Blog Post – Blog & Insights

Search Custom Image Libraries with New Image Similarity Models

Hive

May 4, 2022July 5, 2024

Building a Smarter Way to Search
Image Similarity Models: A Two-Pronged Approach
Why Use Similarity Models for Image Comparison?
Example Image Comparisons and Model Responses
Hive’s Custom Search: API Overview
Final Thoughts

Building a Smarter Way to Search

Hive has spent the last two years building powerful AI models served to customers via APIs. At their core, our current models – visual and text classification, logo detection, OCR, speech-to-text, and more – generate metadata that describes unstructured content. Hive customers use these “content tagging” models to unlock value across a variety of use-cases, from brand advertising analytics to automated content moderation.

While these content tagging models are powerful, some content understanding challenges require a more holistic approach. Meeting these challenges requires an AI model that not only understands a piece of content, but also sees how that content relates to a larger set of data.

Here’s an example: a dating app is looking to moderate their user profile images. Hive’s existing content tagging APIs can solve a number of challenges here, including identifying explicit content (visual moderation), verifying age (demographics), and detecting spam (OCR). But what if we also needed to detect whether or not a given photo matches (or is very similar to) another user’s profile? That problem would fall outside the scope of the current content tagging models.

To meet these broader content understanding challenges, we’re excited to launch the first of our intelligent search solutions: Custom Search, an image comparison API built on Hive’s visual similarity models. With the Custom Search APIs, platforms can maintain individualized, searchable databases of images and quickly submit query images for model-based comparisons across those sets.

This customizability opens up a wide variety of use-cases:

Detecting spam content: oftentimes, spammers on online platforms will use the same content or variants of the original content. By banning a single piece of content and using our custom search solution, platforms can now more extensively protect their users.
Detecting marketplace scams: identify potentially fraudulent listings based on photos that match or are similar to other listings
Detecting impersonation attempts: on social networks and dating apps, detect whether or not the same or similar profile images are being used across different accounts

This post will preview our visual similarity models and explore how to use Hive’s Custom Search APIs.

Image Similarity Models: A Two-Pronged Approach

More than other classification problems, the question of “image similarity” largely depends on definitions: at what point are two images considered similar or identical? To solve this, we used contrastive learning techniques to build two deep learning models with different but complementary ground-truth concepts of image similarity.

The first model is optimized to identify exact visual matches between images – in other words: would a human decide that two images are identical upon close inspection? This “exact match” model is sensitive to even subtle augmentations or visual differences, where modifications can have a substantial impact on its similarity predictions.

The second model is optimized towards identifying manipulated images, and is more specifically trained on (manual) modifications such as overlay text, cropping, rotations, filters, and juxtapositions. In other words, is the query image a manipulated copy of the original, or are they actually different images?

Why Use Similarity Models for Image Comparison?

Unlike traditional image duplicate detection approaches, Hive’s deep learning approach to image comparison builds in resilience to image modification techniques, including both manual image manipulations via image editing software and adversarial augmentations (e.g., noise, filters, and other pixel-level alterations). By training on these augmentations specifically, our models can pick up modifications that would defeat conventional image hashing checks, even if those modifications don’t result in visible changes to the image.

Each model quantifies image similarity as a normalized score between 0 and 1. As you might expect, a pair-wise similarity score of 1.0 indicates an exact match between two images, while lower scores correspond to the extent of visual differences or modifications.

Example Image Comparisons and Model Responses

To illustrate the problem and give a sense of our models’ understanding, here’s how they classify some example image pairs:

This example is close to an exact match – each image is from the same video frame. Both models predict very high similarity scores (although not an exact visual match). However, the model predictions begin to diverge when we consider manipulated images:

Horizontal flip plus filter adjustments

Recoloration plus multiple mask overlay

Layered overlay text

In these examples, the exact match model shows significantly more sensitivity to visual differences, while the broader visual similarity model (correctly) predicts that one image is a manipulated copy of the other. In this way, scores from these models can be used in distinct but complementary ways to identify matching images in your image library.

Hive’s Custom Search: API Overview

Custom Search includes three API endpoints: two for adding and removing images from individualized image libraries, and a third to submit query images for model-based comparison.

For comparison tasks, the query endpoint allows images to be submitted for comparison to the library associated with your project. When a query image is submitted, our models will compare the image to each reference image in your custom index to identify visual matches.

