Operationalizing Contextual AI in Advertising
What Industry Leaders Are Actually Doing at Scale Insights from discussion with leaders from Roku, DAZN, Philo, Intersection and Tavant at Streaming Media Connect on building real-time, AI-driven advertising systems: ▪ Why AI is now foundational to ad operations at scale ▪ How platforms align ads with tone, sentiment, and live context ▪ Where most ad tech stacks break and why ▪ The 4-layer framework powering contextual intelligence Download the Full article First Name * Last Name * Work Email * Phone Number Company * Job Title Download the Article Why This Matters Now Contextual advertising is no longer a targeting tactic, it’s becoming a must-have operational capability. Scale is challenging existing workflows Millions of creatives, real-time signals, and multi-platform delivery are overwhelming traditional systems. Ad operations are deeply fragmented Campaigns span CRM platforms, ad servers, DSPs, and reconciliation layers, creating friction at every step. Real-time decisioning is expected, but rarely operationalized Most teams still act on delayed insights instead of live signals Measurement gaps persist, but decisions can’t wait The gap between what advertisers want to know and what systems can prove remains unresolved “The question is no longer whether contextual advertising works, it’s whether your organization can execute on it and operationalize it at scale.” At scale, AI is not optional, it’s foundational The biggest constraint isn’t technology, it’s fragmented operations. AI-powered accelerators help unify workflows Context now means tone, sentiment, and real-time signals Leaders are optimizing in real time, even without perfect measurement Download the article to explore all The 4 Layers of Contextual Intelligence Learn how leading platforms structure their AI systems to process signals, enrich context, make decisions in real time, and activate across fragmented ecosystems. See how this framework works From the Experts Running AI in Production “AI isn’t giving you the answer. It’s giving you a confidence level.” Roku “The future isn’t about placing ads in a show — it’s about aligning with moments.” Philo See How Industry Leaders Are Scaling Contextual AI Download the Full Article
Can AI be the Key to Driving AVOD’s Success?
In the ever-evolving world of online video viewing, subscription-based streaming has long been the dominant force. However, a new player is emerging and gaining momentum: Advertising-Based Video on Demand (AVOD). This model attracts both new and existing subscribers by offering a low-cost or even no-cost streaming experience, supported by advertisements. AVOD platforms provide a selection of programs that are accompanied by targeted advertisements, making them an appealing choice for a wide range of viewers. Unlike traditional platforms that have witnessed a progressive decline in popularity, AVOD platforms have the advantage of reaching a large and diverse audience. By carefully curating the advertisements shown and avoiding excessive repetition, content distributors can ensure minimal viewer distraction and effectively reduce churn. Considering the rising costs and inflation levels, this approach not only adds value to the customer’s experience but also offers a cost-effective means of generating a high return on investment (ROI). According to Omidia, the future looks bright for AVOD streamers. AVOD is projected to surpass linear television and generate an estimated revenue of $259 billion by 2025. This growth further solidifies the appeal and potential of AVOD as a viable business model in the fast-expanding landscape of online video streaming. The rise of AVOD in recent years has the potential to outshine SVOD (subscription video on demand). Deloitte forecasts that by 2030, a majority of online video service subscriptions will be financially supported, either partially or entirely, by advertisements. The monetization through ads offers unparalleled profitability and fosters deeper engagement with the audience. Interestingly, a survey by TiVo revealed that customers are generally accepting of advertisements when it comes to accessing free content. Overall, the AVOD market is set to experience significant growth in the coming years, further bolstered by advancements in technology like Artificial intelligence (AI) which is poised to drive the expansion of the AVOD market. What impact does AI have on the AVOD market and business outcomes? AI-Powered Predictive Analysis for Business Expansion: AI forecasting software enables complex analysis and facilitates business planning. It provides intelligent data that helps businesses enter new geographic regions and ensures the sustainability of their business models in the long run. Through predictive analysis, content providers can identify the appropriate target audience, understand their demand and potential for growth based on title, genre, and preferences. AI enables better decision-making by offering a clearer picture of audience segmentation and their landscape. Targeted Marketing for Improved ROI: AI solutions can identify the right audience to target and provide customized suggestions based on their behavior – such as frequently watched genres and favorite titles. By offering intuitive recommendations and personalized marketing, AI enhances the customer experience based on preferences. AI-powered insights offer valuable data on customer habits, contributing to improved marketing strategies and, consequently, better