Product badges are ubiquitous in e-commerce platforms, acting as efective psychological triggers to nudge customers to buy specific products, consequently boosting revenues. However, to the best of our knowledge, there has been no attempt to systematically study these badges and their several idiosyncrasies - we intend to close this gap in our current work. Specifically, we try to answer questions such as: How do the products that receive badges difer from those which do not, in terms of price, customer rating, etc.? How long does a product retain a badge on a given platform? If a product is sold on diferent platforms, then does it receive similar badges? We collect longitudinal data from several e-commerce platforms over 45 days, and find that although most of the badges are short-lived, there are several permanent badge assignments, and that too for badges meant to denote urgency or scarcity. Furthermore, it is unclear how the badge assignments are done, and we find evidence that highly-rated products are missing out on badges compared to lower quality ones. Our work calls for greater transparency in the badge assignment process to inform customers, as well as to reduce dissatisfaction among the sellers dependent on the platforms for their revenues.
Today, e-commerce marketplaces like Amazon, eBay, Rakuten, and Alibaba have emerged as
indispensable online platforms helping millions of customers with their purchase needs. At
the same time, they provide livelihood to thousands of sellers and producers worldwide [
When a customer visits an e-commerce website and searches for something to buy, products
matching the search query are returned. Interestingly, the returned product lists often contain
products highlighted with certain badges, such as, ‘Best Seller’, ‘Deal of the Day’, ‘New Arrival’,
‘Discount’, etc. These badges evoke specific psychological reactions in a customer and nudge her
to buy the corresponding products. For example, a ‘Best Seller’ badge means that many other
customers have already bought this popular product and hence, one can go forward and buy it.
Whereas the badge ‘Deal of the Day’ creates a sense of urgency that the customer may miss out
on a wonderful opportunity if that product is not bought immediately. In fact, product badges
have been found to increase purchase conversion rates by as high as 55% [
Recognizing the importance, several Software-as-a-Service (SaaS) providers for smaller
ecommerce businesses (including WooCommerce [
However, since the badges can be assigned to only a handful of products without risking the
dilution of customer attention, it may deprive some products of getting the attention they truly
deserve. Specially in some platforms, getting a badge can bring huge revenue opportunities for
certain products. For example, the products with ‘Amazon’s Choice’ badges can be bought on
Amazon with zero clicks! If a customer uses a voice-activated device, such as Echo or Alexa, and
says “Alexa, order a pillow”, then the ‘Amazon’s Choice’ product would be the default choice
for that keyword [
Despite its ubiquitous presence and associated implications, only a few research studies
have focused on product badges. Adaji et al. [
In this work, we bridge this gap by extensively collecting data from 12 e-commerce platforms over 45 days. Out of these 12 platforms, 10 cater to only niche domains, and the rest 2 are generic which sell a wide variety of products across multiple domains. All of our data is made publicly available1 which, to the best of our knowledge, is the first comprehensive dataset that focuses on product badges, and includes details such as name of the e-commerce platform, query executed, returned results ranked in order with brand names, badges assigned, prices, discounts, ratings, links, and other useful relevant features. Using this data, we attempt to answer the following three research questions:
RQ1. Assignment bias: How do the products that receive badges difer from those which do not? What is so special about these products with badges? RQ2. Longevity: How long does a product retain a badge on a given platform? Does it vary depending on the platform, domain, or type of badges? RQ3. Uniformity: If a product is sold on diferent platforms, does it receive similar badges? Our investigation reveals that though most of the badges are short-lived, there are several products that retained a badge over the full data collection period. Most surprising are the badges denoting scarcity or urgency; one might expect them to be short-lived, yet we find these assignments to be permanent. The same products seem to get diferent types of badges on diferent platforms; so there is no cross-platform consensus in badge assignments. We did not ifnd any clear pattern in the product choices for badge assignment; in fact, products without badges seem to be higher rated than the products with badges. Overall, we take the first step towards exploring the badge landscape in this work, and we hope it can spawn future works looking at product badges in a holistic manner.
In this section, we briefly survey the prior works on e-commerce platforms and digital nudging.
