My first year in this industry, team discussions on product issues basically went like this:
“I think the main problem with this product is quality.” “No, I’ve talked to customers and they’re mainly complaining about logistics.” “That’s because you haven’t read the reviews — reviews are all about functionality problems.”
Two hours of arguing, finally the boss decides: “Let’s fix the packaging first.”
After fixing packaging, negative review rate didn’t budge.
This isn’t a management problem — it’s a cognition problem.
Three Fatal Flaws of “Gut Feeling” Mode
Flaw 1: Everyone’s “Feeling” Is One-Sided
Operations people contact customer service daily, think “customer service is the biggest problem”; Design people read reviews daily, think “appearance is the biggest problem”; Supply chain people handle returns daily, think “quality is the biggest problem.”
Everyone’s feeling is right, everyone’s feeling is incomplete.
Because you haven’t systematically read all reviews — you’re only seeing “what you’ve noticed.”
Flaw 2: Cannot Distinguish “Severity of Problems”
“I think quality is the biggest problem” “I think so too” “Then let’s fix quality first”
Here’s the problem: “Quality” is a very broad word.
Is it material problems? Workmanship problems? Durability problems? Or incomplete functionality?
Without quantitative analysis, you can’t even determine “what to fix first.”
Flaw 3: Cannot Verify Results After Decision
After fixing it, how do you know you fixed it right?
Look at ratings? Ratings are lagging — might take two to three months to reflect. Look at negative review count? Affected by total review volume, can’t see without standardization.
Making decisions by gut feeling, verifying effects by gut feeling — this is a dead end.
Three Core Advantages of “Data Speaks” Mode
Advantage 1: Full Data, Complete Perspective
Use tools to analyze all reviews — no sampling, no omissions.
For example, a product has 500 reviews. Someone “going by gut feeling” might read 50. Someone “using data” reads all 500.
This isn’t 50% vs 100% — it’s “partial” vs “complete picture.”
Advantage 2: Quantified Sorting, Clear Priority
Not “quality problems are biggest” — but “quality-related negatives are 161, accounting for 41% of negative feedback, with 123 concentrated on material brittleness and poor durability.”
This statement can guide action: Fix material first or durability first? Do quality inspection first or change suppliers first?
Advantage 3: Conclusions Verifiable, Effects Quantifiable
Make decisions with data, verification is simple: analyze again after fixing, compare data changes.
Negative rate drops from 28% to 22%, quality-type negatives drop from 41% to 35% — that’s effective improvement.
Not “feel much better,” but “data speaks.”
A Real Transformation Case
I previously coached a seller doing Bluetooth headphones. Problem was “negative review rate has always been high, don’t know what to fix first.”
I had him analyze reviews using the data method.
“Gut feeling” stage judgment:
- Boss thinks it’s “price too high, user expectations mismatched”
- Operations thinks it’s “competitors have more features, we have no advantage”
- Customer service thinks it’s “return process is troublesome, users can’t wait”
“Data-driven” stage results:
- Product issues 61% (mainly “noise” and “insufficient battery”)
- Service issues 26% (mainly “unclear instructions”)
- Experience issues 13% (mainly “uncomfortable to wear”)
Conclusions are completely different.
Based on data guidance, they did three things:
- Changed suppliers to solve noise problem
- Rewrote instructions to solve pairing problem
- Adjusted earbud size to solve wearing problem
Three months later, negative rate dropped from 28% to 18%.
This isn’t voodoo — it’s the power of data.
My Current Decision Process
Now when I make product decisions, it’s basically this process:
Step 1: Data Collection Use tools to collect all reviews — no sampling, full collection.
Step 2: Three-Dimensional Classification Divide negatives into “product issues/service issues/experience issues” three categories. Not simply saying “there are negatives,” but saying “41% are product issues, 3% are service issues.”
Step 3: TopN Sorting Find the N issues with highest proportions and sort them. For example, Top5 issues cover 74% of negative feedback — these 5 issues are the “main battlefield.”
Step 4: Trace to Original Reviews Click into each issue to see original reviews, ensure not misjudged. For example, “quality issues” — what specific type of quality? Material, workmanship, durability?
Step 5: Create Improvement Plan Each issue corresponds to specific improvement actions:
- Material issues → contact supplier/change materials
- Instruction issues → rewrite instructions
- Wearing issues → adjust product design
Step 6: Verify Effects After running changes for a period, analyze again and compare data.
Why Learn to “Use Data”?
Because Amazon competition is escalating.
Previously you could launch any product and make sales — information asymmetry, users didn’t know where to compare prices. Now users can see more products, platform algorithms are smarter, competitors are more intense.
Sellers going by “gut feeling” will gradually be eliminated from competition.
Not because the product is bad, but because they don’t know which part is bad, can’t fix in the right direction.
Sellers using “data” know what to fix first, what to fix later, how to fix — every investment sees returns.
This is the key to closing the gap.
Question in comments: Does your team currently make decisions “by gut feeling” or “by data”? If using data, what tools? How’s the effect?