business outcomes. Enhanced Content Monetization: Well-designed and marketed freemium AVOD content has the potential to attract and retain subscribers across various age groups. AI software analyzes data and standardizes datasets to compare performance across different AVOD platforms. This allows for determining the optimal solution to deploy, the ideal content types, and the optimal display timing. Compared to traditional approaches, AI-powered platforms will have the ability to drive content monetization significantly. AI for Identifying and Retaining Potential Subscribers: AI-based software can also identify users that are more likely to subscribe to the AVOD model. Subscribers who are already comfortable with AVOD platforms and their offerings are more inclined to choose a paid subscription for additional benefits. Conversely, instead of canceling an expensive subscription, customers are more likely to opt for a less expensive ad-supported tier and AI can seamlessly identify such customers and ensure long-term retention. Customer Data Analysis for Personalized Recommendations: AI further utilizes customer behavioral data based on varied touchpoints, including time spent watching a TV show, start and exit times, and advertising data. This substantially improves the viewer experience and increases the amount of time spent viewing suggested content by offering the best recommendations based on preferences and interested segments. What is the future of the AVOD market with AI as the technology engine? The future of the AVOD market, powered by AI, holds immense potential, and promises to revolutionize the media industry. As viewers increasingly turn to online platforms for content consumption, the onus is on platforms of the future like AVOD to innovate and captivate their audience. AVOD models can greatly benefit customers by reducing subscription costs or even offering free services. By leveraging AI, businesses can reshape existing content, optimize the impact of advertisements, gauge customer response, and enhance the overall viewing experience. This seamless transition to AVOD, driven by AI, has the power to disrupt the market and usher in a new era within the media industry. The stage is set for AI to play a transformative role, and the future holds exciting possibilities, as video on-demand technology continues to evolve. Reach out to us at [email protected] or visit here to learn more.
Bayesian Methods for Media Mix Modeling with Carryover and Shape Effects
As discussed in my previous blog posts, a lot of research is being done in ad attribution and media mix modeling. Today I’ll introduce another paper that provides some interesting analysis. Fair warning, you should have a basic idea of Bayesian regression before reading this. You can find a great introduction here. Carryover and Shape Effects The authors’ most exciting contribution is incorporating carryover and shape effects in their media mix model. Carryover effects try to model the impact of media spend over a future period. Since media spend influences consumers on buying a product or service, the impact of such spending doesn’t just last for the time an advertisement is aired but for a more extended period. The authors transform the time series of media spend using a decay function for accounting for such carryover effects. They use the adstock function as described below : wm is a non-negative weight function, and the media spend effect is the weighted average of media spend of the current period and previous L-1 periods. The authors introduce two types of weight functions, geometric decay (where media spend peaks when an advertisement is aired) and delayed adstock (where the impact of media spend rises sometimes after an ad is aired). A visualization describing the effects of the weight functions can be seen below, taken from the paper. Next, the authors discuss shape effects. Shape effects aim to capture diminishing returns on media spend. For example, it is valid to assume that for a specific medium, the rate at which media spends rises is dramatic from 0 to $50 but reduces significantly from $100 to $150. The authors use a Hill function to model shape effects. The discussion of Hill functions is beyond the scope of this blog, but the regression coefficients can be multiplied by the Hill function to get the following form : The hill function for media spend is a point transformation, as opposed to earlier discussed carryover effects. The following graph, taken from the paper, gives a visual representation of diminishing returns, given different parameter values in the Hill function : Both these transformations can be applied to media spend. Depending on individual use cases, one must decide which transformation to apply first. The authors apply the adstock transformation first and then the shape transformation. The final sales at time t, which can be described as y_t , can be modeled using the following equation : To simplify, this equation models sales as a function of some baseline sales τ in addition to transformed media spend effects of control variables, and random noise. Why Bayesian Regression A common question could be: Why estimate these parameters using bayesian regression? The answer lies in the fact that Bayesian regression lets us quantify the uncertainty in our predictions, and more importantly, allows us to set priors on our parameters. For example, it is valid to assume that media spend will never have a negative effect on sales, which allows us to set informative priors on media spend coefficients (constraining them to be non-negative values). The authors then explain their implementation of this model to real-world datasets. They use Gibbs sampling to sample from their model and implement this in STAN. However, multiple techniques for sampling from the posterior distribution and their code can be replicated easily using PyMC3. Please take a look at the fundamentals of Bayesian Regression if this isn’t making much sense. The parameter estimates obtained from the model can then be plugged into a linear optimization algorithm that conditions on a fixed media spend budget to find the best media mix given a set of channels. The linear optimization algorithm introduced by the author is beyond the scope of this post, but I might discuss it in my next one. Stay tuned!
