What Ails Taobao Search?

Author | Zhou Liqing

Editor | Yang Zhou

"What to do when Taobao search isn't working well anymore?" This query yields over 410,000 related notes in the Little Red Book search bar; similarly, "Taobao search clothes are all the same" garners over 540,000 entries.

These posts converge on a common theme: the declining user experience of Taobao search. When users attempt to locate desired products on Taobao, they often find the search results monotonous, with only three or four distinct styles visible even after scrolling through over ten pages.

Many product titles strive for visibility by piling on lengthy descriptions and features, often vastly different from the product itself. This not only significantly hinders effective user search but also exacerbates user fatigue and distrust.

Previously, users could swiftly find needed products on Taobao through simple keyword searches and filter them using mechanisms like sales volume and positive reviews. However, imperceptibly, these user-friendly designs and experiences have quietly vanished.

To enhance search effectiveness, Taobao has introduced advertising services akin to "front row positions" and "priority display" seen in search engines. These services enable merchants to gain higher exposure and sales opportunities during keyword searches, usually appearing prominently in front of consumers.

While this ad-driven search mechanism offers merchants increased exposure to some extent, it also makes search results increasingly utilitarian and homogeneous.

From a broader perspective, the search issue may stem from strategic vacillation within Alibaba. For a long time, there seemed to be lingering uncertainty within Alibaba about whom the "customers" referred to in the company's core values of "customers first".

In the second half of 2020, Daniel Zhang, then CEO of Alibaba, addressed employee queries about who the customer was by stating, "It's meaningless to discuss whether buyers or sellers come first. Neither is appropriate. We need to consider specific situations on a case-by-case basis." This year, Joe Tsai clarified in an interview video, "Our customers are the individuals who shop using our app." This statement provided an answer to Alibaba's question of "who the customer is".

From this perspective, it should serve as a crucial guideline for Taobao to optimize its search experience—focusing on consumers and improving the search experience accordingly. However, the current situation is far from satisfactory, potentially reflecting deviations in the implementation level during top-down communication within Taobao.

01 Caught in a Search 'Loop'

"You won't uncover treasures on Taobao, but you'll definitely get stuck in a loop."

Wang Miao has been busy preparing for winter on Taobao for the past month. Initially, he planned to buy a thick coat to brave the cold winter, so he would open Taobao every evening after work and type "thickened coat" into the search bar.

After some searching, Wang Miao found that the coats he saw daily were always the same few pieces. "FanoStudios, Huang Maimai, BUD, I almost always encounter products from familiar stores. Scrolling down doesn't reveal any new products." This feels like a loop, and no matter how he changes keywords, he can't escape those familiar "faces".

Wang Miao also noticed that when searching for "thickened coat," a significant portion of displayed products came from stores he frequently visited or already followed. For Wang Miao, who sought more new options, this was somewhat frustrating.

"I want to see more products of diverse styles from various stores, but it keeps recommending those I'm already familiar with, instantly limiting my choices," Wang Miao said somewhat helplessly.

On one hand, users encounter the same products during their searches; on the other, the product styles and models they find are severely homogeneous.

Sun Tong shared his experience searching for "padded jackets" with "City Image." On one occasion, while searching for "padded jackets" on the Taobao homepage, he was drawn to a hooded padded jacket with a fur collar, so he clicked to view the details.

But when he exited that product interface and continued refreshing, he found the screen kept showing the same "ZuoShanDiao" padded jackets from different stores, almost identical in style, with only slight price differences. "I originally wanted to see more styles of padded jackets, but all I saw were 'ZuoShanDiao' jackets, like walking into a dead end, unable to find a new direction."

Even more absurd than the "loop" is the mismatch of search functions.

When Wang Miao typed "short down jacket" into the search bar, the results were mixed with many products unrelated to "short down jacket," such as sheepskin coats and fur-lined coats. These products' titles did not contain the keywords.

02 Multiple Failures in Search Functions

Currently, searching for stores on Taobao also poses challenges. Sun Tong mentioned seeing a fashion blogger share a store on Little Red Book that appealed to his style preferences, so he went to search for it on Taobao. However, after typing the store name from memory, he couldn't find it.

"I initially thought the store had closed down but then doubted it because such a popular store wouldn't suddenly disappear." So, Sun Tong privately messaged the blogger, who copied the store name to him, and he finally found the store.

"Especially for stores with English names, searching for them is even more difficult. I found that even subtle differences like case sensitivity and spaces can determine whether or not a store can be found. This is really inconvenient for us consumers."

Moreover, the user experience of "order search" after shopping on Taobao has also garnered much criticism.

