[Steve Side Hustler]

Quick question: Can AI look at my Amazon and Shopify sales and tell me which SKUs are likely to be the best sellers — and which ones I should promote or stop stocking?

I run a small online shop and I’m not technical. I have sales history in spreadsheets and product listings on both platforms. I’m curious what’s realistic to expect from AI, and what I would need to provide so it can give useful recommendations.

What I’m hoping to learn:

• Which types of data matter most (sales history, returns, traffic, price, reviews, etc.)
• How accurate or reliable AI suggestions tend to be
• Beginner-friendly tools or services that can connect to Amazon/Shopify
• Privacy, cost, and how to get started without being technical

If you’ve tried this or can recommend simple apps, workflows, or things to watch out for, please share your experience or next steps. Thanks — I appreciate practical tips and real examples!

[Jeff Bullas]

Quick yes — AI can analyze Amazon and Shopify sales to recommend best-selling SKUs — but the value comes from clean data and a clear question.

Here’s a practical, non-technical path to get actionable SKU recommendations in days, not months.

What you’ll need

• Exports of sales data from Amazon and Shopify (CSV or spreadsheet).
• Basic spreadsheet (Google Sheets or Excel) or a no-code connector (Zapier/Make) if you want automation.
• A simple AI tool that accepts pasted tables or CSVs (chat-based model or an AI spreadsheet add-on).

Step-by-step

1. Export sales for the same date range from both platforms (include SKU, date, units, revenue, refunds, shipping, promo, ad spend if available).
2. Consolidate into one sheet. Make one row per SKU per day or month. Standardize SKU codes (watch for variants: SKU-RED vs SKU_RED).
3. Enrich with margin or cost-per-unit and returns if you can. Add a column for advertising spend per SKU (estimate if needed).
4. Calculate key metrics: units sold, revenue, gross margin, return rate, ad cost per unit, velocity (units/day or units/month).
5. Ask AI to rank SKUs by a combination of velocity, margin, and low return-rate. Use the prompt below.
6. Validate by sampling top recommendations against recent inventory and supplier lead times — run a small test order or promo.

Example columns you should have

• SKU, Date, Channel (Amazon/Shopify), Units, Revenue, CostPerUnit, Returns, AdSpend

Copy-paste AI prompt (use with your AI tool)

“You are an expert retail analyst. I will paste a table with columns: SKU, Channel, Date, Units, Revenue, CostPerUnit, Returns, AdSpend. Please:
1) Consolidate by SKU and give total Units, TotalRevenue, GrossMargin, ReturnRate, AdCostPerUnit, Velocity (units/month).
2) Rank top 10 SKUs to prioritize for restock and growth, explaining why (use metrics).
3) Suggest 3 quick actions to increase sales profitably for the top SKU.
If you need missing values, tell me what to estimate.”

Common mistakes & fixes

• Mixing SKU formats — fix by standardizing codes before analysis.
• Ignoring returns — include returns in net units and margin calculations.
• Using different date ranges — align windows (last 90 days, 12 months rolling).

7-day action plan

1. Day 1: Export and combine data.
2. Day 2: Clean SKUs and add cost fields.
3. Day 3: Calculate metrics in the sheet.
4. Day 4: Run AI prompt and get ranked SKUs.
5. Day 5–7: Test top recommendation with a small ad spend or restock.

Start with clean data, ask a focused question, and run a small test. That’s how you turn AI insight into real sales wins.

[aaron]

Quick win (under 5 minutes): open your Shopify or Amazon report for the last 30 days, sort by Units sold and note the top 5 SKUs — that’s your immediate shortlist for testing.

The problem

Most teams assume AI is the magic — it isn’t. AI gives valuable SKU recommendations only when data is clean, consistent, and the decision rules (what makes a SKU “best”) are explicit.

Why this matters

Picking the wrong SKUs to restock or push with ads ties up cash, increases storage costs, and hurts margins. Get SKU prioritization right and you free working capital and boost profitable revenue.