The Custom Search API will return a similarity score from both the exact visual match model and the visual similarity model on – like those shown in the above examples – for any matching images. Each platform can therefore adapt which of these scores to use (and at what threshold) based on their desired use-case.

Final Thoughts

We’re excited about the ways that our new Custom Search APIs will enable customers to unlock useful insights in their search applications. For Hive, this represents the start of a new generation of enterprise AI that just scratches the surface of what is possible in this space.

If you’d like to learn more about Custom Search APIs or get help designing a solution tailored to your needs, you can reach out to our sales team here or by email at sales@thehive.ai.

BACK TO ALL BLOGS

Introducing Moderation Dashboard: a streamlined interface for content moderation

Hive

April 14, 2022July 4, 2024

Over the past few years, Hive’s cloud-based APIs for moderating image, video, text, and audio content have been adopted by hundreds of content platforms, from small communities to the world’s largest and most well-known platforms like Reddit.

However, not every platform has the resources or interest in building their own software on top of Hive’s APIs to manage their internal moderation workflows. And since the need for software like this is shared by many platforms, it made sense to build a robust, accessible solution to fill the gap.

Today, we’re announcing the Moderation Dashboard, a no-code interface for your Trust & Safety team to design and execute custom-built moderation workflows on top of Hive’s best-in-class AI models. For the first time, platforms can access a full-stack, turnkey content moderation solution that’s deployable in hours and accessible via an all-in-one flexible seat-based subscription model.

We’ve spent the last month beta testing the Moderation Dashboard and have received overwhelmingly positive feedback. Here are a few highlights:

“Super simple integration”: customizable actions define how the Moderation Dashboard communicates with your platform
“Effortless enforcement”: automating moderation rules in the Moderation Dashboard UI requires zero internal development effort
“Streamlined human reviews”: granular policy enforcement settings for borderline content significantly reduced need for human intervention
“Flexible” and “Scalable”: easy to add seat licenses as your content or team needs grow, with a stable monthly fee you can plan for

We’re excited by the Moderation Dashboard’s potential to bring industry-leading moderation to more platforms that need it, and look forward to continuing to improve it with updates and new features based on your feedback.

If you want to learn more, the post below highlights how our favorite features work. You can also read additional technical documentation here.

Easily Connect Moderation Dashboard to Your Application

Moderation Dashboard connects seamlessly to your application’s APIs, allowing you to create custom enforcement actions that can be triggered on posts or users – either manually by a moderator or automatically if content matches your defined rules.

You can create actions within the Moderation Dashboard interface specifying callback URLs that tell the Dashboard API how to communicate with your platform. When an action triggers, the Moderation Dashboard will ping your callback server with the required metadata so that you can successfully execute the action on the correct user or post within your platform.

Implement Custom Content Moderation Rules

At Hive, we understand that platforms have different content policies and community guidelines. Moderation Dashboard enables you to set up custom rules according to your particular content policies in order to automatically take action on problematic content using Hive model results.

Moderation Dashboard currently supports access to both our visual moderation model and our text moderation model – you can configure which of over 50 model classes to use for moderation and at what level directly through the dashboard interface. You can easily define sets of classification conditions and specify which of your actions – such as removing a post or banning a user – to take in response, all from within the Moderation Dashboard UI.

Once configured, Moderation Dashboard can communicate directly with your platform to implement the moderation policy laid out in your rule set. The Dashboard API will automatically trigger the enforcement actions you’ve specified on any submitted content that violates these rules.

Another feature unique to Moderation Dashboard: we keep track of (anonymized) user identifiers to give you insight into high-risk users. You can design rules that account for a user’s post history to take automatic action on problematic users. For example, platforms can identify and ban users with a certain number of flagged posts in a set time period, or with a certain proportion of flagged posts relative to clean content – all according to rules you set in the interface.

Intuitive Adjustment of Model Classification Thresholds

Moderation Dashboard allows you to configure model classification thresholds directly within the interface. You can easily set confidence score cutoffs (for visual) and severity score cutoffs (for text) that tells Hive how to classify content according to your sensitivity around precision and recall.

Streamline Human Review

Hive’s API solutions were generally designed with an eye towards automated content moderation. Historically, this has required our customers to expend some internal development effort to build tools that also allow for human review. Moderation Dashboard closes this loop by allowing custom rules that route certain content to a Review Feed accessible by your human moderation team.

One workflow we expect to see frequently: automating moderation of content that our models classify as clearly harmful, while sending posts with less confident model results to human review. By limiting human review to borderline content and edge cases, platforms can significantly reduce the burden on moderators while also protecting them from viewing the worst content.

