Analyzing Business Advisor Data with AI — Hands-On Tutorial

Do you open your Business Advisor every day, glance at the numbers, think "looks okay" or "not great," and then close it?

Most people do.

The data is right in front of you, but you don't know where to start, let alone what to change.

I've been in e-commerce for six years.

For the first three, I selected products, set prices, and designed product images all by gut feeling — results were hit or miss.

It wasn't until I started using AI to analyze Business Advisor data that I truly tasted the benefits of data-driven operations.

Today I'll walk you through exactly how to turn Business Advisor data into actionable optimization plans using ChatGPT.

I've tested this process across twenty-plus stores in different categories, refined it repeatedly, and the version you're seeing now is the finalized standard operating procedure.

Follow it and you'll get results.

Step one: export your data.

Many people think Business Advisor is view-only, but you can actually export.

Go to the Business Advisor homepage, find the Operations Analysis module, select Traffic Analysis, and click the download button in the top-right corner.

Two key options to note.

First, time range — select the last 30 days.

Thirty days of data is substantial enough to show stable trends without being skewed by a single day's promotion or anomaly.

If you select 7 days, the data fluctuates too much to identify real problems.

Second, metric dimensions — check visitors, page views, average dwell time, bounce rate, product detail page conversion rate, and payment conversion rate.

After exporting, you'll have an Excel table with 30 days of daily data for your store.

I also recommend exporting product performance data, which shows each product's visitors, bookmarks, cart-adds, and transactions — useful for product-level analysis.

Why This Tool Stands Out

Step two: feed it to ChatGPT for analysis.

Open ChatGPT, copy and paste your exported Excel data directly, then add a prompt template.

Here's the optimized template I've refined through extensive testing.

"As an e-commerce operations expert, analyze the following 30-day Business Advisor data for a Taobao store.

First, identify three anomaly points in the store's overall traffic trend.

Second, identify the two aspects of conversion rate most in need of improvement.

Third, compare product performance and indicate which products deserve more investment and which need main image or detail page optimization.

Fourth, provide specific actionable optimization suggestions — each suggestion should include operation steps, expected results, and priority level.

" Then paste your data. ChatGPT will start analyzing, usually delivering results in about ten seconds. The first time I tried it, ChatGPT quickly pointed out a problem with my store: my product detail page bounce rate had been above the industry average for three consecutive weeks, and it was higher on weekends than weekdays. This meant my detail pages weren't appealing to weekend leisure browsers. I had been looking at this data for six months without noticing it.

Step three: execute the optimization suggestions.

AI analysis is step one; what actually makes a difference is execution.

Let me walk through a real case.

I coached a sports suit store with 500 daily visitors and a 2.

1% payment conversion rate.

After feeding 30 days of data to ChatGPT, the AI identified several key issues.

First, the bounce rate was 65%, compared to the industry average of around 55% — 10 percentage points higher, meaning users clicked in but left without continuing to browse.

Second, the main image click-through rate was 3.

2%, far below the industry average of 5.

8%, indicating users weren't willing to click their products in search results.

Third, the top three products accounted for 90% of all store traffic, while the other twenty-plus products had almost no traffic — their traffic distribution was overly concentrated and the new product cold-start mechanism was broken.

Manual analysis of these three issues would require an experienced operator a full day of data review. AI found them in under a minute.

For the high bounce rate issue, ChatGPT suggested optimizing the first three screens of the detail page.

Specifically, screen one should display the product's core selling points — for sports suits, fabric stretch data and breathability test results — not brand stories.

Screen two should use comparison images showing the difference between wearing a sports suit vs.

a regular suit: range of motion comparison, comfort comparison — letting users see the differentiation at a glance.

Screen three should feature real customer photos and usage scenario images, not official staged shots.

Based on ChatGPT's recommendations, we redesigned all three screens.

Screen one got an animated fabric stretch test image with stretch data written directly on it — 42% horizontal stretch.

Screen two got comparison photos of wearing them through large movements: bending, arm-raising, squatting — the sports suit was clearly more flexible. Screen three got five customer photos, three of which were video screenshots for authenticity. Within two weeks, the bounce rate dropped from 65% to 52% — a 13 percentage point improvement, meaning over half of visitors were now willing to scroll further.