My lesson from the field

I’ve seen companies increase profitable inventory turns by 20–40% within 60 days by standardizing SKUs, adding basic cost fields, and using AI to rank by velocity + margin + returns. The trick: make the AI’s objective match your business outcome.

Concrete steps — what you’ll need and how to do it

1. Gather: export CSVs from Amazon and Shopify for the same date range. Columns: SKU, Date, Channel, Units, Revenue, Refunds/Returns, AdSpend (if available).
2. Consolidate & clean: paste into one sheet. Standardize SKU formatting (uppercase, remove spaces/special chars).
3. Enrich: add CostPerUnit and any shipping or fulfillment fees. If you don’t know exact cost, estimate conservatively.
4. Calculate: for each SKU compute Units, NetRevenue (Revenue – Refunds), GrossMargin = NetRevenue – (Units * CostPerUnit), ReturnRate = Returns/Units, AdCostPerUnit = AdSpend/Units, Velocity = Units per 30 days.
5. Run AI: feed the consolidated table to an AI with the prompt below. Ask for a ranked list and rationale.
6. Validate: pick top 3 SKUs, check inventory lead times and run a small marketing test (ad spend or promo) for 7–14 days.

What to expect from AI

A ranked SKU list with the metrics used, suggested weighting, and 3 short actions for the top SKU (e.g., restock, increase ad bid, bundle). Expect to iterate once after you validate with real-world results.

Copy-paste AI prompt

“You are an expert retail analyst. I will paste a table with columns: SKU, Channel, Date, Units, Revenue, Returns, AdSpend, CostPerUnit. Please: 1) Consolidate by SKU and output TotalUnits, NetRevenue, GrossMargin, ReturnRate, AdCostPerUnit, Velocity (units/30 days). 2) Rank top 10 SKUs to prioritize for restock and profitable growth, using a weighted score: 50% Velocity, 30% GrossMargin, 20% (1 – ReturnRate). Show calculations. 3) For the top SKU, give 3 actionable tests to increase profitable sales and the expected KPI impact (units, margin, CAC). If required fields are missing, list what to estimate and conservative defaults.”

Metrics to track

• SKU Velocity (units / 30 days)
• Gross Margin per SKU
• Return Rate
• Ad Cost per Unit
• Inventory Days of Cover / Lead Time

Common mistakes & fixes

• Mixed SKU formats — fix by normalizing (uppercase, strip punctuation).
• Ignoring returns — always use NetRevenue and net units.
• Using different date windows — compare consistent rolling windows (90 days / 12 months).

7-day action plan

1. Day 1: Export and consolidate data.
2. Day 2: Clean SKUs and add cost estimates.
3. Day 3: Calculate metrics in the sheet.
4. Day 4: Run the AI prompt and get ranked SKUs.
5. Day 5: Verify supplier lead times and stock levels for top 3 SKUs.
6. Day 6–7: Run small ad/promo tests on top SKU and measure CPA, units, and margin.

Make the AI’s objective your KPI: velocity + margin + low returns. Clean data, focused prompt, fast test.

Your move.

[Becky Budgeter]

Nice practical tip — sorting the last 30 days by units sold gives a fast, usable shortlist. That’s exactly the right move when you need a quick test without overthinking it.

Here’s a short, practical add-on you can use right away to turn that shortlist into profitable action. Follow these steps: what you’ll need, how to do it, and what to expect.

What you’ll need

• CSV or report exports from Amazon and Shopify for the same date window (start with 30 days for a quick check, 90 days for more stability).
• A spreadsheet (Google Sheets or Excel) and basic fields: SKU, Date, Channel, Units, Revenue, Returns/Refunds, AdSpend (if available), CostPerUnit (estimate if unknown).
• A simple AI tool or spreadsheet formulas to help summarize and rank SKUs.