Setting Human Review Thresholds

To do this, Moderation Dashboard administrators can set custom score ranges that trigger human review for both visual and text moderation. Content scoring in these ranges will be automatically diverted to the Review Feed for human confirmation. This way, you can focus review from your moderation team on trickier cases, while leaving content that is clearly allowable and clearly harmful to your automated rules. Here’s an example rule that sends text content scored as “controversial” (severity scores of 1 or 2) to the review feed but auto-moderates the most severe cases.

Review Feed Interface for Human Moderators

When your human review rules trigger, Moderation Dashboard will route the post to the Review Feed of one of your moderators, where they can quickly visualize the post and see Hive model predictions to inform a final decision.

For each post, your moderators can select from the moderation actions you’ve set up to implement your content policy. Moderation Dashboard will then ping your callback server with the required information to execute that action, enabling your moderators to take quick action directly within the interface.

Additionally, Moderation Dashboard makes it simple for your Trust & Safety team administrators to onboard and grant review access to additional moderators. Platforms can easily scale their content moderation capabilities to keep up with growth.

Access Clear Intel on Your Content and Users

Beyond individual posts, Moderation Dashboard includes a User Feed that allows your moderators to see detailed post histories of each user that has submitted unsafe content.

Here, your moderators can access an overview of each user including their total number of posts and the proportion of those posts that triggered your moderation rules. The User Feed also shows each of that user’s posts along with corresponding moderation categories and any corresponding action taken.

Similarly, Moderation Dashboard makes quality control easy with a Content Feed that displays all posts moderated automatically or through human review. The Content Feed allows you to see your moderation rules in action, including detailed metrics on how Hive models classified each post. From here, administrators supervise human moderation teams for simple QA or further refine thresholds for automated moderation rules.

Effortless Moderation of Spam and Promotions

In addition to model classifications, Moderation Dashboard will also filter incoming text for spam entities – including URLs and personal information such as emails and phone numbers. The Spam Manager interface will aggregate all posts containing the same spam text into a single action item that can be allowed or denied with one click.

With Spam Manager, administrators can also define custom whitelists and blacklists for specific domains and URLs and then set up rules to automatically moderate spam entities in these lists. Finally, Spam Manager provides detailed histories of users that post spam entities for quick identification of bots and promotional accounts, making it easy to keep your platform free of junk content.

Final Thoughts: The Future of Content Moderation

We’re optimistic that Moderation Dashboard can help platforms of all sizes meet their obligations to keep online environments safe and inclusive. With Moderation Dashboard as a supplement to (or replacement for) internal moderation infrastructure, it’s never been easier for our customers to leverage our top-performing AI models to automate their content policies and increase efficiency of human review.

Moderation Dashboard is an exciting shift in how we deliver our AI solutions, and this is just the beginning. We’ll be quickly adding additional features and functionality based on customer feedback, so please stay tuned for future announcements.

If you’d like to learn more about Moderation Dashboard or schedule a personal demo, please feel free to contact sales@thehive.ai.

BACK TO ALL BLOGS

How AI Unlocks Better Sponsorship Measurement

Hive

April 8, 2022July 4, 2024

Dan Calpin, President of Hive, spoke at the 2022 MIT Sloan Sports Analytics conference. Watch Dan’s presentation for insights on how Hive’s AI powers more scalable and comprehensive sponsorship measurement and branded content intelligence, enabling brands to more fully capture the value of their investments and rights holders to better price their assets.

Presentation: How AI-Powered Measurement Can Increase the Value of Your Sponsorships by 30% or More

For more sponsorship measurement insights, check out our Super Bowl LVI brand exposure insights and 2022 March Madness sponsorship analysis.

This analysis leverages Mensio, Hive’s media solution. Mensio uses AI to power faster and more granular sponsorship measurement and branded content intelligence across platforms.

BACK TO ALL BLOGS

OCR Moderation with Hive: New Approaches to Online Content Moderation

Hive

April 8, 2022July 4, 2024

Recently, image-based content featuring embedded text – such as memes, captioned images and GIFs, and screenshots of text – have exploded in popularity across many social platforms. These types of content can present unique challenges for automated moderation tools. Not only does embedded text need to be detected and ordered accurately, it also must be analyzed with contextual awareness and attention to semantic nuance.