2.1. eCommerce platform designs
eCommerce websites deploy multiple techniques to attract and retain customers to increase their
revenue, which include persuasive strategies targeting human psychology [
Even within a platform, there are various means to steer a customer’s attention to a specific
product, which may include explicit identifiers, such as stars, votes, reviews, and badges,
along with some implicit ones, such as the rank of the product in the recommendation list [22].
However, unlike the rest of the explicit identifiers, badges are conferred to a chosen few products.
Zhang [23] looked into how the ‘SuperHost’ badge in Airbnb helps improve sales at a higher
charge. There are external companies that provide specialized badges to e-commerce websites,
such as Shopify [24] which provides payment related trust and security badges, and Yotpo [25]
1https://github.com/architbansal28/Product-Badges
which provides product reviews related badges. Adaji and Vassileva [26] reported how Amazon
persuades its customers to write more reviews by giving the ‘Hall of Fame’ badge. They further
looked into various persuasive strategies that are employed in e-commerce [
Thaler and Sunstein [27] introduced nudging as a tool to achieve societal goals by exploiting some mental shortcuts people take. Following this work, there have been many applications of nudge theory in several domains, including in digital platforms. For example, Mota et al. [28] designed interface nudges to steer more donations to poorer schools in an educational charity platform. Bhuiyan et al. [29] designed nudges to increase consumption of credible news on Twitter, which can help to curb misinformation. A comprehensive survey on digital nudging can be found by Jesse and Jannach [30]. According to their taxonomy, product badges should fall under the category ‘Attracting Attention’, which, in turn, falls under the super-category ‘Increase salience of information’. However, the authors did not talk about product badges other than in a footnote, where they describe product badges as ‘visual highlighting tools’.
Saari et al. [31] tried to categorize diferent persuasive techniques in e-commerce. Unfortunately, product badges have been skipped in this work as well – in fact, product badge does not belong to any of the categories mentioned by the authors, although the ‘visual layout’ category comes close. Another way to nudge customers to buy products borrows ideas from gamification [ 32], such as, through loyalty points, coaxing customers to enter contests, etc. Some researchers have also criticized the interface design choices of e-commerce platforms to push users into making unintended purchases as dark patterns [33].
Overall, we see that psychologically targeting customers and subtly persuading them towards purchasing specific products is prevalent in e-commerce; however, product badges seem to be a comparatively new phenomenon and not well explored in the literature. In this work, we attempt to bridge this gap.
As mentioned in the introduction, in this work, we attempt to explore the badge assignments across diferent e-commerce platforms. For this purpose, we looked into platforms that serve a niche domain vis-a-vis generic platforms that sell a variety of products across domains, including the niche ones. We are also interested to know whether the same products sold on diferent platforms get similar badges. Therefore, we restricted ourselves to websites that target a specific country; otherwise the products (brands) being sold might be quite diferent. In this study, we have chosen websites that are operational in India.
We concentrate our study on four niche domains: (i) baby products, (ii) cosmetics, (iii) fashion, and (iv) home decor. These domains were chosen based on two factors. Firstly, we tried to make the niche domains as diverse as possible. Though there may be small overlaps, e.g., websites related to baby products and fashion may both include garments for babies; we carefully selected the queries to collect data from diferent websites so that the returned product lists are diferent across domains. Secondly, the selected domains should have websites that are popular in India.
For each domain, we picked two or three niche websites, totaling ten websites across four domains, as listed in Table 1. As generic websites, we chose Amazon and Snapdeal, which sell products across all these domains.2 Next, we selected ten diferent queries for each of these four domains so that the corresponding e-commerce websites (niche and generic) returned a considerable number of products as results. These queries are also reported in Table 1.
To automatically collect the product listing against each query, we used the Selenium WebDriver (selenium.dev) to automate the process of firing the search query to these websites and retrieve the results returned. Further to ensure that the site does not realize that the same user is firing the queries and therefore somehow get influenced while assigning badges, we cleared the cookies and launched a fresh search every time we fired a query. We considered the results up to a maximum of 100 products, and also stored other metadata associated with the products, such as product name, price, discount, average customer rating, number of ratings, and most importantly, the badge assigned to a product (if available). We call one instance of the data collected across all websites as a snapshot. Our dataset contains data collected over 45 days during June–August 2021, with 2 snapshots per day at a gap of 12 hours, i.e., altogether there are total 90 snapshots. Overall, the data consists of 62, 889 unique products across platforms, out of which 27, 016 (42.9%) products got some badge in at least one of the snapshots.