Step-by-step: how to do it

1. Export matching date ranges from both platforms and combine into one sheet. Keep columns consistent.
2. Clean SKUs by normalizing format (uppercase, remove stray spaces or symbols) so the same product isn’t split into duplicates.
3. Add cost—enter CostPerUnit. If you don’t know exact cost, use a conservative estimate (better to underestimate margin at first).
4. Calculate key metrics per SKU: Total Units, Net Revenue (Revenue – Refunds), Gross Margin = NetRevenue – (Units*CostPerUnit), Return Rate = Returns/Units, Ad Cost per Unit = AdSpend/Units (if available), Velocity = Units per 30 days.
5. Rank SKUs by a simple score you choose (example: 50% Velocity, 30% Gross Margin, 20% (1 – Return Rate)). You can do this with sheet formulas or ask your AI to consolidate and apply the weighting—don’t expect perfection, just a prioritized shortlist.
6. Validate the top 3: check inventory on hand, supplier lead times, and run small, measurable tests (a modest ad push or a short promo) for 7–14 days to confirm demand at profitable margins.

What to expect

You’ll get a ranked list with clear reasons (velocity, margin, returns), a realistic short-list for testing, and quick actions to try (restock, raise ads for profitable SKUs, or bundle slow movers). Expect to iterate once after real-world tests — the data will help you refine cost estimates and weights.

Tip: start with the 30-day quick check to choose a test SKU, then expand to 90 days if seasonal spikes are likely. Would you like a simple example weighting to try in your spreadsheet?

[aaron]

Strong start on the 30‑day sort — it’s the right way to get a shortlist fast. Now turn that shortlist into profit by ranking SKUs on contribution margin and stock position, not just units.

The gap

Units and revenue can mislead. Amazon/Shopify fees, ad spend, returns, and stockouts flip “bestsellers” into cash traps. The job isn’t to find what sells — it’s to find what sells profitably and can be supplied without stockouts.

Why this matters

Every restock and ad dollar should move toward SKUs with high contribution margin, stable velocity, and enough coverage to avoid going dark. That’s how you grow revenue and free working capital at the same time.

Field lesson

Teams that switch to fee-adjusted contribution margin and lead-time-aware velocity reliably improve ad efficiency and inventory turns. The insider trick: exclude SKUs that will stock out within lead time from growth pushes — treat them as “restock priority,” not “scale priority.”

Do / Do not

• Do include platform fees, fulfillment, discounts, and returns in margin.
• Do normalize SKUs across variants and channels before ranking.
• Do compute velocity on 30 and 90 days to smooth spikes.
• Do factor inventory on hand and supplier lead time (days of cover).
• Do use a weighted score anchored to your KPI (profit, not revenue).
• Don’t promote SKUs that will stock out before lead time — you’ll pay for momentum you can’t fulfill.
• Don’t mix date ranges or average percentages; use weighted sums.
• Don’t ignore refunds — use net units and net revenue.

What you’ll need

• Exports from Amazon and Shopify for the same window (30 and 90 days)
• Columns: SKU, Date, Channel, Units, Revenue, Discounts, Refunds/Returns, AdSpend, CostPerUnit, PlatformFees (FBA/referral or payment/fulfillment), ShippingCost, InventoryOnHand, LeadTimeDays (estimate if needed)
• A spreadsheet and a chat-based AI that accepts pasted tables

How to do it

1. Consolidate: Standardize SKU case and formatting; one row per SKU per day (or month for simplicity).
2. Calculate per SKU (by channel first, then blended): NetRevenue = Revenue – Discounts – Refunds; TotalCOGS = Units * CostPerUnit; TotalFees = PlatformFees + ShippingCost + AdSpend; ContributionMargin = NetRevenue – TotalCOGS – TotalFees; Velocity30 and Velocity90 (units per 30/90 days); ReturnRate = Returns/Units; AdCostPerUnit = AdSpend/Units.
3. Inventory lens: DaysOfCover = InventoryOnHand / (Velocity30/30). StockoutRisk = DaysOfCover < LeadTimeDays.
4. Score: For SKUs without StockoutRisk, Score = 50% Velocity30 (normalized) + 30% ContributionMargin (normalized) + 20% (1 – ReturnRate). Flag StockoutRisk SKUs as “Restock priority.”
5. Decide: Top 10 by Score = growth candidates; Top 10 by (Velocity90 x Margin) with StockoutRisk = restock priority list.
6. Act: Increase bids/promo only on growth candidates; place POs for restock priorities first.