Emojis have historically been another obstacle for automated moderation. Thanks to native support across many devices and platforms, these characters have evolved into a new online lexicon for accentuating or replacing text. Many emojis have also developed connotations that are well-understood by humans but not directly related to the image itself, which can make it difficult for automated solutions to identify harmful or inappropriate text content.

To help platforms tackle these challenges, Hive offers optical character recognition (OCR)-based moderation as part of our content moderation suite. Our OCR models are optimized for the types of digitally-generated content that commonly appears on social platforms, enabling robust AI moderation on content forms that are widespread yet overlooked by other solutions. Our OCR moderation API combines competitive text detection and transcription capabilities with our best-in-class text moderation model (including emoji support) into a single response, making it easy for platforms to take real-time enforcement actions across these popular content formats.

OCR Model for Text Recognition

Effective OCR moderation starts with training for accurate text detection and extraction. Hive’s OCR model is trained on a large, proprietary set of examples that optimizes for how text commonly appears within user-generated digital content. Hive has the largest distributed workforce for data labeling in the world, and we leaned on this capability to provide tens of millions of human annotations on these examples to build our model’s understanding.

We recently conducted a head-to-head comparison of our OCR model against top public cloud solutions using a custom evaluation dataset sourced from social platforms. We were particularly interested in test examples that featured digitally-generated text – such as memes and captioned images – to capture how content commonly appears on social platforms and selected evaluation data accordingly.

In this evaluation, we looked at end-to-end text recognition, which includes both text detection and text transcription. Here, Hive’s OCR model outperformed or was competitive with other models on both exact transcription and transcription allowing character-level errors. At 90% recall, Hive’s OCR model achieved a precision of 98%, while public cloud models ranged from ~88% to 97%, implying a similar or lower end-to-end error rate.

OCR Moderation: Language Support

We recognize that many platforms’ moderation needs extend beyond English-speaking users. Hive’s OCR model supports text recognition and transcription for many widely spoken languages with comparable performance, many of which are also supported by our text moderation solutions. Here’s an overview of our current language support:

Language	OCR Support?	Text Moderation Support?
English	Yes	Yes (Model)
Spanish	Yes	Yes (Model)
French	Yes	Yes (Model)
German	Yes	Yes (Model)
Mandarin	Yes	Yes (Pattern Match)
Russian	Yes	Yes (Pattern Match)
Portuguese	Yes	Yes (Model)
Arabic	Yes	Yes (Model)
Korean	Yes	Yes (Pattern Match)
Japanese	Yes	Yes (Pattern Match)
Hindi	Yes	Yes (Model)
Italian	Yes	Yes (Pattern Match)

Moderation of Detected Text

Hive’s OCR moderation solution goes beyond producing a transcript – we then apply our best-in-class text moderation model to understand the meaning of that speech in context (including any detected emojis). Our backend will automatically feed text detected in an image as an input to our text moderation model, making our model classifications on image-based text accessible with a single API call. Our text model is generally robust to misspellings and character substitutions, enabling high classification accuracies on text extracted via OCR even if errors occur in transcription.

Hive’s text moderation model can classify extracted text across several sensitive or inappropriate categories, including sexuality, threats or descriptions of violence, bullying, and racism.

Another critical use-case is moderating spam and doxxing: OCR moderation will quickly and accurately flag images containing emails, phone numbers, addresses and other personal identifiable information. Finally, our text moderation model can also identify promotions such as soliciting services, asking for shares and follows, soliciting donations, or links to external sites. This gives platforms new tools to curate user experience and remove junk content.

We understand that verbal communication is rarely black and white – context and linguistic nuance can have profound effects on how meaning and intent of words are perceived. To help navigate these gray areas, our text model responses supplement classifications with a score from benign (score = 0) to severe (score = 3), which can be used to adapt any necessary moderation actions to platforms’ individual needs and sensitivities. You can read more about our text models in previous blog posts or in our documentation.

Our currently supported moderation classes in each language are as follows:

Language	Classes
English	Sexual, Hate, Violence, Bullying
Spanish	Sexual, Hate
Portuguese	Sexual, Hate
French	Sexual
German	Sexual
Hindi	Sexual
Arabic	Sexual

Emoji Classification for Text Moderation

Emoji recognition is a unique feature of Hive’s OCR moderation model that opens up new possibilities for identifying harmful or harassing text-based content. Emojis can be particularly useful in moderation contexts because they can subtly (or not-so-subtly) alter how accompanying text is interpreted by the reader. Text that is otherwise innocuous can easily become inappropriate when accompanied by a particular emoji and vice-versa.

