E-commerce & Retail

Reconcile 8,400 SKUs across Shopify and Amazon before your standup

NRF says shoppers will return $850 billion of merchandise in 2025 — 19.3% of every online order. Your team is wrangling Shopify exports, Amazon Seller reports, and Klaviyo CSVs in spreadsheets to figure out which SKUs are bleeding. SheetAI lives in those sheets, reads the cells in place, writes the formulas, and explains every move.

Try it on your sheet See real workflows

Audit-ready trailNo data trainingReversible per action

7 min

To reconcile a multi-channel SKU file

Median time to join a 5-tab Shopify + Amazon + Meta workbook into a single profitable-SKU view, on representative DTC books with 6,000–10,000 SKUs.

19.3%

Of online orders are now returned

NRF 2025 Retail Returns Landscape — the online return rate ran 3.5 points above the all-channel average of 15.8%.

$1.73T

Lost annually to inventory distortion

IHL Group 2025 — $1.157T to out-of-stocks plus $572B to overstocks, after $172B of improvements in the prior year.

Online return rates hit 19.3% in 2025 — that's $215 billion of e-commerce merchandise heading back through the warehouse.

Returns are the second checkout your finance team never planned for. The fix is not another dashboard — it is a sheet that can join your Shopify orders, Amazon FBA returns, and ad spend on the same row, and tell you which SKUs are quietly destroying margin.

Source: NRF — 2025 Retail Returns Landscape

Modern DTC and marketplace operators run their business out of CSVs. Shopify exports, Amazon Seller Central reports, Meta ads breakdowns, Klaviyo flow stats, Recharge subscription data — they all land in Excel or Google Sheets, and a human stitches them together every Monday. The 2025 NRF returns benchmark put the cost of that stitching in plain numbers: $850 billion in merchandise back through the door, and a 19.3% return rate online. IHL Group puts inventory distortion at $1.73 trillion globally. SheetAI does not replace your reporting tool. It replaces the part of your week where someone copies a Helium10 export into a tab, runs a VLOOKUP, and prays the columns lined up. Every action it takes is reversible, every formula it writes is standard Excel, and your file never leaves your account.

The state of e-commerce ops in 2026

Four numbers, sourced from 2025 retail and DTC benchmarks, that explain why your Monday report takes nine hours — and where AI is and is not helping yet.

Online return rate (2025)

19.3%

NRF 2025: 19.3% of online sales returned vs. 15.8% all-channel. 9% of those returns are fraudulent. Apparel runs 24%+.

NRF 2025

Global inventory distortion (lost sales)

$1.73T

$1.157T to out-of-stocks (customers walk), $572B to overstocks (deep markdowns). 6.5% of global retail sales.

IHL Group 2025

Meta ads CPM increase, 2025 vs. prior year

+20%

Meta CPM rose 20.03% across 2025; holiday peaks ran +66%. CAC up 40–60% over the last two years across DTC.

Right Side Up / Triple Whale

Retailers using or piloting AI in 2025

~80%

McKinsey: 78% of businesses use AI in at least one function. 86% of retailers have implemented AI somewhere — but most of it never touches the operator's spreadsheet.

McKinsey / Envive

The pattern: returns and stockouts are eating margin, ad costs are climbing 20% a year, and AI adoption has crossed 80% — yet most of that AI lives in a vendor dashboard the operator does not open. SheetAI is built for the inverse: AI that reads the SKU sheet you already work in.

Anatomy of a weekly merchandising review

Where the hours actually go on a weekly DTC merch review, before any automation. We mapped this against the median multi-channel brand running Shopify + Amazon + at least one paid channel. If your Monday looks like this, you are not behind — you are the median.

Phase 1 — The export marathon

›Pull Shopify orders, refunds, and inventory CSVs
›Download Amazon Seller Central business reports and FBA returns
›Export Meta + Google Ads breakdowns and Klaviyo flow stats

⚠ Six tools, eight CSVs, three different SKU naming conventions. The 2025 Shopify benchmark says merchants run six apps on average; Plus stores run 30+ across the stack.

Phase 2 — The SKU join

›Match Shopify variant IDs to Amazon ASINs to internal SKUs
›Pivot order rows up to SKU-day grain
›Reconcile FBA fees, refunds, and storage charges to net revenue

⚠ A VLOOKUP across 8,400 SKUs that breaks the moment a column moves. Half the merch managers we talk to keep a manual mapping tab.

Phase 3 — Returns and margin truth

›Attach 30-day return rates back to each SKU
›Subtract refunds, fees, and storage to get a true contribution margin
›Flag anything below the merchandising target

⚠ 19.3% of online orders come back. The "winning" SKU on gross revenue is often the one funding the warehouse return desk.

Phase 4 — Channel attribution

›Pull ROAS by campaign from Meta, Google, and TikTok
›Tie campaigns to SKU sales using UTMs and last-click
›Build a CAC-by-channel-by-cohort view

⚠ CAC is up 40–60% in two years per LoyaltyLion. Three attribution models disagree. The marketer picks the one that defends the budget.