Copy‑paste AI prompt

“You are a senior retail analyst. I will paste a table with columns: Date, Channel, SKU, Units, Revenue, Discounts, Refunds, CostPerUnit, PlatformFees, ShippingCost, AdSpend, InventoryOnHand, LeadTimeDays. Tasks: 1) Aggregate by SKU across channels and compute: NetRevenue, TotalCOGS, TotalFees, ContributionMargin (NetRevenue – TotalCOGS – TotalFees), Velocity30 (units/30d), Velocity90, ReturnRate, AdCostPerUnit, DaysOfCover = InventoryOnHand / (Velocity30/30), StockoutRisk = DaysOfCover < LeadTimeDays. 2) Exclude StockoutRisk SKUs from growth ranking and label them ‘Restock priority’. 3) For remaining SKUs, build a weighted score: 0.5*normalized Velocity30 + 0.3*normalized ContributionMargin + 0.2*(1 – ReturnRate). 4) Output two lists: a) Top 10 Growth SKUs with Score and the 3 metrics behind it; b) Top 10 Restock Priority SKUs with estimated stockout date (today + DaysOfCover). 5) For the #1 Growth SKU, propose 3 tests (budget, expected units, target CAC/AdCostPerUnit) and a breakeven check. 6) If any fields are missing, state conservative defaults to proceed.”

Worked example (simple)

• SKU A: Units 1,000; NetRevenue $25,000; TotalCOGS $10,000; TotalFees $7,000; ContributionMargin $8,000; Velocity30 1,000; Inventory 1,200; LeadTime 20d; DaysOfCover = 36d; No stockout risk.
• SKU B: Units 1,100; NetRevenue $24,000; TotalCOGS $9,900; TotalFees $9,200; ContributionMargin $4,900; Velocity30 1,100; Inventory 300; LeadTime 25d; DaysOfCover = 8d; Stockout risk.

Result: Promote SKU A (higher contribution and safe cover). Restock SKU B first; don’t scale ads until PO is placed.

Metrics to track weekly

• Contribution margin per SKU and in total
• Velocity30 vs Velocity90 (trend and stability)
• Days of cover vs lead time (stockout window)
• Ad cost per unit and blended CAC
• Return rate and refund dollars

Common mistakes & quick fixes

• Missing platform fees — pull FBA/referral/payment fees into a single “PlatformFees” column.
• Variant splitting — map child SKUs (size/color) to one parent if you plan at the parent level.
• Seasonality noise — use both 30 and 90 days; if they disagree, favor 90d for purchasing.
• Promoting stock‑constrained SKUs — move to “Restock priority,” not “Growth.”

7‑day action plan

1. Day 1: Export Amazon + Shopify (30/90 days). Standardize SKUs.
2. Day 2: Add CostPerUnit, PlatformFees, ShippingCost, AdSpend.
3. Day 3: Compute ContributionMargin, Velocity30/90, ReturnRate.
4. Day 4: Add InventoryOnHand, LeadTimeDays; compute DaysOfCover and StockoutRisk.
5. Day 5: Run the AI prompt. Split lists: Growth vs Restock Priority.
6. Day 6: Place POs for top restock SKUs; set ad caps on them.
7. Day 7: Launch a controlled test on top Growth SKU (e.g., +15% bids, 7‑day budget) with targets: +20% units, stable or better AdCostPerUnit, positive ContributionMargin.

Expectation: a defensible, fee‑adjusted top‑10 growth list, a restock list with dates, and a 7‑day test that proves profit, not just volume.

Your move.

[Jeff Bullas]

Nice point — exactly: exclude stock‑constrained SKUs from growth pushes. That one move alone stops wasted ad spend and prevents disappointed customers.

Here’s a practical playbook to get from combined Amazon + Shopify exports to a ranked, profit‑first SKU list you can act on in a week.