Hive OCR is able to detect and classify any emojis supported by Apple, Samsung, or Google devices. Our OCR model currently achieves a weighted accuracy of over 97% when classifying emojis. This enables our text moderation model to account for contextual meaning and connotations of emojis used in input text.

To get a sense of our model’s understanding, let’s take a look at some examples of how use of emojis (or inclusion of text around emojis) changes our model predictions to align with human understanding. Each of these examples is from a real classification task submitted to our latest model release.

Here’s a basic example of how adding an emoji changes our model response from classifying as clean to classifying as sensitive. Our models understand not only the verbal concept represented by the emoji, but what the emoji means semantically based on where it is located in the text. In this case, the bullying connotation of the “garbage” or “trash” emoji would be completely missed by an analysis of the text alone.

Our model is similarly sensitive to changes in semantic meaning caused by substitutions of emojis for text.

In this case, our model catches the sexual connotation added by the eggplant emoji in place of the word “eggplant.” Again, the text alone without an emoji – “lemme see that !” – is completely clean.

In addition to understanding how emojis can alter the meaning of text, our model is also sensitive to how text can change implications of emojis themselves.

Here, adding the phrase “hey hotty” transforms an emoji usually used innocuously into a message with suggestive intent, and our model prediction changes accordingly.

Finally, Hive’s OCR and text moderation models are trained to differentiate between each skin tone option for emojis in the “People” category and understand their implications in the context of accompanying text. We are currently exploring how the ability to differentiate between light and darker skin tones can enable new tools to identify hateful, racist, or exclusionary text content.

OCR Moderation: Final Thoughts

User preferences for online communication are constantly evolving in both medium and content, which can make it challenging for platforms to keep up with abusive users. Hive prides itself on identifying blindspots in existing moderation tools and developing robust AI solutions using high-quality training data tailored to these use-cases. We hope that this post has showcased what’s possible with our OCR moderation capabilities and given some insight into our future directions.

Feel free to contact sales@thehive.ai if you are interested in adding OCR capabilities to your moderation suite, and please stay tuned as we announce new features and updates!

BACK TO ALL BLOGS

New and Improved AI Models for Audio Moderation

Hive

April 6, 2022July 4, 2024

Live streaming, online voice chat, and teleconferencing have all exploded in popularity in recent years. A wider variety of appealing content, shifting user preferences, and unique pressures of the coronavirus pandemic have all been major drivers of this growth. Daily consumption of video and audio content has steadily increased year-over-year, with a recent survey indicating that a whopping 90% of young people watch video content daily across a variety of platforms.

As the popularity of user-generated audio and video increases, so too does the difficulty of moderating this content efficiently and effectively. While images and text can usually be analyzed and acted on quickly by human moderators, audio/video content – whether live or pre-recorded – is lengthy and linear, requiring significantly more review time for human moderation teams.

Platforms owe it to their users to provide a safe and inclusive online environment. Unfortunately, the difficulties of moderating audio and video – in addition to the sheer volume of content – have led to passive moderation approaches that rely on after-the-fact user reporting.

At Hive, we offer access to robust AI audio moderation models to help platforms meet these challenges at scale. With Hive APIs, platforms can access nuanced model classifications of their audio content in near-real time, allowing them to automate enforcement actions or quickly pass flagged content to human moderators for review. By automating audio moderation, platforms can cast a wider net when analyzing their content and take action more quickly to protect their users.

How Hive Can Help: Speech Moderation

We built our audio solutions to identify harmful or inappropriate speech with attention to context and linguistic subtleties. By natively combining real-time speech-to-text transcription with our best-in-class text moderation model, Hive’s audio moderation API makes our model classifications and a full transcript of any detected speech available with a single API call. Our API can also analyze audio clips sampled from live content and produce results in 10 seconds or less, providing real-time content intelligence that lets platforms act quickly.

Speech Transcription

Effective speech moderation needs to start with effective speech transcription, and we’ve been working hard to improve our transcription performance. Our transcription model is trained on moderation-relevant domains such as video game streams, game lobbies, and argumentative conversations.

In a recent head-to-head comparison, Hive’s transcription model outperformed or was competitive with top public cloud providers on several publicly available datasets (the evaluation data for each set was withheld from training).

Each evaluation dataset consisted of about 10 hours of recorded English speech with varying accents and audio quality. As shown, Hive’s transcription model achieved lower word error rates than top public cloud models. This measures the ratio of incorrect words, missed words, and inserted words to the total number of words in the reference, implying Hive’s accuracy was 10-20% higher than competing solutions.

