India’s InMobi is Disrupting Online Advertising by Making 15 Billion Decisions Per Day
InMobi runs an ad-network that decides the value of over 15 billion ad transactions per day using machine learning. In just a decade, InMobi has positioned itself as a challenger to ad tech incumbents Google and Facebook. Yet in the race to be the best ad-networker, InMobi and others will need to balance product innovation with protecting their advertisers, publishers, and selves from consumer data misuse and brand deterioration.
In the world of online advertising, data is king when it comes to better targeted ads. The industry has been long dominated by the likes of Google and Facebook[1], but India’s new unicorn InMobi is using machine learning (ML) to threaten their incumbency.
Founded in 2008 in Bengaluru, India, InMobi runs an ad-network which connects advertisers with publishers to serve ads to consumers. Today, InMobi handles over 15 billion transactions per day, and they expect that volume to quadruple over the next year[2]. Their secret sauce is in how they add value for advertisers using ML. Each time a consumer visits one of their publisher’s platforms (i.e. an app or website), InMobi receives a packet of data about the consumer and the opportunity from the publisher. This data is fed through InMobi’s ML algorithms to assign a value to the opportunity for each advertiser based on a system which identifies similar customer profiles and their likelihood of engaging. That valuation decides the bid that each advertiser makes on the ad opportunity, and the winning bidder gets to serve their ad. All of this happens in real-time, and InMobi’s ML algorithms are constantly evolving to produce better results based on new data[3].
In the race to add the most value for advertisers, InMobi continues to invest in its technology to improve its valuation algorithms and ad-network. Their key investments include:
- Detecting Publisher Fraud with ML: Following instances of publisher fraud in recent years (i.e. the Methbot Operation which generated fake ad views[4]), InMobi invested in developing ML technology to detect fraud in their network. Reducing fraud increases the ROI for advertisers as fewer ads are wasted on fake viewers.
- Ensuring Ad Quality: In March 2018 InMobi announced a partnership with advertising security experts at GeoEdge to help ensure only high quality advertisements are allowed in InMobi’s ad-network[5]. This protects InMobi’s publishers by reducing the likelihood that offensive ads will be shown on their platforms.
- Partnering with Microsoft: In June 2018, InMobi announced a partnership with Microsoft to collaborate on technology and go-to-market strategies[6]. In the long term, this partnership will allow InMobi to develop and scale new capabilities more quickly for their products.
- Acquiring Pinsight Media: In October 2018, InMobi acquired Pinsight Media for its cutting-edge data management technology as well as its ability to distill marketing insights and profile customers using massive data sets[7]. In being able to provide marketing insights, InMobi wants to move up the marketing value chain from simply executing the advertising strategy to actually influencing it.
While InMobi is making valuable investments in protecting advertiser dollars and publisher integrity, as they increasingly automate aspects of the ad-network, they may want to consider focusing on mitigating the risks of removing humans from the process. For example, in 2017, InMobi settled with the FTC after being charged with tracking consumers’ locations without their permission. InMobi’s main competitors have faced issues too with their ad-networks: in 2016, Facebook inadvertently allowed the spread of misinformation ads around the 2016 US presidential election[8], and in 2017 Google had major advertisers pull their ads from YouTube after learning they were being shown next to offensive content[9]. These recent mishaps in the ad-networking space serve as a reminder that protecting consumers’ data and advertisers’ brands will be important pieces of building a sustainable business. To this end, there are a few actions I would recommend:
- Work with Governments: Governments take privacy seriously. As InMobi rushes to use customer data in inventive new ways, partnering with governments now can help them ensure they remain compliant with local regulations.
- Benchmark Competitors: With the public eye turned to Facebook due to their recent ad scandals, InMobi can benefit from the public information that Facebook is sharing about new practices they are implementing.