What you’ll need

• CSV exports from Amazon and Shopify for the same windows (30 & 90 days).
• A spreadsheet (Google Sheets or Excel) with these columns: SKU, Date, Channel, Units, Revenue, Discounts, Refunds, AdSpend, CostPerUnit, PlatformFees, ShippingCost, InventoryOnHand, LeadTimeDays.
• A chat AI or spreadsheet formulas to summarize and score.

Step-by-step (do this)

1. Consolidate: paste both exports into one sheet. Normalize SKUs (uppercase, strip spaces/symbols).
2. Aggregate: one row per SKU with totals for Units, Revenue, Refunds, AdSpend, InventoryOnHand, etc.
3. Compute core metrics: NetRevenue = Revenue – Discounts – Refunds; TotalCOGS = Units * CostPerUnit; TotalFees = PlatformFees + ShippingCost + AdSpend; ContributionMargin = NetRevenue – TotalCOGS – TotalFees; Velocity30 & Velocity90; ReturnRate = Refunds / Units; AdCostPerUnit = AdSpend / Units.
4. Inventory lens: DaysOfCover = InventoryOnHand / (Velocity30/30). Flag StockoutRisk = DaysOfCover < LeadTimeDays.
5. Score & split: exclude StockoutRisk SKUs from growth ranking. For remaining SKUs build Score = 0.5*norm(Velocity30) + 0.3*norm(ContributionMargin) + 0.2*(1 – ReturnRate). Rank and pick top 10 growth candidates.
6. Act: place POs for Restock priority; run controlled ad tests on Growth candidates only.

Copy-paste AI prompt (use as-is)

“You are a senior retail analyst. I will paste a table with columns: Date, Channel, SKU, Units, Revenue, Discounts, Refunds, CostPerUnit, PlatformFees, ShippingCost, AdSpend, InventoryOnHand, LeadTimeDays. Please: 1) Aggregate by SKU and compute NetRevenue, TotalCOGS, TotalFees, ContributionMargin, Velocity30, Velocity90, ReturnRate, AdCostPerUnit, DaysOfCover, StockoutRisk. 2) Label SKUs with StockoutRisk as ‘Restock priority’. 3) For remaining SKUs, score them: 0.5*normalized(Velocity30) + 0.3*normalized(ContributionMargin) + 0.2*(1 – ReturnRate). 4) Return: Top 10 Growth SKUs with Score and the 3 metrics behind it; Top 10 Restock Priority SKUs with estimated stockout date. 5) For the #1 Growth SKU propose 3 tests (budget, expected units, target AdCostPerUnit) and a simple breakeven check. If any fields are missing, list conservative defaults to proceed.”

Simple worked example

• SKU A: Units 1,000; NetRevenue $25k; COGS $10k; Fees $7k; Contribution $8k; Velocity30 1,000; Inventory 1,200; LeadTime 20d; DaysOfCover 36d → Growth candidate.
• SKU B: Units 1,100; NetRevenue $24k; COGS $9.9k; Fees $9.2k; Contribution $4.9k; Velocity30 1,100; Inventory 300; LeadTime 25d; DaysOfCover 8d → Restock priority.

Common mistakes & fixes

• Missing platform fees — add a PlatformFees column or estimate conservatively.
• SKU variants split — map child SKUs to a parent when evaluating core demand.
• Promoting stock‑constrained SKUs — flag them and move to restock flow instead of ads.
• Seasonal noise — use both 30 & 90 day velocities; favor 90d for purchasing decisions.

7‑day action plan

1. Day 1: Export data and normalize SKUs.
2. Day 2: Add CostPerUnit, PlatformFees, ShippingCost, AdSpend.
3. Day 3: Calculate metrics (ContributionMargin, Velocity30/90, ReturnRate).
4. Day 4: Add inventory & lead times; compute DaysOfCover and StockoutRisk.
5. Day 5: Run the AI prompt and get Growth vs Restock lists.
6. Day 6: Place POs for Restock priority; cap ads for low‑cover SKUs.
7. Day 7: Run a 7‑day controlled test on top Growth SKU (+15% bids or small budget), measure units, AdCostPerUnit, and margin.

Start with the shortlist, add the inventory lens, then test. Small, measured steps turn AI insight into cash — fast.

[Author]

Posts