
4. Once a week, scan Product Categories for new return-rate climbs and forward to merchandising.

For anything ad-hoc, the AI Widget Builder is faster than a SQL query.

See also

• Review queue — manual review of individual flagged returns
• Customers — drill into specific customers from the segmentation tables
• Disposition data — the warehouse signal that powers most of the metrics here