As discussed in the last section, we explore a total of 12 e-commerce websites, and all of these assign multiple types of badges to diferent products, where these badge names are often unique to the site. To perform a cross-platform study, we group diferent badges into coherent categories. Badges are designed to trigger specific psychological reactions from customers, so that they take 2It is worth noting that for platforms with a global presence, we have considered their India specific website, e.g., Amazon.in. Badges assigned at diferent e-commerce websites and their respective categories. Here ‘X’ and ‘Y’ represent diferent numbers used in the actual badges – for example, ‘Only 5 Left in Stock’, ‘Buy 3 Get 2’, etc.
Best Seller, Kids Gift Ideas
Only X Left in Stock Amazon Deal of the Day, Limited Time Deal, Deal is X% Claimed, Prime Day Deal Amazon’s Choice Prime Day Launch
Few Left, Last Sizes Left - Special Price Bewakoof Flash Sale
Buy X Get Y, Buy X for Y, Color of the Month
Best Sellers, Trending, Glam Award e.l.f. Cosmetics New
Holy Grail
Bestseller, #1 Mom’s Pick FirstCry XNeLweft!
Super Saver Hopscotch X Left
Flash Sale, Freshness Unplugged Season Sale, Brand Day Sale LimeRoad NOefewr: Buy X Get Y Free, Ofer: Freebie, Best Price
Exclusive Max Fashion BNueywX Get Y, Buy X at Y, Buy X at Y% Of, Flat X% Of My Baby Babbles Sale
Bestseller
Sale Nykaa New
Ofer Featured
Best Seller Pepperfry CNleewarance Sale! 30/100 Nights Trial
Trending, Top Seller, X Orders in Last Y Days, X People Just Ordered, X% Positive Feedback Snapdeal X Left!
Featured
Best Seller Urban Ladder Only X Left
New Arrival
Social Proof
Scarcity
Urgency
Endorsement
Exclusivity
Scarcity
Urgency
Promotional
Social Proof
Recency
Endorsement
Social Proof
Scarcity
Recency
Promotional
Scarcity
Urgency
Recency
Promotional
Exclusivity
Recency
Promotional
Urgency
Social Proof
Urgency
Recency
Promotional
Endorsement
Social Proof
Urgency
Recency
Promotional
Social Proof
Scarcity
Endorsement
Social Proof
Scarcity
Recency
a mental note of the products with the badges and get swayed towards buying them. Following
the lines of Cialdini [
Social Proof: These badges indicate that the product is popular among customers, e.g., ‘Best Seller’, ‘Trending’.
Scarcity: This type of badges communicate that the product is only available in a small quantity, e.g., ‘Few Left’, ‘Only 5 Left in Stock’.
Urgency: These badges convey that one needs to buy the product quickly, else she will miss a great opportunity, e.g., ‘Deal of the Day’, ‘Limited Time Deal’.
Recency: This type of badges indicate that the product has been introduced recently, e.g., ‘New Arrival’, ‘New’.
Promotional: This type of badges indicate that by buying the product now, a customer will receive a special ofer or discounted price, e.g., ‘Buy 1 Get 1’, ‘Super Saver’.
Endorsement: These badges carry the endorsement from the platform or some expert, e.g., ‘Amazon’s Choice’, ‘Featured’.
Exclusivity: This type of badges indicate that the product is available exclusively on a platform or for a chosen set of customers, e.g., ‘Exclusive’, ‘Prime Day Launch’.
It is worth noting that a type of badge may invoke a mix of emotions. For example, a Scarcity badge also indicates that the product should be bought urgently and hence can also be put in Urgency category. Furthermore, a Scarcity badge also tacitly implies that the product has been selling fast, and thus serves as a Social Proof. Similarly, a badge like ‘10 People Just Ordered’ can belong to both Social Proof and Recency categories. To reduce the chance of misinterpretation, three annotators were asked to independently assign categories to the badges we encountered in our dataset, by examining their nomenclature, and we choose the category of a badge based on the majority opinion. Table 2 lists all badges present on diferent websites and their respective category information. Note that we did not include the ‘Sponsored’ badge in our study because it is assigned in exchange for a fee, and thus does not shed any insight on the badge assignment policies. For similar reasons, we also did not consider ‘Out of Stock’ and ‘Sold Out’ badges in our study.