Core Features Breakdown

Next, we addressed the main image click-through rate.

ChatGPT pointed out that 3.

2% CTR means fewer than 4 out of every 100 users seeing your product actually click in.

The problem was that the main image didn't convey core information — it was just a plain product photo with no selling point labels.

Following AI suggestions, we created four different main images for AB testing.

Image one: white background with promotional tag.

Image two: model wearing the product with fabric stretch label.

Image three: model wearing the product with price discount label.

Image four: video animation cover.

After a week of testing, image two had the highest CTR at 5.

1%, nearly 60% higher than the original 3.

2%.

Scaling up, we standardized all main store images to the model-plus-core-selling-point-label format, and the CTR stabilized around 4.5%. Though still below the industry average of 5.8%, it was steadily improving. This CTR boost also brought more free organic search traffic — when the system sees better CTR, it judges your products more popular and gives you more impressions.

For the over-concentration of traffic, ChatGPT suggested adding a cross-sell recommendation module at the bottom of best-selling product detail pages, manually configuring five or six promising new products.

This way, traffic naturally flows from best-sellers to other products.

Simultaneously, create a separate traffic-driving SKU for these new products with a low threshold price to attract clicks.

After implementing this, within a month, the second-tier products saw visitor increases of 120%, 85%, and 60% respectively.

The traffic structure shifted from 90% concentrated on three products to 70% on five products — healthier distribution with reduced individual product risk.

Even better, these new products, boosted by bestseller traffic, gradually developed their own organic search rankings, creating a positive cycle.

Through these three optimizations, three months later the sports suit store saw a qualitative leap. Daily visitors grew from 500 to 680 — about 35% growth. Payment conversion rate improved from 2.1% to 2.8%. Monthly sales nearly doubled from under 100,000 yuan to nearly 200,000 yuan. Most satisfying was the owner saying he no longer had to guess what to change every day — AI gave him direction, and he just needed to execute. That feeling is far more reassuring than blind trial and error.

Step-by-Step Tutorial

Beyond the three core optimizations, data analysis also revealed an interesting pattern. Looking at the 30-day trend, weekday conversion rates were about 20% higher than weekends. This is because weekday shoppers have stronger purchase intent, while weekend visitors are just browsing. For this, ChatGPT suggested increasing coupon投放 on weekends and making detail pages more engaging with videos instead of static images. After adjustment, weekend conversion rates improved 15%, narrowing the gap with weekdays. This shows that data-driven operations shouldn't just look at averages but also at changes over time.

Another trick few people know: combine Business Advisor data with competitor data.

Manually collect information from direct competitors — prices, review counts, main image styles, title keywords — and feed it all to the AI.

ChatGPT can then analyze not just your own weaknesses but also give specific comparison-based suggestions.

One of my students, a home goods seller, discovered after competitor analysis that the top three competitors all used lifestyle-scene main images while he was still using plain white backgrounds.

After switching, his main image CTR doubled.

Competitor analysis used to require extensive manual collection and comparison.

With AI, you just throw the data in and it automatically analyzes everything — efficiency improved more than tenfold.

Data-driven operations isn't really about whether you can read data — it's about whether you can find clear optimization directions from data. Before AI, a team needed a dedicated data analyst at 10,000 yuan per month minimum, and analysts could only produce reports — the owner still had to come up with specific suggestions. Now with AI, you spend ten minutes exporting data and feeding it, and you get a analysis report worth its weight in gold. Plus AI offers a more comprehensive perspective rather than fixating on one dimension — it synthesizes bounce rate, CTR, conversion rate, average order value, and other metrics for systematic recommendations.

Usage Tips

One final practical tip: make this analysis a weekly ritual. Every Monday morning, spend 15 minutes exporting last week's data and feeding it to ChatGPT to check for new anomalies — a sudden one-day conversion drop, or a category with significantly reduced visitors. Store operations are like sailing: data is your radar, AI is your navigation system. You can reach your destination without navigation, but with it, you sail faster and avoid hidden rocks. Try it today — open Business Advisor, export 30 days of data, feed it to ChatGPT. You might discover issues and opportunities you've been overlooking.

Summary & Recommendations