Audio Moderation

Hive’s audio moderation tools go beyond producing a transcript – we then apply our best-in-class text moderation model to understand the meaning of that speech in context. Here, Hive’s advantage starts with our data. We operate the largest distributed data-labeling workforce in the world, with over five million Hive annotators providing accurate and consensus-driven training labels on diverse example text sourced from relevant domains. For our text models, we leaned on this capability to produce a vast, proprietary training set with millions of examples annotated with human classifications.

Our models classify speech across five main moderation categories: sexual content, bullying, hate speech, violence, and spam. With ample training data at our disposal, our models achieve high accuracy in identifying these types of sensitive speech, especially at the most severe level. Our hate speech model, for example, achieved a balanced accuracy of ~95% in identifying the most severe cases with a 3% false positive rate (based on a recent evaluation using our validation data).

Thoughtfully-chosen and accurately labeled training data is only part of our solution here. We also designed our verbal models to provide multi-leveled classifications in each moderation category. Specifically, our model will return a severity score ranging from 0 to 3 (most severe) in each major moderation class based on its understanding of full sentences and phrases in context. This gives our customers more granular intelligence on their audio content and the ability to tailor moderation actions to community guidelines and user expectations. Alternatively, borderline/controversial cases can be quickly routed to human moderators for review.

In addition to model classifications, our model response object includes a punctuated transcript with confidence scores for each word to allow more insight into your content and enable quicker review by human moderators if desired.

Language Support

We recognize that many platforms’ moderation needs extend beyond English-speaking users. At the time of writing, we support audio moderation for English, Spanish, Portuguese, French, German, Hindi, and Arabic. We train each model separately with an eye towards capturing subtleties that vary across cultures and regions. Our currently supported moderation classes in each language are as follows:

We frequently update our models to add support for our moderation classes in each language, and are currently working to add more support for these and other widely spoken languages.

Beyond Words: Sound Classification

Hive’s audio moderation model also offers the unique ability to detect and classify undesirable sounds. This opens up new insights into audio content that may not be captured by speech transcription alone. For example, our audio model can detect explicit or inappropriate noises, shouting, and repetitive or abrasive noises to enable new modalities for audio filtering and moderation. We hope that these sound classifications can help platforms identify toxic behaviors beyond bad speech and take action to improve user experience.

Final Thoughts: Audio Moderation

Hive audio moderation makes it simple to access accurate, real-time intelligence on audio and video content and take informed moderation actions to enforce community guidelines. Our solution is nimble and scalable, helping platforms of all sizes grow with peace of mind. We believe our tools can have a significant impact in curbing toxic or abusive behavior online and lead to better experiences for users.

At Hive, we pride ourselves on continuous improvement. We are frequently optimizing and adding features to our models to increase their understanding and cover more use cases based on client input. We’d love to hear any feedback or suggestions you may have, and please stay tuned for updates!

BACK TO ALL BLOGS

2022 March Madness: Sponsors Generate $410M+ in Media Value

5 brands - AT&T;, Capital One, Coca-Cola (including Powerade), Nike, and Spalding - collectively earned more than half of the total media value generated during the 2022 Men’s and Women’s tournaments

Hive

April 5, 2022July 5, 2024

At a glance:

2022 March Madness generated more than $410M in equivalent media value for brands exposed within telecasts of the Men’s and Women’s tournaments (excluding commercials)
Exposure was concentrated within official NCAA sponsors, equipment providers, and the apparel brands outfitting participating teams, which collectively earned 93% of equivalent media value
Across the Men’s and Women’s tournaments, Capital One earned the most brand exposure among NCAA Corporate Champions, as did Buick among NCAA Corporate Partners, Nike among uniform sponsors, Spalding among equipment providers, and State Farm among all other brands
Among uniform sponsors, Nike outfitted more than half of all teams across the Men’s and Women’s tournament (69 of 136 participating teams) and earned the most cumulative exposure among uniform sponsors (61% in the Men’s tournament and 48% in the Women’s)
Official NCAA corporate partners – especially Coca-Cola – benefited from a surge in search engagement during tournament games, according to data from EDO which connects televised brand exposure from sponsorships and advertisements to online search activity

The 2022 NCAA Division I Men’s and Women’s Basketball Tournaments are now in the books after 134 games, millions of busted brackets, and two new national champions. This year, Hive and Elevate Sports Ventures teamed up to determine the final scores for the brands who invested alongside March Madness.