How to Decipher Customer Journey with Relevant Advertisement Attribution
Every advertiser has a unique context. Are you enabling the right choices? A large multinational company is launching a new lifestyle product and is looking to gain the attention of high-income, mid-career women across tier-one cities. Their advertising campaign has unique requirements, and they want the best slots suited to their product promotion. How would you guide them to make the best choices and drive premium returns for your platform? You need deep insights on advertising performance across segments that demonstrate the optimal ROI to win the trust of the advertisers. You can promote specific segments, drive higher revenue, and make better pricing decisions based on quantifiable metrics unearthed with actionable analytics. Accuracy in advertising attribution is your key for precise audience targeting, campaign optimization and improved performance that boosts the bottom line while bolstering top-line growth. Campaigns leave important cues. Are you listening? The new product line you launched last week has caught the imagination of young shoppers. Your e-commerce site has an upsurge in traffic. You used different media to advertise and promote the product line, and it has worked. The ad spots on prime-time TV and jingles on FM radio are on for a week now. You placed inserts in newspapers with QR code for discount coupons. The redemption of those coupons is doing well too. Your social media campaign is running in parallel, and you are ready for another round of emails to roll out referral offers. You are convinced that your ad spend has delivered the desired results. It is important to trace your customer journey through all the different touchpoints up to the conversion or buying stage. Your ad spend needs to be rationalized and focused on the media mix that delivers optimum results. In short, you need to analyse ad attribution, to zero in on your campaign effectiveness. Ad attribution unlocks the significance of every touchpoint to conversion. Personalized campaigns demand an intimate understanding of customer behavior, as well as customers’ channel and platform preferences. Your campaign ROI depends on your knowledge of the customers. Advertising attribution processes allow you to trace your customers’ actions across multiple touchpoints to reveal the levels of interaction that brought them to the point of sale. This data is crucial for evaluating past campaign performance and intelligently planning the next ones for better outcomes. Advertising attribution is a quantitative measure of each touchpoint in nudging the customer journey towards conversion. Single-touch ad attribution – for example, measuring the first click or last click for a given promotion on any platform – can deliver straightforward analysis if the action is definitive, such as a discount offer for the first 50 customers within a day, advertised on Facebook. The marketer can assign success to the specific promotion. Multi-touch journeys are deciphered through various ad attribution models. A multi-touch customer journey is more difficult to attribute. For instance, a customer may have seen a newspaper insert ad, noticed a similar advertisement on a social platform, received a promotional offer through a friend, checked out the company website, and then received remarketing enforcement before making a purchase. Now, every touchpoint is a nudge forward and must be accordingly scored. This multi-touch attribution is a flexible scoring model for marketers to assign due credit to each interaction for a comprehensive performance insight. Linear models assign equal value to each touchpoint. Shapely values, such as U and W-shaped scoring models, give more credit to the first and last touchpoints and first, middle, and previous touchpoints, respectively. Some marketers prefer a time decay model that treats the touchpoints closer to purchase as more important than touchpoints at the beginning of the journey. Advanced algorithmic and statistical models leverage AI and ML for ad attribution. The complexity of advertising data requires advanced custom models to assign performance metrics to every touchpoint adequately. Data-driven and statistically evolved, these attribution models leverage AI-based algorithms to score the customer purchase decision milestones appropriately. Machine learning algorithms guide the marketers in deciphering conversion probability to plan promotions accurately. The right choice of ad attribution model drives campaign performance. The pressure on marketing budgets has created a greater need for campaign precision. Marketers must choose the right attribution model to improve campaign performance and sales lift. However, there are no absolutes in this game. No statistical model, no matter how evolved and data-rich can guarantee 100% accuracy. Marketers must consider models that align with their customer journey and campaign intricacies. An advanced AI and ML-powered analytics platform can algorithmically design attribution models to deliver timely and accurate metrics. Making the right choice for ad attribution is intrinsic to campaign success. Marketers can leverage these attributes to design just-in-time campaigns with higher confidence. Which ad attribution model would you bet on for your campaign analysis? Please share your thoughts with us at [email protected]; or to learn more about Tavant’s media solutions, click here.
