How Machine Learning will Benefit Advertisers and Publishers in 2019
More so than previous years, 2018 brought significant challenges and opportunities to the digital media world.
In 2019, publishers and advertisers will leverage advancements in artificial intelligence (AI), namely machine learning, to limit ad fraud, ensure brand safety and comply with data protection regulations.1,2,3
Here’s what machine learning is, and the benefits it will bring to both advertisers and publishers this year.
What is Machine Learning?
Machine learning is a method of achieving AI4 that uses algorithms to gradually “learn” from data without explicitly being told how. 5
For example, publishers and advertisers can use machine learning for predictive analytics. By examining previous customer behavior in real-time, they can identify trends to recommend relevant content or ads that match a user’s interests. 6
Empowered by this technology, publishers and brands will be able to overcome new and persistent challenges in 2019. 7,8
Machine Learning Combats Ad Fraud
According to Juniper Research, advertisers lost $19 billion to fraudulent activities in 2018 and this figure is expected to reach $44 billion by 2022. 9
Technologies that were in their infancy in 2018 will become more important in 2019, including blockchain, which uses a cryptography-encrypted pipeline to ensure that users are served the correct ads with less fraudulent attacks. 10
When combined with advanced machine learning systems, blockchain will reduce ad fraud even more efficiently in 2019. 11
For example, AdBank has submitted a patent for a machine learning fraud detection system with plans to implement it on their ad network. The system uses multiple machine learning algorithms to review the blockchain data structure and conduct automated checks of suspicious transaction behavior. 12
“We are really excited to be gaining considerable traction not only in the crypto community but also with publishers and advertisers alike by becoming their champions,” said Angelo Dodaro, CMO, and Co-founder of AdBank. “Every day we have new advertisers and publishers approaching us to find out how they can use Adbank to grow their revenue and eliminate fraudsters and middlemen. It is so rewarding to provide real solutions for companies and fuel their growth.” 13
Machine Learning Addresses Brand Safety Concerns
In early 2017, brand safety became a primary issue for advertisers when programmatic ads from Mercedes-Benz, Waitrose and Marie Curie appeared next to Islamic State and pro-Nazi videos. 14 Later in the year, ads from Deutsche Bank, Adidas, Amazon and eBay were served next to pedophilic video content. 15
According to a report by CMO Council, 72 percent of marketers are concerned about brand safety in programmatic adverting, and more than a quarter have inadvertently had their content appear alongside compromising content. 16
Thankfully, modern AI-powered algorithms can detect context and content in nuanced ways to ensure that ads are not served inappropriately. 17
In December 2018, Integral Ad Science released a Brand Safety and Suitability solution that uses machine learning to review and re-calibrate models that ensure brand safety and suitability. 18
“Brand safety can’t be solved alone, it is going to take all the stakeholders in our industry, working together, to find a solution,” said Harmon Lyons, SVP of Global Business Development, IAS. “We saw such great success because YouTube, IAS and prominent brands were all willing to listen to each other, work together and create a solution that advertisers can trust to verify their investment.” 19
Learn how advertisers are addressing brand safety concerns
Machine Learning Ensures Data Protection Compliance
Earlier this year, the General Data Protection Regulation (GDPR) disrupted the digital media industry by requiring organizations to get consent from European users before collecting cookies that include personal data. 20
According to a report by Sizmek, 77 percent of brand marketers expect their audience targeting with third-party data to be limited due to the regulation. 21
While GDPR only impacts companies in the United States that do business with European users, similar data protection regulations are expected to be enforced at a state level in 2019.
