Introduction

Artificial intelligence and machine learning continue to show promising results for a wide-range of cybersecurity applications. At Bolster, we depend heavily on the latest advancements to achieve ultra-fast, ultra-accurate phishing and scam verdicts at scale. One of the most important tasks that is performed with our technology is logo detection to assist in detecting fraudulent websites and other online infringements in real time. The logo detection algorithms that we've developed and employ are extremely powerful, hardened through our efforts with leading global brands. In this blog we'll introduce a set of use-cases leveraging logo detection and geared towards solving some of the thornier online brand protection problems.

  1. Multi-channel brand monitoring
    Monitoring your brand presence across a myriad of social media platforms, apps stores, marketplaces and websites can be particularly challenging given the scale of the undertaking. Equipped with our logo detection capabilities, you'll be able to take inventory of logo usage across all Internet channels. What's more, in the process of detecting logos, any URLs discovered are scanned in real-time for malicious activity. If phishing or fraud activity is detected on any of the URLs, the system will automatically issue takedown requests with the respective hosting providers. In trickier cases dealing with trademark or copyright infringements, our highly-accurate detection, complete with high-resolution screenshots and logo detection, facilitates evidence gathering, submissions and takedown processes.
  2. Unauthorized & counterfeit sales
    Logo detection can play a significant role in helping you find pathways to unauthorized or counterfeit product sales. Many companies conducting business online struggle with unauthorized sales of legitimate products. And others companies struggle with counterfeit products flooding various e-commerce channels. Both conditions, unauthorized sales and counterfeit sales, can be extremely damaging to your business and your brand's reputation. Here logo detection can play an invaluable role in helping you discover, and dismantle, illicit product placement across the Internet at scale. Use logo detection, for example, to detect fake social media ads promoting fraudulent offers or links to marketplace offers. And use logo detection to also scour marketplaces for unauthorized placements.
  3. Brand impression monitoring
    The vast majority of social media content is rich media including graphics, video, audio. This means that systems relying solely on natural language processing for infringement detection will miss a significant amount of information and incidents. Consequently, for marketers, brand managers and security personnel, visibility into brand placement, performance, and infringements is significantly obscured. Logo detection helps here as an added discovery mechanism. During a big online event, for example, a team equipped with logo detection can extract more meaningful insights in terms of brand placement and impressions across more online channels. The improved fidelity in turn allows teams to more effectively measure ROI.
  4. Copyright infringement detection
    Company logos are easy to get a hold of. A simple screenshot or search on Google is typically all it takes to gain access to a logo file for usage. A common practice by bad actors on the Internet is to take advantage of well-known, trusted, brands to help artificially establish legitimacy through false association. A malicious site might represent themselves as an authorized reseller with a vendor's logo, or might include a customer logo strip of endorsements. Either way, it's prohibitive action that chips away at a brand's reputation and the trust placed in it. Logo detection can help Legal and Marketing teams significantly with both the discovery of infringements like these and associated enforcement processes.
  5. Tracking brand sentiments
    The Internet has become the go-to source for customer feedback and reviews. And in this day and age, all it takes is a few unhappy customers posting harmful or disparaging reviews to 'cancel' a product, or even an entire company. The challenge given all the different online channels and platforms on the Internet is to find where product discussions are occurring, especially the ones involving unhappy customers. Ultimately, the heavy lifting in this scenario is performed by content moderation engines powered by nature language processing, but logo detection is often times used to help locate these posts in the first place. Pairing the two, natural language processing and logo detection, boosts the overall discovery and assessment of customer feedback and sentiment.

Get started today with Bolster!

The online world is fast-moving. And when scammers and counterfeiters target your brand, your detection and remediation windows are narrow. Left unchecked, the business you've built, reputation and all, is at risk. And the risk can result in lasting damages. Trust is after all one of the hardest things to earn, and repair. Legal, Marketing and Security must come together and fully embrace technology to combat the many risks. AI must be at the core of the solution, and logo detection a critical weapon to keep your brand safe and your team sane.

Learn more:

Bolster Brand Protection solutions

Underlying technologies:
Fast Image Matching at Scale
Logo detection using Yolo v5 and Pytorch

Request a demo & free trial: Free trial