To understand the role that badges play while buying products online, we conducted a survey that 203 people participated in. Out of 203 participants, 81% reported that they buy from ecommerce sites on a mid to high frequent basis. We asked each participant 10 questions where we showed her a pair of items – exactly one of these items had a product badge, and asked “Suppose you are looking for <query string>. Which of the following would you prefer to buy?” It is important to note that the pair of items had similar images, ratings, and prices, so that the badge becomes the primary diference between the two while not being explicitly so. We had a couple of questions to test the participants’ perception of the availability of items with Scarcity badges, and another couple of questions to test the preference for one type of badge over another. We shall get to the details of these in the subsequent sections, where we motivate the research questions posed by us on the basis of the survey responses.
Our survey showed that badges indeed have a positive influence on customers; we noticed that the number of participants who chose products with badges varied from 65% to 88% across products. Since having a badge helps a product get the desired traction with the customers, it is No Badge
Badge No Badge
Badge
Prices (in INR) 0 10000 20000 30000 40000 0 20 40 60 pertinent to ask whether the products with badges are any diferent from those which did not get badges. This can also help predict the underlying badge assignment policies employed by diferent websites. We look into the following four attributes: price, discount, average customer rating (CR), and number of customer ratings to compare the products. To clarify the attribute ‘discount’, if the original price of a product is but currently it is available at a lower price , then the discount is − . It is important to note that e-commerce websites can give discounts without any explicit Promotional badges. Moreover, in such cases, we consider the ‘price’ of the product to be its current selling price, i.e., in our example.
Since the badge assignment to a product can vary across snapshots, we followed two schemes while creating the list of products with ‘Badge’ and with ‘No Badge’: 1. we include a product in the ‘Badge’ list if it has received a badge in at least one snapshot; 2. we include a product in the ‘Badge’ list only for those snapshots where it has received a badge, and do not include for snapshots where it did not get a badge. Figure 1 shows the comparison between these two types of products in both schemes for the FirstCry website. Interestingly, there is no significant diference in the comparison results between these two schemes. For both, products with badges ofer higher discounts; a few of them are higher priced, although the median price is similar for both types. Surprisingly, we found that the average rating is higher for products without badges, raising concerns about whether lower quality products are being promoted using the badges. We found a similar trend for other platforms as well.
In addition to the earlier experiments, we dived into the recommendation lists to see which products received multiple badges in a platform. We found that 9 out of 12 websites assigned 2, 3, or 4 badges to a single product (in diferent snapshots or, in some rare cases, even in the same snapshot). For example, in Amazon, Harpa Women’s A-Line Dress received 4 badges: Amazon’s Choice (Endorsement), Best Seller (Social Proof ), Only X Left in Stock ( Scarcity), and Deal of the Day (Urgency). Similarly, in Nykaa, Lakme Eyeconic Liquid Eyeliner received 4 badges: Featured (Endorsement), Bestseller (Social Proof ), Sale (Urgency), and Ofer ( Promotional). These results are captured in Table 3.
In Table 4, we provide the number of instances, accumulated over all platforms, where two badge categories were assigned to the same product. Thus, if a product receives badges of three diferent categories, then we will add it to the count of all the three possible category-pairs. Note that no product was assigned both Endorsement and Exclusivity badges, and hence only this pair is missing in Table 4. From this table, it seems that websites take a two-pronged approach – Urgency and Promotional – to quickly sell specific products. Furthermore, we conjecture that popular products having Social Proof badges may often sell fast, and when the resources are depleted, Scarcity badges are assigned to these. We think Recency and Promotional badges probably go hand-in-hand to mitigate the cold start problem in recommendations. Similar opinions may be formed for other category-pairs with high counts.
After categorizing the badges, we turn our attention to the stability of badge assignments at diferent e-commerce platforms. Knowing how long a product holds a particular badge (termed as ‘Longevity’) is important for multiple reasons: (i) if a Scarcity or Urgency category badge is given to a product for long, it can indicate potentially unfair practice by the platform; retaining the same badge to particular products (ii) may lose its value for the repeat customers, and (iii) may deny opportunities to other products.