Is Face Recognition Technology Shaping the Future?
Face recognition is one of the fastest growing technology these days, and there are various companies not only involved in R&D but same time developing numerous applications in different fields. There have been various biometric ways in use to recognize a person’s identity like finger-scan, hand-scan, retina-scan, and face recognition. Face recognition is not an altogether new technology, but artificial intelligence and machine learning techniques are continually making it better. It is the latest way to identify people. Face recognition is rapidly gaining momentum with many business benefits such as enhanced user experience, cost-effective without manual intervention. In some cases, people can be recognized even without his/her knowledge by taking his live photo or video. How does it work? Face recognition uses deep learning algorithms which is an advanced form of machine learning, to compare a digital image to the stored faceprint to verify an individual’s identity. Every human face has approximately 80 nodal points that are nothing but the peaks and valleys that make up the different facial features and help to distinguish individuals. These nodal points are measured for each face and create a numerical code, called a faceprint, representing the face in the database. Some of the features measured by Face Recognition Technology are: The distance between the eyes The width of the nose The shape of the cheekbones The depth of the eye sockets The length of the jawline In Face recognition application, these measurements are retained in an application database and used as a comparison for any person image needed to be identified. Use Cases: While the use case of face recognition can be endless, here are a few that are in production already and widely being used or being researched most – Human face-based Attendance Management System for any company Implemented in Tavant and being used as a pilot project Identify suspicious person at any public places or restricted places Already deployed at some of the airports/subways in Japan, UAE, and China Face recognition-based check-in for better customer experience and save time Recently Delta airlines rolled out face recognition technology at Detroit Metro Airport Unlocking mobile, PC or any personalized device For example, Apple’s iPhone X includes Face ID technology that allows users to unlock their phones with a faceprint mapped by the phone’s camera. Entertainment Industry For example, the Kinect motion gaming system leverages face recognition to differentiate among players. Targeted Advertisement Smart advertisements in airports are now able to identify the gender, ethnicity, and approximate age of a passer-by and perform targeted advertising according to the person’s demographic. Improved User Experience MasterCard, Amazon, and Alibaba have rolled out face recognition payment methods often referred to as selfie pay Technology Adoption Challenges: With every technology adoption, in addition to benefits, there are a couple of challenges as well which needed to be known before implementing any use case: Security: Your facial data can be easily gathered and likely to be stored privately or in the public domain, often without your knowledge and permission. It is possible that hackers could access and steal this data and may misuse it. Privacy: Face recognition technology is being used more widely. That means your facial data could end up available in a lot of places. You probably even would not know who has access to it. Freedoms: Government agencies and even unauthorized entity may track your personal data. It would not be easy to stay anonymous. Safety: Also, face recognition can also be misused and lead to online harassment and stalking. For example, what if someone takes your picture in public and uses face recognition software and finds out exactly who you are. Mistaken identity: Face recognition data can be prone to error, i.e., misidentify or failing to identify, even with very low probability, which can implicate people for crimes they haven’t committed. Limitation: Face recognition application cannot differentiate identical twins. Also, it will not give the desired output in dim light. The Road Ahead: Face recognition technology is still in an evolving phase and gradually being adopted in various applications ranging from social media to critical government applications. Many top players like Amazon, Google, and IBM have come up with their offerings that can be used by anyone and investing in research to improve the speed and accuracy of the process. Companies must unleash the potential of face recognition technology and its implementation applicability to gain a better competitive advantage in their business.