- Add Content Tagging to ML Algorithms: A relevant advancement in the field is the ability to use ML to automatically tag content with topics. For example, and ad featuring a blackjack table may be tagged with “casino”, “blackjack”, and “cards”. Given the massive volume of content flowing through InMobi’s ad-network, automated tagging could be used to gain insight to which types of content perform best for various customer profiles. That information could allow InMobi to better value the opportunity for each advertiser at a more granular level based on their available content. Microsoft recently patented a content tagging system[10].
As InMobi continues to grow, it will be important to find the right balance between investing in new features and investing in risk mitigation. To this end, the questions I hope they ask themselves are:
- What are the major risks InMobi faces and how severe might the consequences be?
- What is more important for InMobi: investing in new features, or investing in minimizing risk?
(Word Count: 798)
Sources
[1] Grazia Cecere, Fabrice Le Guiel, Vincent Lefrere, “Economics of free mobile applications: Personal data as a monetization strategy,” via Google Scholar, Accessed November 2018
[2] Ravi Krishnaswamy, (@InMobi), “’InMobi handles over 15 billion transactions per day & will grow more than four times in the coming year,’ Ravi Krishnaswamy, CTO, InMobi talks about the transformative technology roadmap he is driving at InMobi and the @Microsoft partnership with @alnoor_peermohd from @bsindia.”, Twitter, 26 July 2018, 12:10 AM, https://twitter.com/InMobi/status/1022378544238919681
[3] Artificial Intelligence in Industry, “The Future of Advertising and Machine Learning – Audience Targeting, Reach, and More,” August 2018, podcast, https://soundcloud.com/techemergence/the-future-of-advertising-and, accessed November 2018
[4] WhiteOps, “The Methbot Operation,” https://cdn2.hubspot.net/hubfs/3400937/Resources/WO_Methbot_Operation_WP.pdf, accessed November 2018.
[5] “InMobi Partners with GeoEdge to Regulate Ad Quality Standards and Improve User Experience,” InMobi press release, March 14, 2018, on InMobi website, https://www.inmobi.com/company/press/inmobi-partners-with-geoedge-to-regulate-ad-quality-standards/, accessed November 2018
[6] “InMobi Forms Strategic Partnership with Microsoft to Power New Cloud-Based Enterprise Platforms for Marketers,” InMobi press release, June 26, 2018, on InMobi website, https://www.inmobi.com/inmobi-msft-announcement, accessed November 2018
[7] “InMobi Acquires U.S. based Advertising and Data Company, Pinsight Media,” InMobi press release, October 17, 2018, on InMobi website, https://www.inmobi.com/inmobi-acquires-pinsight-media, accessed November 2018
[8] Nicholas Fandos and Kevin Roose, “Facebook Identifies an Active Political Influence Campaign Using Fake Accounts,” https://www.nytimes.com/2018/07/31/us/politics/facebook-political-campaign-midterms.html, accessed November 2018.
[9] Daisuke Wakabayashi and Sapna Maheshwari, “YouTube Advertisers Exodus Highlights Perils of Online Ads,” https://www.nytimes.com/2017/03/23/business/media/youtube-advertisers-offensive-content.html, accessed November 2018.
[10] Microsoft Technology Licensing LLC, “Automatic content tagging”, Google Patents, https://patents.google.com/patent/US20180052874A1/en, accessed November 2018
Michael,
Great piece. I would argue that given the current climate, investing in risk minimization is on par with investing in new features. Normally I would favor new features over risk minimization in a burgeoning industry, because risk tolerance is much higher when there is high growth potential and a reliable competitive edge. However, public sentiment toward companies like InMobi is turning, and with public sentiment comes government regulation that ushers in the direct targeting of companies to set an example. Instead, a potential increased investment in consumer data and advertiser’s brand protection could become a source of competitive advantage by itself. In other words, InMobi has an opportunity to brand themselves as a possible market leader in consumer and business data protection. The potential for AI to detect fraud is obviously a proven concept, so tapping into this “other” emerging market is also a convenient way to diversify InMobi’s business opportunities.