In fact, on being asked about when they expect a product with Scarcity badge to be sold out, 15% of our survey participants answered “in a few hours”, and 44% answered “in a few days”.
Longevity of a badge for a product (ℒ, ) is defined as ℒ, = # snapshots with having badge total number of snapshots .
For example, if Product A has badge X for 60 snapshots out of 90, then ℒ , Average longevity of a badge (ℒ ̂ ) is calculated over all products having this badge: is 6900 = 0.67. ℒ ̂ =
∑ ℒ, ∑ ℐ(ℒ, > 0) , where ℐ is the indicator function.
Box plots in Figure 2 present the range of ℒ, values across products and badges in diferent platforms for the query ‘men casual shirt’, with ℒ ̂ values shown in the parentheses. While Figure 2(a) shows the ℒ, values across all badges in a platform for the query (values in parentheses denoting average ℒ ̂ ), all other sub-figures show ℒ, values for individual badges in respective platforms. We see that Max Fashion, on average, has badges with the highest longevity; in some cases, the longevity is 1 for Max Fashion and Bewakoof. Diving in, for Max Fashion, New (Recency) badges, and for Bewakoof, Buy X Get Y (Promotional) badges are the longest lasting. Amazon, on the other end, gave away badges with the least average longevity. However, the Best Seller (Social Proof) badges in Amazon have high longevity, including 1 in some cases; in contrast, Deal of the Day (Urgency) and Only X Left in Stock (Scarcity) badges have low longevity, thereby bringing down the average longevity for this platform.
After exploring the longevity of product badges at a query level, we aggregate the observations across all queries for a platform to identify macro trends. Specifically, we first consider all products with low (0 < ℒ, ≤ 0.2), medium (0.2 < ℒ, ≤ 0.5), and high (ℒ, > 0.5) longevity. Figure 3 show these products across diferent websites under diferent domains. Our data indicates that the longevity of badges for the majority of products is rather small and there is a large churn in the assigned badges. When we look into the individual domains, it is hardest to retain a badge in Fashion, while in Cosmetics, the badges are retained fairly long.
We also look into the longevity of the product badges based on the categories that these belong to. Figure 4 shows an overview of the same. 0.0 0.2
Additionally, we use another metric Consistency ( , ) as a measure of how long a product retains badge continuously. This is related to Longevity, but here the focus is on the contiguous time stretch. We measure , as , = ( # contiguous snapshots with having badge ) total number of snapshots .
For example, if a platform assigns badge to product in snapshots 1, 2, and 3 but not in 4, , would be 34 = 0.75.
Along with Longevity, we also computed Consistency, and we were particularly interested in badges belonging to the Scarcity and Urgency categories. Ideally, for these two categories, Home Decor
Fashion 100% 80% ts 60% c u d o r P 40% # 20% 0% 100% 80% ts 60% c u d o r P 40% # 20% 0% 100% 80% s tc 60% u d o r P 40% # 20% 0% 100% 80% s tc 60% u d o r P 40% # 20% 0%
Amazon Snapdeal Urban Ladder Pepperfry Social Proof Scarcity Urgency Recency Promotional Endorsement Exclusivity Social Proof Scarcity Urgency Recency Promotional Endorsement
Exclusivity Baby Products
Cosmetics Home Decor
Fashion Low
Medium Longevity
High
Low
Medium Longevity
High consistency values should be low. In fact, we captured multiple daily snapshots to track the outflux of these ‘rare’ products. However, the actual scenario is quite diferent.
