Breaking news! Big changes to Amazon’s front desk

Amazon's front desk has changed again.

Some sellers have found that the front-end search box of Amazon has changed. Amazon has quietly launched the image search function. After consumers upload pictures, the platform will identify multiple items and recommend similar products. Most categories such as clothing, home decoration, electronic products, kitchenware, etc. can be identified, and sellers are happy to get a new traffic entrance.

On the other hand, perhaps because consumers are not satisfied with Amazon's revised review display, Amazon has recently reversed its actions and gradually resumed displaying the number of product reviews. A large number of search pages display "product star rating + five-pointed star + number of reviews." Buyers' shopping choices are more intuitive, and the status of old products has been preserved.

Amazon launches new features

Yesterday, Amazon launched an important feature - image search. Currently, this feature is only available on the US site, and has not yet been displayed on sites such as the UK, Germany, and Japan.

Click the camera symbol to open the photo search page, upload the picture at the " Upload a Photo " button, and Amazon will identify related items, such as T-shirts, shorts, shoes, etc. Click one of the items, and recommendations for similar styles will appear, along with the number of product colors or patterns, brand, number of reviews and star ratings, and price (including struck-out price). If you drop the mouse, five options will appear: sorting method, gender, Prime, price, and size.

The sorting methods include selected, price from low to high, price from high to low, average customer reviews, and latest models; the price ranges include less than $25, $25 to $50, $50 to $100, $100 to $200, and $200 and above, and users can check one or more price ranges at the same time; the sizes range from XXS to 6XL and even more, allowing users to lock in the right product in one step.

This new feature is applicable to multiple categories. The editor tried to upload photos of daily life scenes, and water cups, mobile phone holders, kitchen utensils, Rubik's cubes, etc. were all recognized. However, in some photos, the recognized items did not have separate display buttons, and the same products were directly displayed on the same page.

Clothing, home furnishings, etc. are the key application categories of the image search function, and Amazon also provides some examples of celebrity styles.

Take an indoor photo of a girl in a skirt as an example. After opening the photo, the page directly gives product recommendations for related elements in the photo, such as butterfly decorations, carpets, bookshelves, etc., and displays the color or pattern number, brand, average star rating and quantity, and price of the product. At the same time, search buttons are added for other products in the photo, including skirts, shoes, stools, etc., and you can click to search for similar styles.

Compared with traditional keyword search, photo search has unique advantages.

When consumers see a style they like online, whether it is clothing accessories or home decoration, they can search the entire image and copy the style they want; when shopping offline or in daily life, if they see a product they like, they can search for the same item by taking a photo without text description or searching for the brand, and they can also compare prices, which makes shopping much more efficient.

Amazon's launch of the image search function may have been inspired by Taobao, which has always been at the forefront of e-commerce shopping exploration and has already mastered the image search function.

When you open Taobao's photo search function, the platform will remind users to "aim at the product and automatically identify it." It can identify multiple products in the camera without taking a photo, such as pillows, facial masks, and water cups in the bedroom, and list the identified items at the top of the mobile page. Users can click on specific items to search for the same item.

Users can also upload pictures in the photo search function, and the software will also identify multiple products, allowing users to select specific items to search for the same item and give related product recommendations, such as DIY accessories on a water cup . In the search results display, users can choose to display the same item by comprehensive, hot sale or price (high or low price), and can choose to display it in a flat or vertical manner.

At present, Amazon is slightly inferior to Taobao in terms of object recognition accuracy. If there are 10 objects in a photo, Taobao can recognize almost all of them in seconds, but Amazon can only recognize a small part of them, such as two or three objects or even less. The following is the recognition effect of the same picture on Taobao and Amazon, the upper picture is Taobao, and the lower picture is Amazon:

When using Amazon image search on a PC and clicking on similar products recommended by the platform, the page will not jump directly to the original page, but will open a new web page. Buyers can return to the original page at any time to choose from many similar products.

After the image search function was launched, sellers have another traffic entrance. Sellers can pay attention to popular styles or popular new products on Tik Tok or other social platforms, and appropriately arrange products to capture traffic from the new entrance; in addition, as prices are divided into 5 intervals, some sellers may slightly reduce prices to enter a lower price range, such as adjusting the price from US$25.5 to US$24.99.

In the image search channel, the number of product reviews will be displayed as usual, and products with more reviews will regain their advantage. In fact, a large number of Amazon pages have started to display the number of reviews again.

Amazon resumes showing review counts

In mid-August, Amazon made a major adjustment to the review display, canceling the traditional display of the number of reviews and instead displaying the proportion of five-star reviews, such as 40% and 80%, hoping to create more opportunities for new products. This surprised sellers who have always attached great importance to the accumulation of reviews. Everyone’s first reaction was that the number of reviews is not that important, new products are expected to break through, and the moat of old products has disappeared.

But before long, some sellers noticed that some categories began to display the number of reviews again. As of this morning, the categories recommended on the homepage of Amazon.com, such as back-to-school supplies, electronic products, game accessories, home kitchen, toys, and health care, all displayed "product rating + five-pointed stars + number of reviews", and many keyword search pages also displayed the same.

Buyers are used to placing orders based on the number of reviews, so Amazon's sudden change in the review display was not well received by the outside world at the beginning. Perhaps buyers showed that they were not satisfied with this during their shopping, and now Amazon has restored the number of reviews.

Momentum Commerce has observed this change.

Starting from August 15, the percentage of search results pages without the number of reviews on Amazon in the United States began to rise sharply . Between August 15 and 25, 24.5% of Amazon US search results did not show the number of reviews. But on August 26, Amazon saw a dramatic pullback, and in the next three days, only 4.9% of Amazon US searches showed no number of reviews.

While this number is still well above the average prior to August 15, it suggests that Amazon may have seen some negative effects from the review display changes, such as a drop in traffic or conversions.

As of the end of August, search result pages with no reviews still accounted for about 5% of Amazon’s US searches. They were not concentrated in certain price ranges or categories, but appeared slightly more frequently in terms with higher search frequency rankings. However, these differences may also disappear completely over time. The moat of the number of reviews for old products has been preserved for the time being.