Phase 5 — The decision pack

›Repricing recommendations across marketplaces
›Inventory reorder list with safety stock
›Narrative for the founder review and the Monday standup

⚠ The pack is a copy of last week with the numbers updated. The narrative is written from scratch every Sunday night.

Cautionary tale

Why "AI inside the sheet" matters more than you think

In October 2018, Sears filed for Chapter 11 — a 132-year-old retailer that once ran the largest mail-order catalog in the world. The post-mortem in Digital Commerce 360 was damning: between 2013 and 2017, Sears closed nearly a third of its stores, and online sales fell from $2.6 billion to $1.3 billion in the same window. The company had the data — every transaction, every SKU, every return — but the systems to make sense of it never got the investment. Operators were stitching together exports in spreadsheets while Amazon was reading their cells in real time.

The lesson: You do not lose the e-commerce game in one quarter. You lose it the way Sears did — by letting the work sit in tabs that nobody trusts, while the competition reads the same data ten times faster. SheetAI does not replace your operator's judgment. It removes the eight hours a week they spend reformatting CSVs so they can spend that time on the merchandise call.

Source: Digital Commerce 360 — How Sears Failed in the E-commerce Era

Who it's for

If your Monday standup is a re-run of last Monday's spreadsheet, this section is for you.

DTC Founders

Need a real-time view of contribution margin by SKU and channel without waiting on the head of e-comm to refresh the deck.

Heads of E-commerce

Drown in Shopify, Amazon, and Klaviyo exports every Monday — and still cannot trust the SKU-level numbers in the founder review.

Merchandising Managers

Reconcile 8,000+ SKUs across marketplaces by hand, with return rates and FBA fees attached to nothing.

Performance Marketers

Reconcile platform-reported ROAS against Shopify revenue across three attribution models and a 7-day window.

Operations Leads

Manage reorder points, safety stock, and warehouse transfers from a workbook that breaks every time a column moves.

Real e-commerce workflows

The exact prompt, the formula it writes, and the result you'd defend in a founder review.

ecommerce_workbook.xlsx — SheetAI

You ask

“In a new sheet called SKU_Audit, list every SKU from Shopify!A:A and Amazon!A:A. Pull units sold last 30 days, gross revenue, refunds, FBA fees, and 30-day return rate. Add a contribution margin column and a flag for any SKU below 20%.”

SheetAI does

✓Builds a unique SKU list from both channels using a SKU mapping tab.
✓Sums units and revenue by SKU using SUMIFS across both channels.
✓Joins refunds and FBA fees back to the SKU row.
✓Calculates contribution margin and flags SKUs below the threshold.

Formula written

=IFERROR((SUMIFS(Shopify!E:E,Shopify!A:A,A2)+SUMIFS(Amazon!E:E,Amazon!A:A,A2)-SUMIFS(Returns!D:D,Returns!A:A,A2)-SUMIFS(Fees!C:C,Fees!A:A,A2))/SUMIFS(Shopify!D:D,Shopify!A:A,A2),0)

Result

8,412 SKUs reconciled across two channels. 247 flagged below 20% margin — sorted by lost-margin dollars so the merchandising call has a top-30 list, not a 10-tab spreadsheet.

Everything DTC operators need, in one chat box

Turn fragmented marketplace exports into one defensible operator view. Reconcile SKUs, audit returns, attribute spend, and propose reorders with AI precision — inside the workbook your team already runs.

✓Multi-channel SKU reconciliation

✓Return-rate and refund analysis

✓Blended ROAS and CAC by cohort

✓Reorder points and safety stock math

✓Repricing across marketplaces

✓Customer LTV by channel and cohort

Plays well with your stack

Shopify exports (orders, products, inventory, refunds)
Amazon Seller Central reports (FBA, ads, business, returns)
Klaviyo CSV exports (customers, flow stats, segments)
Meta Ads breakdowns (campaign, ad-set, ad-level)
Google Ads exports (campaign + keyword + product)
Recharge subscription data (active, churned, MRR)

What a weekly e-commerce review looks like

A representative DTC + Amazon brand running ~8,000 SKUs and 4 paid channels, before and after SheetAI. The "before" mirrors the median Monday for a head of e-commerce; the "after" reflects what our DTC customers report after their second week on the platform.

Before SheetAI

~34 hours

Sun
Pre-pull exports for Monday — Shopify, Amazon, Meta, Klaviyo
~3h
Mon
SKU map repair, VLOOKUPs across 8,400 rows
~9h
Tue
Refunds + FBA fees attached, contribution margin built
~7h
Wed
Reconcile platform ROAS to Shopify revenue, three attribution models
~6h
Thu
Reorder list, safety stock math, supplier email drafts
~5h
Fri
Founder review pack rebuilt from scratch
~4h

✓

With SheetAI

~5 hours

Mon
AI joins 8,412 SKUs, flags 247 below margin, sorts by $-impact
~1h
Tue
Return-reason cohort built; top 10 cost drivers surfaced
~1h
Wed
Blended ROAS and CAC-by-cohort drafted; you edit, not write
~1.5h
Thu
Repricing and reorder packs exported with full lineage
~1h
Fri
Founder review on the call, not a rebuild
—

~85% reduction in weekly merch-review hours, on a representative DTC + Amazon brand.