For Bewakoof, we found that 9 products (all under ‘women kurti’ or ‘women jackets’) had a Scarcity badge for all the snapshots. There were 26 products that had a consistency value of more than 0.6 – all of these products are for ‘women’; among products for ‘men’, the highest Amazon Snapdeal Nykaa e.l.f. Cosmetics
Amazon Snapdeal LimeRoad Max Fashion Bewakoof Social Proof Scarcity Urgency Recency Promotional Endorsement Exclusivity Social Proof Scarcity Urgency Recency Promotional Endorsement Exclusivity consistency value was 0.52. Following a similar pattern, in FirstCry, there were 11 products that had a Scarcity badge for all the snapshots, and all of these products appeared under ‘baby girl dresses’. For gender-agnostic searches in ‘Baby Products’, products under ‘stroller’ typically had the highest consistency values, which were around 0.4, barring one (Tify & Tofee Portable Stroller with Canopy) that had a consistency value of 0.69 in FirstCry. In Hopscotch, the number of products that had a Scarcity badge all throughout was 26 – however, most of these belonged to ‘onesies’ and ‘baby footwear’. Moreover, the consistency values for the products under ‘baby boy clothes’ and ‘baby girl dresses’ were comparable for this website. Contrarily, the generic websites typically had lower consistency values, especially Amazon, which had a ‘sofa set’ with the highest consistency value of 0.71. Snapdeal, however, had a handful of products (8) with consistency values above 0.9; two of these were ‘face primer’ and the rest belonged to ‘sofa set’ or ‘dining set’. Although we found that the highest consistency values in the generic websites mostly belonged to the ‘Home Decor’ domain, surprisingly, the niche website Urban Ladder had lower consistency values, with the highest one being 0.61 belonging to a ‘table lamp’.
In contrast to Scarcity badges, Urgency badges had relatively lower consistency values. For the websites Nykaa, LimeRoad, and Pepperfry, the average consistency values were 0.2, 0.3, and 0.32, respectively. For Bewakoof, one product under ‘men polo t-shirt’ had a consistency value of 0.92. My Baby Babbles, however, had high consistency values (>0.88) for almost all products under ‘baby shampoo’, ‘baby towels’, and ‘baby diaper’. Three products in Amazon – all moisturizers from Mamaearth – had Urgency badge in all 90 snapshots. See Table 5 for more examples.
Our survey had specific questions where we showed multiple similar products, each with a diferent badge category. Almost 47% of the participants chose the products with Social Proof badge (e.g., ‘Best Seller’). Endorsement badge (e.g., ‘Amazon’s Choice’) came second with 24%; none of the other categories seem to have a significant preference over others. Thus, getting diferent badges may have a direct impact on the sales of the sellers. Hence, we investigate whether products sold on multiple platforms are assigned similar badges. We found that if a product is present on multiple platforms, the product names are similar (if not exactly the same) across websites, and contain the brand name and product specifications like color, fiber (for apparel), etc. For example, ‘Lakme Insta Eye Liner, Black, Water Resistant, Long-Lasting, 9 ml’ on Amazon and ‘Lakme Insta Eye Liner - Black’ on Nykaa. To account for the possibility of not finding an exact name match, we match the products on diferent platforms based on parameters like brand, color, etc. Moreover, even the names of the badges are often unique across the platforms; thus, we look for matches in the badge category.
Surprisingly, out of the 29 possible platform-pairs, we found that only 10 of these had at least one product that was assigned a badge on both, as mentioned in Table 6. In this table, we mention for each platform-pair, the number of queries (out of 40 for Amazon–Snapdeal pair, and out of 10 for every other pair) for which there is at least one common product that has received the same badge category, the number of diferent badge categories that these products belong to, and two scores:
, which we explain with the following example.
Let us consider that there are two platforms, and . For simplicity, let there be a single badge category ℬ. Now let the products having a ℬ-type badge and their corresponding Longevity in the two platforms be as follows: ∶∶ { ∶ 0.3, ∶ 1.0, ∶ 0.4, ∶ 0.2} , and ∶∶ { ∶ 0.2, ∶ 0.6, ∶ 1.0, ∶ 0.1} . We compute as the ratio of the number of common products to the number of all products in the two platforms belonging to the same badge category. In this example, = |{∶∈ ∧∈ }|
|{∶∈ ∨∈ }| On the other hand, we define
= |{,,,,, }| |{,}| to the Longevity of all products in the two platforms. Here imply a higher
, or vice versa. 0.3+1.0+0.4+0.2+0.2+0.6+1.0+0.1 = 13..58 = 0.3947. Note that a higher
0.3+0.2+0.4+0.6
Predictably, the maximum number of queries for which there is some common product with an identical badge category occurs for the generic platform-pair: Amazon–Snapdeal. In terms = 26 = 0.33. as the ratio of the Longevity of common products = Σ{∶∈∧∈} Σ{∶∈∨∈} ℒℬ, ℒℬ,
= does not necessarily Uniformity of badges from diferent categories.