Shang Wen shared his experience with "City Image": "The keyword I searched for is clearly in the title, but I can't find the order when searching for it."

Shang Wen bought a pair of shorts this summer, and his friend also wanted the same pair, so Shang Wen confidently went to search for it in his historical orders.

"The pants are made of denim, so I searched for 'denim pants' in my historical orders, but I couldn't find the order at all, even though I did buy them. I had to search for just 'pants,' dare not add even the word 'jeans,' and then I found it."

Shang Wen said this situation is not occasional and that now when searching for historical orders, he only dares to use the most basic keywords like "dress" and "pants," fearing any slight mistake will cause him to miss the order he's looking for.

Compared to Shang Wen, Zhang Lan's order search experience was the opposite.

The year before last, Zhang Lan bought two bags from the same Taobao store. Last week, on a whim, he wanted to check if the store had any new products, so he searched for "bags" in his historical orders. However, after repeated searches, he couldn't find the orders for those two bags. It was only when he searched for "tote bags" based on the bag style that he finally found the related order and then the other through the store name.

Although Zhang Lan and Shang Wen encountered different issues, they both faced a common dilemma: the search keyword was clearly in the product title, but they couldn't find the corresponding order.

Many netizens have also reported that the historical order list is not arranged in normal chronological order but is completely disorganized. When browsing orders, one often suddenly encounters very old orders among recent ones.

When consumers inquire about order search issues with customer service, they often receive a mechanical response, "Sorry for the poor service experience," without practical solutions.

03 What's the Root Cause?

Taobao search is not an isolated case; major e-commerce platforms' search functions seem troubled. Some netizens bluntly stated, "Search on 1688 is even more absurd. Slightly detailed search terms directly yield no results."

On Little Red Book, there are over 40,000 related notes about "Pinduoduo search keywords not matching results," and there have even been absurd situations like "searching for hats yielding dresses and food."

Specifically analyzing Taobao, from a technical perspective, employees responsible for the search and recommendation subdivisions at Taobao revealed that the search core intelligent engine Ha3 and the search ranking core BERT are both operating normally. Taking the issue of homogeneity as an example, in technical implementation, the system must perform deduplication processing, a basic operation to ensure search result diversity.

However, consumers still frequently encounter repeated search results. "This is likely due to deviations in the search strategy." The search and recommendation team provides technical capabilities, but it's the algorithms and operational strategies that ultimately determine how search results are presented.

Similarly, regarding order search, in cases where searching for "pants" only yields some orders, from a technical level using the inverted index, the possibility of failure is extremely low. Theoretically, the search is fully capable of displaying all related "pants" orders. The aforementioned person stated the problem likely lies in the operational strategy, such as during the word segmentation phase, perhaps using the full name "jeans" from the start, thereby narrowing the search scope.

Looking back at Taobao search's development history, early Taobao search engines won user favor with their precision.

A seasoned 17-year Taobao user recalled, "The initial Taobao search engine could extremely accurately search for all items under corresponding terms based on keywords. For example, when you typed 'blue square-necked puff-sleeve shirt,' it could really find all matching shirts on the Taobao platform and sort them precisely by sales volume, positive reviews, and other dimensions."

Times have changed, and today's Taobao search is no longer purely function-oriented.

According to Taobao merchants, Taobao analyzed data and concluded that at least a dozen interactions with buyers are needed for optimal conversion results. Under this logic, advertising's share in the entire recommendation system has gradually increased.

Jiang Xia, an e-commerce practitioner, told "City Image": "As users' time is fragmented across many platforms, e-commerce platforms increasingly focus on efficient traffic utilization. Therefore, platforms tend to recommend products more likely to result in sales. If the product also has a higher average transaction price, it's naturally a better choice."

Under this premise, merchants' behaviors are also influenced by platform strategies.

"When placing advertisements, merchants will also focus on recently added-to-cart and favorited customers for additional investment to increase the conversion probability."

Jiang Xia further pointed out, "To ensure traffic acquisition and conversion, merchants are gradually becoming homogeneous in product development, tending to copy popular products or make minor innovations based on them, which has a higher success rate. Those originally differentiated products, especially those unable to effectively improve conversion rates, struggle to obtain sufficient traffic, making it difficult for merchants to recover development costs. Therefore, fewer merchants are willing to try new things."

Merchants all creating "similar popular products" undoubtedly exacerbates platform product homogeneity, further deteriorating the user search experience.

As a result, users may be trapped in the strategic framework set by the platform and merchants, with their access to product types becoming increasingly limited and shopping experience compromised.

(The characters in the article are pseudonyms)