What SheetAI will not do

An e-commerce tool that is honest about its limits is the only kind worth installing. Some decisions belong to humans, full stop.

Change live prices on Shopify or Amazon

SheetAI proposes a repricing list. A human pushes the button — through your repricer, your Shopify admin, or your Amazon Seller Central session. We do not hold your marketplace credentials and we do not write to live storefronts.

Refund a customer or process a return

Refunds touch payment processors, fraud rules, and your customer-experience policy. SheetAI flags the SKUs and reasons; a human approves the refund through Shopify, Amazon, or Recharge.

Send messages to your customers

No email, SMS, or chatbot replies on your behalf. If you want a Klaviyo flow built, SheetAI can draft the segment and the copy in a sheet — you review and load it into Klaviyo yourself.

Change your ad bids without review

It will draft a "pause / keep / increase" list with the rationale per campaign. The bid change happens in Meta or Google Ads under your operator's login — not via an autonomous agent on your account.

“

Monday used to start at 6am with a stack of exports. Now my head of e-comm walks in at 9, the SKU audit is already drafted, and we spend the standup arguing about the merchandise — not arguing about whether the spreadsheet is right.

— Founder, $40M ARR DTC apparel brand (Shopify Plus + Amazon FBA)

Frequently asked

Things DTC and marketplace operators ask before they switch.

Does SheetAI connect directly to Shopify and Amazon?

›

SheetAI works on the exports you already pull — Shopify CSV/XLSX, Amazon Seller Central reports, Klaviyo extracts, ad-platform breakdowns. We deliberately do not hold your storefront credentials, which is a security plus and the reason brands trust us with margin data. The reads happen on your sheet, in your account.

Can it handle 8,000+ SKUs without choking?

›

Yes. SheetAI reads ranges on demand rather than loading the whole file into context, so it scales to hundreds of thousands of rows. For very large reconciliations, the Python tool runs the join server-side and writes results back to your sheet.

Will the formulas work in Excel and Google Sheets?

›

Every formula SheetAI writes is standard Excel / Google Sheets syntax — SUMIFS, INDEX/MATCH, IFERROR, VLOOKUP, FILTER. There are no proprietary functions to install, and your file remains portable. The Excel add-in and Google Sheets add-on let your operator run the same prompts inside Microsoft 365 and Google Workspace.

How is this different from Triple Whale or Polar?

›

Those are dashboards. SheetAI is the assistant that lives in the sheet your operator works in when the dashboard does not answer the question. If you want a fixed view of blended ROAS, use a dashboard. If you want "join Amazon FBA fees back to my Shopify SKU list and tell me which SKUs lose money after returns," use SheetAI.

Can it tell me which SKUs are unprofitable after returns?

›

Yes — that is one of the canonical workflows above. SheetAI joins your orders, refunds, FBA / Shopify fees, and ad spend on a SKU column and computes contribution margin per SKU. It explains the formula it used so a buyer or merch manager can audit the math.

Does it support multi-marketplace repricing?

›

Yes, as a proposal step. SheetAI compares your current price to a competitor median (e.g. from a Helium10 export), computes a margin-preserving new price, and outputs a list. The actual price change happens in your repricer or marketplace admin — see "What SheetAI will not do" above.

Is my customer data safe?

›

Your spreadsheet stays in SheetAI. The AI receives only the cell ranges relevant to the action you requested — never the entire file by default. We do not train models on customer data, and Pro plans include enterprise-grade encryption at rest and in transit. Klaviyo and order-level data is treated with the same controls as financial data.

We are a 5-person ops team. Does this work for a small brand?

›

Especially. The operators we talk to at $5–50M brands do not have a full-time analyst — the head of e-comm is the analyst at 9pm on Sunday. SheetAI gives one operator the leverage of three by taking the mechanical exports-to-pivot work off the desk.

How does this differ from copy-pasting into ChatGPT?

›

SheetAI works against live cell references — it writes formulas, not screenshots of formulas. Edits are reversible per action, every change is annotated, and the model can iterate on its own output by reading the result back. Pasting an Amazon Seller export into a chat tool gives you advice; SheetAI gives you a finished sheet your buyer can sign off on.

Sources & further reading

Every benchmark and statistic on this page is drawn from publicly available research. We cite our sources because we read theirs.

Related teams using SheetAI

Adjacent functions that hit the same kinds of problems.

Stop wrangling Shopify exports.
Start growing the brand.

Open SheetAI, drop in your Shopify and Amazon CSVs, and watch the SKU audit build itself. Free forever for the first 20 credits a day — no card required to find out whether it works on your data.

Start free All use cases

No credit card required — Free forever tier available