, we see that the pair Amazon–Nykaa has the maximum uniformity. If we consider only those platform-pairs where both websites are niche, then we find Nykaa–e.l.f. Cosmetics to have the maximum uniformity in terms of both scores. These observations suggest that there is highest uniformity in the Cosmetics domain compared to other domains considered here. Exclusivity badges, there was no overlap between any platform-pairs, and hence these categories do not appear in Table 7. Based on , Social Proof badges are the most common – this result may seem intuitive because popular brands should be in demand irrespective of platforms, and the longevity of Social Proof badges is normally higher. On the other hand, Scarcity badges seem to be assigned most uniformly across platform-pairs and queries.
This work has several takeaways for the diferent stakeholders in an e-commerce business, namely, the customers, the sellers, and the marketplace platform provider. For the customers, ifrstly, they should be aware of the diferent emotional reactions that the badges try to trigger – being aware should save them from falling an easy prey; for example, some items may not be as scarce as the Scarcity badges would like one to believe. Secondly, products with badges are not necessarily better in quality than those without them, and hence, they should ideally buy after careful comparison.
For the sellers, although they may vie for product badges to get that extra boost in sales, they should bear in mind that badge assignment, especially when it comes to the Endorsement category, does not necessarily reflect that the products with badges are superior to those without. Moreover, securing a badge in a platform is also unlikely to guarantee similar success in another platform.
For the e-commerce platforms, first and foremost, they should make the badge assignment process transparent to achieve a level playing field for all the sellers, and thus prevent dissatisfaction among smaller sellers. Furthermore, they should closely regulate badge assignments to restrain some products from retaining a badge for a prolonged period, especially for Scarcity, Urgency, and Recency badges. If standardization of product badge assignment policies can be achieved, it would allow more uniformity across platforms as well, which would provide extra impetus to the sellers to ensure that their products are of high quality, and this, in turn, would benefit the customers. [21] Z. Chen, H. Cao, F. Xu, M. Cheng, T. Wang, Y. Li, Understanding the role of intermediaries in online social e-commerce: An exploratory study of beidian, Proc. ACM Hum.-Comput.
Interact. 4 (2020). [22] M. Schuckert, X. Liu, R. Law, Stars, votes, and badges: How online badges afect hotel reviewers, Journal of Travel & Tourism Marketing 33 (2016) 440–452. [23] C. Zhang, Home sharing economy: Reputation badge and hosts competition, SSRN (2018). [24] shopify, Anyone, anywhere can start an ecommerce website, https://www.shopify.in, 2006. [25] yotpo, ecommerce marketing platform, https://www.yotpo.com, 2021. [26] I. Adaji, J. Vassileva, Evaluating personalization and persuasion in e-commerce, in:
PPT@PERSUASIVE, volume 1582 of CEUR Workshop Proceedings, 2016, pp. 107–113. [27] R. H. Thaler, C. R. Sunstein, Nudge: Improving decisions about health, wealth, and happiness, HeinOnline, 2008. [28] N. Mota, A. Chakraborty, A. J. Biega, K. P. Gummadi, H. Heidari, On the desiderata for online altruism: Nudging for equitable donations, Proc. ACM Hum.-Comput. Interact. 4 (2020). [29] M. M. Bhuiyan, M. Horning, S. W. Lee, T. Mitra, Nudgecred: Supporting news credibility assessment on social media through nudges, Proc. ACM Hum.-Comput. Interact. 5 (2021). [30] M. Jesse, D. Jannach, Digital nudging with recommender systems: Survey and future directions, Computers in Human Behavior Reports 3 (2021) 100052. [31] T. Saari, N. Ravaja, J. Laarni, M. Turpeinen, K. Kallinen, Psychologically targeted persuasive advertising and product information in e-commerce, ICEC, 2004, pp. 245–254. [32] J. Karac, M. Stabauer, Gamification in e-commerce - A survey based on the octalysis framework, in: HCI, 2017, pp. 41–54. [33] A. Mathur, G. Acar, M. J. Friedman, E. Lucherini, J. Mayer, M. Chetty, A. Narayanan, Dark patterns at scale: Findings from a crawl of 11k shopping websites, Proc. ACM Hum.-Comput. Interact. 3 (2019).