Amazon reviews to Google Sheets, insights your team uses
You finally get a steady stream of Amazon reviews… and then you realize you can’t actually use them. They’re scattered across pages, written in different styles, and the worst ones (1–2 stars) hide the real issues behind rants.
Product Managers feel it when prioritization turns into guesswork. A brand owner sees it when returns creep up. And support leads get stuck answering the same complaint because nobody turned Amazon review insights into a clear list of problems. This Amazon review insights automation fixes that.
You’ll see how the workflow pulls low-star reviews, summarizes them with Gemini, sorts them into categories your team can act on, then logs everything in Google Sheets and pings Slack when it’s done.
How This Automation Works
See how this solves the problem:
n8n Workflow Template: Amazon reviews to Google Sheets, insights your team uses
flowchart LR
subgraph sg0["Manual Start Flow"]
direction LR
n0@{ icon: "mdi:play-circle", form: "rounded", label: "Manual Start Trigger", pos: "b", h: 48 }
n1@{ icon: "mdi:swap-vertical", form: "rounded", label: "Configure Workflow Settings", pos: "b", h: 48 }
n2@{ icon: "mdi:swap-horizontal", form: "rounded", label: "Execute Actor Fetch Dataset", pos: "b", h: 48 }
n3@{ icon: "mdi:swap-vertical", form: "rounded", label: "Iterate Through Reviews", pos: "b", h: 48 }
n4["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/slack.svg' width='40' height='40' /></div><br/>Post Slack Completion Notice"]
n5@{ icon: "mdi:robot", form: "rounded", label: "Send Model Prompt", pos: "b", h: 48 }
n6["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/code.svg' width='40' height='40' /></div><br/>Parse JSON Script"]
n7@{ icon: "mdi:database", form: "rounded", label: "Generate Spreadsheet", pos: "b", h: 48 }
n0 --> n1
n5 --> n6
n3 --> n4
n3 --> n5
n6 --> n7
n7 --> n3
n1 --> n2
n2 --> n3
end
%% Styling
classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef disabled stroke-dasharray: 5 5,opacity: 0.5
class n0 trigger
class n5 ai
class n2 decision
class n7 database
class n6 code
classDef customIcon fill:none,stroke:none
class n4,n6 customIcon
The Challenge: Turning Amazon Reviews Into Decisions
Reading negative reviews sounds simple until you do it every week. You open Amazon, filter to 1–2 stars, click through pages, and try to remember what you saw. One review says “cheap,” another says “broke,” another blames shipping, and suddenly you’re stuck: is this a quality issue, a packaging issue, or a customer expectation issue? Even if you take notes, they live in someone’s doc, not in a shared system your team trusts. Honestly, the biggest cost is momentum. You lose it every time you start over.
It adds up fast. Here’s where it usually breaks down.
- You spend about 2 hours a week just collecting and skimming reviews, then still need another pass to find patterns.
- Complaint themes don’t get categorized consistently, so “shipping damage” and “arrived broken” end up as separate problems.
- Important feedback gets lost because it’s trapped in long paragraphs instead of a spreadsheet row your team can sort and filter.
- Teams stop checking reviews regularly because the process is annoying, so issues linger longer than they should.
The Fix: Scrape Low-Star Reviews, Summarize, Categorize, Log
This workflow starts with a single run in n8n, then does the review grunt work for you. It sends an Amazon product URL to an Apify scraper that fetches reviews already filtered to 1–2 stars, so you’re focusing on complaints instead of praise. Next, it loops through each review and asks Google Gemini to produce structured output: a sentiment score, an issue category (Quality, Design, Shipping, and so on), a short summary, and a practical improvement suggestion. A small parsing step cleans the AI response into neat JSON so it stays consistent. Finally, every review becomes a new row in Google Sheets, and Slack gets a “done” message once the batch finishes.
The workflow begins with your configured product URL and scraping settings. Gemini turns messy text into uniform fields you can filter in Sheets. Then the automation writes everything to your tracker and confirms completion in Slack so nobody has to babysit the run.
What Changes: Before vs. After
| What This Eliminates | Impact You’ll See |
|---|---|
|
|
Real-World Impact
Say you review one product listing every Monday and you look at 60 low-star reviews. Manually, even 2 minutes per review to read, interpret, and jot a note is about 2 hours, plus another 30 minutes to summarize themes for the team. With this workflow, you spend about 5 minutes setting the URL and kicking it off, then wait while it processes. You come back to a Google Sheet that’s already sortable by category and sentiment, and Slack tells you when it’s complete.
Requirements
- n8n instance (try n8n Cloud free)
- Self-hosting option if you prefer (Hostinger works well)
- Apify to scrape Amazon reviews into a dataset.
- Google Sheets for a shared, filterable review tracker.
- Google Gemini API key (get it from Google Cloud Console).
Skill level: Intermediate. You’ll connect accounts, paste an API key, and edit one URL plus a prompt.
Need help implementing this? Talk to an automation expert (free 15-minute consultation).
The Workflow Flow
You trigger a run on demand. It starts from a manual trigger, which is perfect when you want to review one ASIN today, then a competitor tomorrow.
Apify collects the right reviews. The workflow sends your Amazon product page to the Apify actor and pulls back a dataset filtered to 1–2 star feedback, so you aren’t wasting time on generic praise.
Gemini turns text into structured insights. Each review is processed in a loop, and Gemini returns a sentiment score, a category, a summary, and an improvement suggestion. A parsing step cleans the output so it fits the spreadsheet reliably.
Google Sheets becomes the source of truth. Every structured response is appended as a new row, then Slack sends a completion notice once the batch is finished.