Things to Consider Before Replacing Google DSM
July 2019 Isn’t That Far Away Deprecation of DoubleClick Sales Manager (DSM) has caused some consternation in the media and publishing industry. After July 31, 2019, DSM users will no longer have access to the tools and data in the system, which enable publishers to manage digital channel sales. The reactions to this vary from frustration to denial. And like other deprecations, it’s something publishers must get past. For millions of affected organizations, the search for a replacement order management system has already begun, whether they are happy about it or not. Amidst the short timeline, you need to ask a few right questions while choosing a replacement for Google DSM. Below is a list of factors, a company should consider 1. Accomplishing the purpose served by Google DSM: Streamline business by providing media workflows with minimal to zero changes to the existing process a. Create proposals b. Create products and categories c. Manage rate cards 2. Automation and accuracy: Eliminate manual processes and errors by automating the business rules a. Data validations b. Calculate the metrics based on budget and product c. Dashboard for monitoring and reporting d. Define targeting rules 3. Integration with DFP: Implement a bi-directional integration with DFP for ease of managing campaigns a. Tracking the campaign status and basic reporting within the tool b. Integration should be updated with new releases of DFP APIs 4. Integration with Salesforce: Provide the ability to exchange information with CRM a. Access lead information b. Update opportunities automatically 5. Number of integrations: Create a seamless experience by connecting to in-house or third-party services a. Ad-exchanges/products: Provide a choice of ad platforms for bridging campaigns b. Data providers: Access audience segments c. Inventory Forecasting: Plan budget judiciously d. CDN: Upload assets e. Payments: Allow invoicing and acceptance of payments 6. Designed for digital ecosystem: The UX should be built keeping in mind the digital ad ops team a. Frictionless navigation b. Accessible from desktops, laptops, tablets 7. Cloud or on-premise: Can support your deployment model a. One click builds and deployment b. Easy to roll-out upgrades 8. Customers: Extent of experience with other advertising companies a. Domain expertise b. Depth and breadth in terms of technology choices available 9. Cost: The migration away from Google DSM needs to be cost-effective a. Application maintenance and enhancement b. No hidden costs 10. Customization and extensibility: Make changes based on the roadmap and additional requirements a. Add new modules b. Integrations with new services or platforms c. Modify existing workflow or UX based on the custom needs of the team 11. Value-added services: This can act as a differentiator amongst multiple available options a. Proposal templates b. Advance reporting and analytics c. Insights into the progress of media proposals d. Advance UX controls to improve operational efficiency Step Forward Don’t wait until 2019 to plan your transition. So as DSM gets ready to kick the bucket, what a perfect opportunity to upgrade to a smarter and more powerful solution. Backed by more than a decade of experience in building digital solutions for top media companies, Tavant deeply understands the specific needs of the advertising industry and have successfully delivered solutions to complex business problems ranging from media sales to advertising and reporting. Companies can now utilize Tavant’s media planning and sales manager for complete control and flexibility over your order management process.