| | Payment & Financing Services

Already sold out? How bots derail limited sales in e-commerce and what you can do about it

Limited and rare items are highly coveted among fans and collectors – such as sneakers by certain brands. Interest in a product grows, the more exclusive and unique it is. The fashion and lifestyle industry uses artificial scarcity, also known as drop, to boost sales and provide customers with an exclusive brand experience. Resellers often exploit this to resell the sneakers – to continue with the example – for several times their original value. An unwanted side effect which more and more companies are taking technical steps to counter. These technical solutions can be supported by case management to filter out suspicious orders and ensure legitimate customers are still served.

Sneakers from the “Yeezy” brand or fashion products by “Supreme” successively utilize this principle of supply and demand. The goods are only sold at certain locations, at certain times and in a limited quantity. And this practice is highly effective! In just a matter of seconds, these branded products are completely sold out. Similar offers are also available online. Besides fans, resellers also snap up these branded products. To do this, they are increasingly using automated programs (bots) that help them secure the sought-after items just seconds after the sale starts. Mass use of these bots also results in the poor performance of store pages, since a huge quantity of orders are submitted within a short period of time. Consequently, real fans who are willing to buy are left with empty shopping carts. Even if the online shop is quickly sold out, for companies it is impotant that limited products end up in the hands of genuine customers and fans in order to promote brand loyalty. For this reason, successfully running limited drops also involves preventing store visits by bots. Below I have summarized the types of bots used in practice and how you can recognize them.

What types of bots exist?

Add-to-cart bots:  Bots that add the desired product into the shopping cart ready for order submission


Notification bots:  These bots inform the user about offers and the availability of items


Ticket bots:  Bots that draw (multiple) tickets online in order to participate in a prize draw or pre-selection for purchase


Purchase bots:  Bots that handle the complete purchase process for the user


Fraud prevention and protective measures – the path to success

Using a range of technical analyses, it is possible to identify bots and block them at various points of the order process. We utilize technology to apply these methods.

Behavior pattern analysis Here, the user behavior on the website is analyzed for suspicious activity. A normal user clicks through the product range and is unsure exactly what they wish to buy. By contrast, bots move deliberately and with precision, allowing them to be identified.


Speed analysis The speed at which bots connect to and open webpages is much faster than a person can achieve.


User-agent header analysis Many bots use similar user-agent headers. Here, it pays to compare them with known signatures or those already flagged as suspicious in order to verify order queries.


IP address analysis IP addresses are compared with those from known botnets and either blocked or allowed.


Long-term analysis The behavior of some bots can be categorized based on connection requests, their behavior and their activity time. For example, if certain connections are made at certain times of day and they only visit the same websites every time, these IP addresses may be considered to be bots.


Browser analysis JavaScript libraries can force requesting systems to solve mathematical problems in order to determine whether it is a normal browser or a bot. A browser can provide answers, while bots are equipped with minimal functions and are therefore unable to respond to such requests.


Anomaly analysis Requests can be checked for word sequences, typing errors, format, double spaces and other imperfections that only arise in a certain bot network.


This allows a large number of bot orders to be stopped. However, a “gray zone” covers around 10% of all orders and is difficult to classify. Here, we use case management for our customers. Read more about this soon!