For example, the California Consumer Privacy Act of 2018 grants consumers the right to request the data that businesses collect and to ask businesses not to sell their data. 22 As more states implement complex data protection regulations, advertisers will need to be increasingly cautious or risk fines. 23
One way for advertisers to overcome GDPR challenges is through contextual targeting. In fact, 87 percent of marketers said they plan to increase contextual targeting in the next 12 months. 24
One implementation of this method is semantic contextual targeting, which uses machine learning to interpret the content of webpages in a nuanced way. Unlike keyword targeting, semantic contextual targeting analyzes the sentiment behind content and can interpret when words have multiple meanings. In turn, advertisers will be able to target relevant context categories and place ads next to content that exhibits positive sentiment without the use of third-party data. 25
In 2018, concerns about brand safety, ad fraud and GDPR ushered digital media into a new era of innovations in machine learning – innovations that are keeping publishers and advertisers ahead of the curve.
2019 promises to build upon these advancements in the foreseeable future.
1. “The Outlook for AI & Its Practical Applications.” ExchangeWirecom, 19 Dec. 2018.
2. “The Battle Against Ad Fraud in 2019.” ExchangeWirecom, 17 Dec. 2018.
3. “Predictions for Artificial Intelligence in 2019.” InformationWeek, 2018.
4. McClelland, Calum. “The Difference Between Artificial Intelligence, Machine Learning, and Deep Learning.” Medium.com, Medium, 4 Dec. 2017.
5. Phillips, Mary-Katharine. “What Is Machine Learning?” Twipe, 15 Nov. 2018.
6. Recchia, Chad. “Three Everyday Examples Of How Machine Learning Has Impacted Advertising.” Forbes, Forbes Magazine, 2 Feb. 2018.
7. “Artificial Intelligence and News Trends in 2019.” What’s New in Publishing | Digital Publishing News, 3 Dec. 2018.
8. “The Ultimate GDPR Resource Guide for Publishers.” What’s New in Publishing | Digital Publishing News, 26 Nov. 2018.
9. “Ad Fraud to Cost Advertisers $19 Billion in 2018, Representing 9% of Total Digital Advertising Spend.” Juniper Research, 2018.
10. Lubek, Sven. “Top 2019 Digital Marketing Trends and Predictions.” Chief Marketer, Chief Marketer, 8 Jan. 2019.
11. Bhattacharya, Joydeep. “An Online Ad Fraud Counterattack Built with Blockchain, AI and Cryptography.” B2B News Network, 21 May 2018.
12. “Patent Update: Adbank’s AI Ad Fraud Detection System.” Medium.com, Medium, 27 Nov. 2018
13. “Adbank Is Strategically Changing The Online Advertising Industry With the Blockchain.” Influencive, 27 Dec. 2017.
14. Mostrous, Alexi. “Big Brands Fund Terror through Online Adverts.” The Sunday Times, The Sunday Times, 9 Feb. 2017.
15. Mostrous, Alexi, and Katie Gibbons. “YouTube Adverts Fund Paedophile Habits.” The Sunday Times, The Sunday Times, 24 Nov. 2017.
16. Brand Protection From Digital Content Infection.” CMO Council, 2018.
17. “Is Artificial Intelligence the Remedy for Brand Safety Woes?” Appier, 8 Nov 2018.
18. Integral Ad Science, 2018.
19. IBID
20. Nadeau, Michael. “What Is the GDPR, Its Requirements and Facts?” CSO Online, CSO, 23 Apr. 2018.
21. “New Sizmek Research Reveals Most Marketers Concerned About Walled Gardens and Data.” Sizmek, 16 Oct. 2018.
22. Sweeney, Erica. “California Adopts Privacy Law Aimed at Protecting Consumer Data.” Marketing Dive, 2 July 2018.
23. “Blockchain in Marketing: Will 2019 Be a Year More of Hope or Hype?” Marketing Tech News, 2019.
24. Jackson, Marinn. “In a Post-GDPR World, Contextual Targeting Is Reborn.” AW360, 11 Sept. 2018.
25. “Could AI Combined with Context Negate the Need for Third Party Data?” Marketing Tech News, 2018.
26. Icon by Eucalyp from Flaticon.com. Licensed by Creative Commons.