Amazon's allure is undeniable. Between
For many merchants, however, it's an effort that almost always ends in failure. Intense competition leads to high, often insurmountable cost-per-acquisition (CPA) and low average order values (AOV), which, in turn, sifts out new entrants that don't have the resources for sustained experimentation.
Recently, artificial intelligence (AI) has flipped the script. One of AI's biggest advantages is its incredible capacity for massive data analysis, and it's a feature that's seriously underutilized by merchants worldwide.
Painstaking competitor research that would take weeks can now be done in a day. Insights that you'd need a trained eye for can now be served up on a platter. And human error and personal biases that might skew results – poof, (mostly) gone!
Let me show you how.
Here's a 5-step guide for leveraging AI to supercharge Amazon conversions – by creating better ads, visual materials, and product descriptions and even improving your product through extracting invaluable insights from competitor reviews.
Note that since the principles and methods discussed in this article can be transposed to other activities, you may find these insights applicable even if you’re not selling on Amazon.
What we'll be doing
For this exercise, our focus is on user reviews. We aim to conduct an in-depth analysis of what customers say about your competitor's products, as it's typically one of the most revealing and honest sources of information.
By looking at reviews, we learn not only what the customers think of the product, but also discover what's important to them – why they bought it, whether or not their expectations were met and in what way, what they like/dislike about the product, and so much more. Any online seller should treat customer reviews as the goldmine they are.
Ultimately, here's what we're doing: we want to feed as many relevant customer reviews as possible into the AI and then ask targeted questions to gain insights about the buyers, competitor product shortcomings, and more.
With that done, we'll have identified what matters most to customers, areas where you can outshine your competition, and who your target audience is – all valuable information that you can weave into your advertising, product copy, and visuals to maximize your product's appeal to Amazon shoppers.
Let's break it down step-by-step.
Step 1: Identify your competitors
Pop on Amazon and find the 3 biggest competitors for your product with at least 200 written reviews each. That's it – that's the first step. Realistically, if you've had any ambition of selling on Amazon, you should already have this completed.
Step 2: Scrape user reviews
Step 2 is about getting those user reviews in a format that the AI can read (in our case - .CSV or .txt).
Luckily, you don't need to go through all of the reviews and manually copy them into a spreadsheet, as there are dozens of Amazon review scraping tools available online at various price points. Personally, I've been using Apify, however, recent pricing changes have me considering alternatives.
Do note that since Amazon is fighting against these tools, their availability and functionality may change in the future – so make the most of this opportunity while it's easily accessible. Also, I’m not affiliated with any of the tools I mention in the article.
Typically, scrapers are super straightforward and you need little-to-no technical know-how to use them – simply paste the competitor's product URL in the scraper and then download the data in a suitable format.
As a result, you're going to have massive, organized spreadsheets with competitor product reviews.
Step 3: Upload scraped data to AI
Next up, you'll want to upload them to an AI tool, namely
Formerly, Claude was by default the better option since it allowed substantially higher input and output allowances, however, ChatGPT has caught up in this regard (with 120k context windows) and may serve just as well.
Which one should you use?
Consider the following:
- ChatGPT is available worldwide but requires the paid version (~$20) for the file upload feature.
- Claude is only available in the US and the UK. While the free version of Claude may be enough for you, you need a VPN to access it if you're based outside the US and the UK.
Either tool works, so your call!
Once you've registered, simply upload your scraped user reviews to your AI of choice.
Step 4: Condition the AI to give you optimal insights
Here's where we currently stand: we have hundreds of product reviews from your biggest competitors uploaded into a tool that can comprehensively analyze them in moments – the AI can identify patterns, find outliers, discover hidden details, and even enhance the data with publicly available information.
There's only one problem – we need to help the AI get its gears turning in the right direction for the best insights. At this point, we may ask all sorts of questions, but unless we “condition” it to “think” the way we want it to, those answers will be all over the place and, hence, suboptimal.
So, how do we condition it?
By asking the right questions.
What are the right questions?
There's no definitive group of right questions. That said, for our purposes, I recommend asking the following set – it's one I've developed through much trial and error and I've found to (currently) generate the best results. Do note that it's important you stick to the order below, as each consecutive answer will inform how the AI answers the next question.
Here are 8 AI prompts that will ultimately help you boost your conversions:
-
“Based on the reviews provide a buyer persona table with persona type, description, optimal age range, lifestyle description, motivation to buy, education level and percentage of how many reviews fit each persona”
-
“Create a new table with each buyer persona and add what they love, what they didn't like or even hated, what emotions they have expressed - positive and negative, what technical aspects they liked and didn’t like”
-
“Create a table with the 10 most popular positive emotional things expressed in the reviews and add percentage of how many reviews fit each emotional thing”
-
“Create a table with the 10 most popular negative emotional things expressed in the reviews and add percentage of how many reviews fit each emotional thing”
-
“Create a table with the 10 most popular positive technical things expressed in the reviews and add percentage of how many reviews fit each technical thing”
-
“Create a table with the 10 most popular negative technical things expressed in the reviews and add percentage of how many reviews fit each technical thing”
-
“Are there any paradoxes in the reviews, false negatives?”
-
“Based on everything we have learned about the product and the buyers, develop a marketing executive summary”
And just like that – you've got detailed buyer personas, you can pinpoint competitor product or service weak points, you know what info you need to include in your assets, and so much more.
Step 5: Create better assets
Equipped with intimate information about your main competitor's customers and their expectations, you can put it to use in optimizing your own strategy and resources – create better ads, improve product features, emphasize the benefits that matter, use the language your customers speak, include features that matter, etc.
Is your buyer persona moms with several kids? Have a mom with several kids in your visuals.
Are people fawning over a particular feature? Give it center stage in your copy.
Are people complaining about long delivery times? Offer them a short delivery time guarantee.
The comprehensive analysis enables you to speak your audience's language like never before and that's exactly where the value lies. Of course, you still need to pair the insights with high-quality copywriting, product photos, etc., but now you can do it while being much better informed.
Is it really that easy?
In broad strokes, yes. You can replicate this process to gain invaluable insights about your product category.
There are a few caveats though.
First, knowing what questions to ask and – just as importantly – HOW to ask them of the AI is a skill that may take some time to hone. In this type of undertaking, the devil's in the details, and while general questions can certainly give you insights to boost sales, getting the absolute maximum out of the reviews will require you to dive into the nitty-gritty and think outside the box. So, while the blueprint detailed in this article will certainly help boost conversions, expect it to evolve in time.
Ultimately, it's a matter of time, creativity, and trial & error.
Secondly, do be cautious with the AI. It can misunderstand things – both the data and your requests. Do not hesitate to challenge it, for example, by asking how it reached certain conclusions. There's nothing worse than basing your strategy on false premises, so really take the time to build solid foundations.
Lastly – this isn't a magic bullet. You need to continuously put in the work, learn, improve, optimize, and innovate to stay ahead of the curve. We're still a ways off from tools that do it all for you automatically.
That said, the work is well worth it. These customer review insights can help you improve your messaging across the board, giving your marketing efforts a boost not only on Amazon, but for your other channels, too – your store, your ads, your customer support, and beyond.
The technology's out there and easily available. Now I've shown you how to use it.