The following insights were generated using Hive’s AI-powered media intelligence platform, Mensio, which provides always-on measurement of in-content brand exposure for more than 7,000 brands across 24/7 programming from national TV channels and regional sports networks. Mensio is trusted by a diverse set of leading brands, rights holders, and agencies to measure the value of and share of voice from sponsorship activations, product placement, and other in-content exposures.

Official NCAA partners lead in-game brand exposure

Through 134 total tournament games – 67 from each of the Men’s and Women’s tournaments – brands earned more than 187 hours of cumulative brand exposure and more than $410 million in equivalent media value from their in-content exposure (excluding traditional commercials).

“Given the tradition and extraordinary momentum behind March Madness, and the mounting attention by brands on collegiate athletes and athletics, the university sports ecosystem is primed for mature brand exposure analysis of this nature,” said Kyle Folts, Vice President, Elevate Sports Ventures, Insights. “At Elevate Sports Ventures, we believe measuring exposure at scale empowers sponsors to efficiently and effectively make data-driven decisions that optimize their partnerships.”

With significant NCAA branding and deliberate assets for sponsorship placements, in-game brand exposure was concentrated within two tiers of official NCAA sponsors, equipment providers, and uniform sponsors (see Figure 1).

Official NCAA sponsors, equipment providers, and uniform sponsors collectively earned 93% of total time on screen and 94% of total equivalent media value in the Men’s tournament, and 78% of total time on screen and 66% of total equivalent media value in the Women’s tournament. The difference in mix between tournaments was driven by additional sponsorship assets available in the First and Second Rounds of the Women’s tournament; during those games, State Farm’s logo was placed on the stanchion arm along with a collection of other brands which were visible on the pole pads at the base of the stanchion and varied by arena.

Figure 1.

AT&T, Capital One, Coca-Cola, Nike, and Spalding among March Madness exposure winners

The three NCAA Corporate Champion brands – AT&T, Capital One, and Coca-Cola (including Powerade) – along with top earners Nike and Spalding collectively earned 50% of the total televised screen time and 56% of the total equivalent media value earned by brands across the 2022 Men’s and Women’s tournaments.

Across the 2022 Men’s and Women’s tournaments, Capital One earned the most cumulative brand exposure among NCAA Corporate Champions. Buick was the most exposed brand among NCAA Corporate Partners, while Nike led the uniform sponsors, Spalding the equipment providers, and State Farm led all other brands (see Figure 2).

While in-stadium exposure is carefully scripted based on a brand’s contractual terms, the natural variability of gameplay and broadcast coverage can result in brands with similar assets earning different values from their placements.

“Most digital signage in arenas is allocated to brands for a fixed duration. However, brands get the most value from the subset of that exposure which is visible to the larger audience watching the event at home, which can often vary across brands based on gameplay,” said Dan Calpin, President of Hive – Enterprise AI. “The ability to measure this exposure in near real-time, especially during a season or multi-week event like March Madness, creates an opportunity to better align exposure with where brands get value.”

Figure 2.

Nike leads exposure among apparel brands; Air Jordan outperforms in both tournaments

While most NCAA sponsors enjoy exclusivity among their competitive set during the tournament, apparel brands are unique in that uniform sponsorships are contracted with teams – resulting in a competition for exposure among Nike, Under Armour, Adidas, and Nike’s Air Jordan brand.

More than half of the teams in the tournament (69 of 136) wore Nike uniforms; however, the exposure earned by apparel brands are a combination of both how many of their teams make the tournament, and how many games those teams ultimately play.

In the Men’s tournament, Under Armour outfitted the second-most teams (14; behind Nike’s 35) but saw only one of those teams advance to the Sweet Sixteen and none beyond that. Meanwhile, Air Jordan only outfitted six Men’s teams but strong performance resulted in the brand representing 25% of the field in the Sweet Sixteen, Elite Eight, and Final Four, as well as one half of the National Championship game (see Figure 3).

Figure 3.

In the Women’s tournament, the distribution of games was identical to the distribution of teams. Nike outfitted 50% of the teams, and those teams played in 50% of all games (see Figure 4).

image

Official NCAA Sponsors earned higher search engagement during telecasts

NCAA sponsors engaged viewers beyond traditional ads and in-game signage. According to data from data, measurement, and analytics software company EDO, top sponsors experienced a higher than average Search Engagement Rate (SER) in the minutes adjacent to their on-screen exposure. SER is EDO’s proprietary metric based on the increase in online search activity for a brand or product in the minutes immediately following a televised exposure (controlling for impressions, duration, and other factors).

