TechBytes with Shouvick Mukherjee, Chief Technology Officer, Amobee
Chief Technology Officer, Amobee
In December 2017, Amobee launched their Inventory Accountability Program to build a strong anti-fraud community. This signaled the start of active participation from adtech companies in creating a brand safe and fraud-free digital ecosystem. To understand how Amobee is building an advanced brand safety roadmap for better inventory accountability, we spoke to their Chief Technology Officer, Shouvick Mukherjee.
Tell us about your role in Amobee and the team you handle.
As Chief Technology Officer, I lead Amobee’s data science, engineering and technology teams. My responsibilities cover Amobee’s stack across our DSP, DMP, DataMine and Brand Intelligence analytics. Since I’ve been on board, we’ve worked to integrate our offerings to create a comprehensive, cohesive marketing solution for both brands and agencies. Every day, I work with internal stakeholders across the company – including sales, product management and marketing – to help build the technology, at scale, to set our business on a path for continued growth. My role is a culmination of more than 20 years of experience at Fortune 500 companies, focusing on leveraging data and omnichannel digital relationships to connect clients with their consumers.
What led Amobee to launch an advanced brand safety and technology inventory accountability program?
Brand safety and transparency are going to remain incredibly important for marketers as we move through 2018 and beyond. We’ve been working diligently with other leaders across the industry to ensure our industry-leading technology and partnerships eliminate fraud before it reaches our platform and equips clients with pre-bid brand safety, contextual targeting and protection in the fight against fraud.
What are the core technological tenets of the Amobee Inventory Accountability Program?
Amobee’s global brand safety and fraud prevention solution leverages an early detection system using advanced technology and partnership integrations to prevent fraud from reaching the Amobee platform and ensures Amobee DSP customers have access to the highest quality programmatic inventory. In addition, we are among the first DSPs to block fraudulent mobile apps as part of the platform-wide offering. Amobee is also the first DSP to offer multiple goal optimizations, coupling key performance indicators with a built-in viewability algorithm that delivers view rates up to 30 percent higher than other DSPs and minimizes the time required to adjust campaigns while they’re in-flight.
In addition to our robust brand safety offering, we’re also participating in the IAB Tech Labs ads.txt initiative to help mitigate inventory fraud concerns and allow marketers to focus on creating compelling campaigns that resonate with their target audiences. In support of the ads.txt initiative, Amobee is committed to cleaning the supply chain and blocking unauthorized and counterfeit inventory made available through non-certified sellers in order to promote a safe and transparent buying ecosystem.
Amobee’s global fraud prevention offering includes the following —
- Platform-wide fraud blocking in partnership with DoubleVerify to identify and eliminate any bid opportunities that are classified as bot, site or mobile app fraud before reaching the platform
- A patent-pending diagnostic tool built by Amobee to identify and automatically block suspicious activity created by very short-lived bots
- Post-Bid Monitoring across 90+ metrics at both high and granular levels
- Inventory targeting tools that can be applied at the individual campaign level target curated supply, including the ability to activate premium publisher deals and guaranteed buys
- Rigorous brand safety tools provide the ability to block fake, inflammatory, objectionable or off-brand content
- Anti-fraud certification programs including Trustworthy Accountability Group for both piracy and fraud
Would rigorous brand safety regulations negatively impact customer experience standards?
To the contrary, we feel that brand safety only enhances customer experience, both for our clients and the consumers they’re looking to reach. Our long-term, advanced brand safety initiative ensures we’re able to align brand and agency clients with the highest quality inventory to support campaign objectives by delivering ads that are in view and seen by a real audience in a brand safe environment.
How does your intelligence technologies complement existing customer experience platforms?
We’re working on some really interesting ideas in this space. Right now, marketing technology platforms expect the customers to deeply understand the stack; what we’re doing is giving them more choices in leveraging data, optimizing the performance of the campaign by providing a better understanding of the audiences and networks. This gives marketers more in-depth insights in real time so they can make informed decisions while a campaign is in flight or for post-flight delivery to improve their day-to-day activity in our platform. We’re thinking about and providing solutions so marketers can use this wealth of data to improve the work they’re doing each day.
How do you leverage AI/ML technologies at Amobee?
Amobee has had machine learning and artificial intelligence baked into our DNA for a long time. And our data science team keeps innovating the existing algorithms to improve the performance of our platform and has had its work recognized through the publication of white papers and appearances at technology conferences.
The Amobee Marketing Platform relies heavily on AI and ML to deliver the best value to advertisers. On the DSP side, we have developed a multitude of algorithms to improve the ROI for advertisers. Some of this goes back to our brand safety initiative, with the Amobee DSP automatically filtering the high-quality inventory for our advertisers based on campaign performance using ML models. This removes the burden from our advertisers during the campaign setup phase to manually optimize the supply path for best ROI. The algorithm uses ML models to predict inventory quality and then uses matching algorithms to allocate the right inventory to advertisers.
A fundamental challenge for the marketers during campaign setup phase is budget allocation which requires the marketer to distribute the total campaign budget across multiple line-items.
They may also need to tweak the allocations periodically to achieve their performance and reach goals. To help marketers right size budget allocation and distribute the total campaign budget across multiple line items, the Amobee DSP has the ability to automatically allocate the budget based on performance and spend potential. It also considers channel specific KPI goals to optimize delivery for each channel. This AI-based algorithm is moving budgets on a continuous basis across line-items to provide the best result for the advertiser.
In targeted display advertising, the primary goal is to identify the best opportunities to display an ad to an online user who is most likely to take a desired action such as purchasing a product or signing up for a newsletter. An optimization aware DSP should bid different amount based the predicted value from the specific impression instead of fixed CPM values. The Amobee DSP has used traditional ML algorithms to predict the value of each impression in real-time based on user features, contextual features and advertiser features. We have come up with innovative ways to deal with the extreme data-sparsity and high feature dimensionality issues.
We also dynamically explore multiple bidding strategies to deliver the best ROI for the advertiser. As marketers need becomes more complex, they expect us to optimize on multiple KPI goals, such as optimizing both viewability and action rate for a video campaign.
To handle these complex needs from advanced advertisers, the Amobee DSP has implemented generalized multi-goal optimization algorithms which has lead to significant ROI improvements for our customers. We have also started using Deep Learning methods for predicting user intent accurately from browsing pattern of users and are seeing significant improvement in action prediction.
What are your predictions for AI-driven digital branding technologies in 2018?
Some of the most important AI-driven features will revolve around helping marketers for brands and agencies do their day-to-day job smarter and better. To that end:
- We can leverage AI to ensure the data coming in and data going out is more accurate and predictions are better and more precise.
- Building algorithms that provide feedback and data in almost real-time that gives our platform the ability to get that information back to the marketers to significantly improve their campaigns in-flight quicker and more effectively.
- With the shift of advertising dollars from traditional TV to streaming services on the rise, the unique ability to understand audience behaviors through contextual AI across linear and connected TV will provide marketers deep insight into cross-platform media planning for connected TV, mobile apps, and the web.
- The emerging tech surrounding speech-enabled communications are going to be critical in personalizing the experience for the user to the platform they’re operating as well as working with voice-activated platforms to make sure digital advertising can be personalized for the consumer. AI will play a massive role in learning how to personalize the marketing for ads in real-time so marketers can better target audiences.
Thanks for chatting with us, Shouvick.
Stay tuned for more insights on marketing technologies. To participate in our Tech Bytes program, email us at firstname.lastname@example.org