You can easily modify the category list or the language of the summaries based on your needs. See the full implementation guide below for customization options.
Step-by-Step Implementation Guide
Step 1: Configure the Manual Trigger
This workflow starts on demand, so you’ll trigger it manually during setup and testing.
- Add the Manual Start Trigger node as the trigger for the workflow.
- Keep the default settings (no fields required) in Manual Start Trigger.
- Optionally leave Flowpast Branding as a reference note in the canvas.
Step 2: Connect Apify
Set review scraping parameters and run the Apify actor to collect review data.
- In Configure Workflow Settings, set maxReviews to
10and apifyActorId tojunglee/amazon-reviews-scraper. - Open Execute Actor Fetch Dataset and set Operation to
Run actor and get dataset. - Set Custom Body to
{ "startUrls": [ { "url": "https://www.amazon.co.jp/dp/B09JDGYSQW" } ], "maxReviews": 10, "filterByRatings": [ "oneStar", "twoStars" ], "proxyConfiguration": { "useApifyProxy": true } }. - Credential Required: Connect your Apify OAuth2 API credentials in Execute Actor Fetch Dataset.
Step 3: Set Up Send Model Prompt
Configure Gemini to analyze each review, then parse the JSON response for spreadsheet-ready output.
- In Send Model Prompt, set Model to
models/gemini-2.0-flash-lite. - Set the prompt content in Send Model Prompt to
=あなたは熟練の商品開発マーケターです。 以下のAmazonレビューを分析し、情報をJSON形式のみで出力してください。 { "sentiment_score": (1〜5の数値。1が最悪、5が最高), "category": ("Quality", "Price", "Function", "Design", "Shipping", "Other" から最も当てはまるものを1つ選択), "summary": (ユーザーの不満点を日本語で30文字以内に要約), "improvement": (具体的な改善案を日本語で50文字以内で提案) }. - Credential Required: Connect your Google Gemini credentials in Send Model Prompt.
- In Parse JSON Script, keep the JavaScript Code as provided to clean and parse the model response.
```json and ``` to prevent parsing errors.Step 4: Configure Output/Action Nodes
Create a spreadsheet, process each review in batches, and send a Slack completion notice.
- In Generate Spreadsheet, set Resource to
spreadsheet. - Credential Required: Connect your Google Sheets credentials in Generate Spreadsheet.
- Use Iterate Through Reviews to batch items coming from Execute Actor Fetch Dataset and Generate Spreadsheet.
- Set the Slack message in Post Slack Completion Notice to
=✅ Amazon レビュー分析が完了しました！ 処理したレビュー数: {{ $('Iterate Through Reviews').item.json.batchSize }} Notionデータベースを確認してください。. - Choose the target channel in Post Slack Completion Notice using Channel (replace the placeholder value).
- Credential Required: Connect your Slack OAuth2 credentials in Post Slack Completion Notice.
Step 5: Test and Activate Your Workflow
Run a manual test to confirm the review scraping, AI parsing, spreadsheet creation, and Slack notification all complete successfully.
- Click Execute Workflow to run Manual Start Trigger.
- Verify Execute Actor Fetch Dataset returns review items and Iterate Through Reviews begins batching.
- Confirm Send Model Prompt returns JSON, Parse JSON Script parses it, and Generate Spreadsheet creates the spreadsheet.
- Check that Post Slack Completion Notice posts to the selected Slack channel with a batch count.
- Once successful, toggle the workflow to Active for production use.
Watch Out For
- Apify credentials can expire or your actor rental can hit limits. If scraping suddenly returns empty results, check your Apify run history and token permissions first.
- If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
- Default prompts in AI nodes are generic. Add your brand voice early or you’ll be editing outputs forever.
Common Questions
About 30 minutes if your accounts are already set up.
Yes. You won’t write code, but you will connect Apify, Gemini, and Google Sheets credentials and paste in a product URL.
Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in Gemini API usage and your Apify actor costs, which depend on how many reviews you scrape.
Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.
You can adjust the Apify “Run an Actor” settings to target a different product URL, change the 1–2 star filter, or increase how many reviews are collected per run. In the Gemini “Message a model” prompt, swap the category list to match how your team talks (for example, split “Quality” into “Durability” and “Materials”), and change the output language if you don’t want Japanese. If you already have reviews in a spreadsheet, you can skip scraping entirely and feed the loop from Google Sheets instead of Apify.
Usually it’s a permissions issue on the Google account connected in n8n, or the spreadsheet ID is wrong. Also check that your header columns match what the workflow is trying to append (sentiment_score, category, summary, improvement). If the sheet is owned by another account, make sure the connected Google user has edit access, not view-only.
On n8n Cloud, capacity depends on your plan’s monthly executions, and self-hosting removes the execution cap (your server becomes the limit). Practically, most teams run this in batches of a few dozen to a few hundred reviews at a time so the AI analysis stays stable and you don’t hit rate limits. If you need to process thousands, run smaller batches and schedule it.
Often, yes, because this workflow needs scraping, looping through many items, and structured AI output handling. n8n is comfortable with that kind of “batch job” logic, and self-hosting keeps costs predictable when volume grows. Zapier and Make can work, but long loops tend to get expensive and fiddly. If you’re only processing a handful of reviews occasionally, those tools can be fine. If this is going into a real weekly operating rhythm, n8n is usually the calmer choice. Talk to an automation expert if you want a quick recommendation for your exact setup.
Once this is running, your team stops “reading reviews” and starts making decisions from them. Set it up, let it collect the signals, and use your time for fixes that actually move the product forward.
Need Help Setting This Up?
Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.