SER for the three NCAA Corporate Champions – AT&T, Capital One, and Coca-Cola – was 1.69x that of the Primetime Broadcast and Cable average, meaning the trio generated 69% more search engagement than average (see Figure 5). Across the board, March Madness advertisers experienced 26% more search than the average Primetime Broadcast and Cable program. Of particular note is Coca-Cola, generating 126% more online search than average, likely due to their “Coca-Cola® with Coffee” launch campaign.

“NCAA programming typically performs exceptionally well in EDO data,” said Laura Grover, Head of Client Solutions at EDO. “In 2021, for example, NCAA March Madness programming comprised four of the top ten most engaging sports programs across all of TV. Further, the Men’s Championship Game was the second strongest sports environment for driving ad engagement in 2021, trailing only Super Bowl LV.” Grover continued, “This year’s games provided a similarly engaging environment, and Coca-Cola experienced considerable success pairing the NCAA environment with the launch of Coca-Cola with Coffee. That campaign has proven to be Coca-Cola’s strongest of the past year.”

Figure 4.

About Hive

Hive is the leading provider of cloud-based AI solutions for content understanding, which are trusted by hundreds of the world’s largest and most innovative organizations. The company empowers developers with a portfolio of best-in-class, pre-trained AI models, serving billions of customer API requests every month. Hive also offers turnkey software powered by proprietary AI models and datasets, enabling industry-leading applications for critical business needs. Collectively, Hive’s solutions are transforming legacy approaches to content moderation, brand protection, sponsorship measurement, context-based ad targeting, and more. For more information, visit thehive.ai or follow on LinkedIn.

About Elevate Sports Ventures

Elevate Sports Ventures is a best-in-class sports and entertainment consulting firm, providing proven, innovative solutions to organizations across the global sports and entertainment landscape. Elevate taps into the extensive resources, relationships, and expertise of its partners to innovate and execute comprehensive strategies and solutions in Venue Renovations, Sales and Marketing, Stadium Licenses, Premium Ticketing, Corporate Hospitality, Customer Research, Strategy and Analytics, Sales Training, and more. Formed in partnership between the San Francisco 49ers and Harris Blitzer Sports & Entertainment (HBSE) in 2018, Elevate welcomed Oak View Group (OVG), Ticketmaster and Live Nation as partners in June, 2018. For more information, visit: www.ElevateSportsVentures.com or follow @ElevateSV on Twitter or LinkedIn.

BACK TO ALL BLOGS

2022 March Madness: Week 1 Sponsorship Analysis

The first of a two-part analysis from Hive and Elevate Sports Ventures unveils which brands were early winners from their televised March Madness exposure

Hive

March 23, 2022September 12, 2024

At a glance

Through the first week of games, 2022 March Madness generated more than $165M in equivalent media value for brands exposed within telecasts of the Men’s and Women’s tournaments (excluding commercials)
The NCAA tournament offers fewer opportunities for sponsorship placements, resulting in roughly 15% less total screen time for brands per game compared to this year’s conference tournaments; exposure has been deliberately concentrated within official NCAA sponsors, as well as the apparel brands outfitting participating teams
Across the Men’s and Women’s tournament, Spalding earned the most time on screen during the first week of games due to prominent placement on the basket stanchion arms
Among uniform sponsors, Nike outfitted more than half of all teams across the Men’s and Women’s tournament (69 of 134 participating teams)
Nike’s Air Jordan will likely over-perform on uniform exposure in the Men’s tournament going forward, as the brand outfits 9% of all Men’s tournament teams but represents 25% of the Sweet Sixteen contenders
40 national advertisers placed ads across both the Men’s and Women’s tournaments through Sunday, March 20th; 22 additional brands exclusively aired commercials during the Men’s tournament and 12 additional brands exclusively aired commercials during the Women’s tournament

While there are still two weeks left of March Madness, 104 of the 134 games in the 2022 NCAA Division I Men’s and Women’s Basketball Tournaments are now in the books. As one of the most watched sporting events on the television calendar, March Madness has already generated tremendous exposure for the brands associated with it.

This is the first of a two-part analysis of March Madness 2022, completed in collaboration between Hive and Elevate Sports Ventures. For a full analysis of brand exposure within and around the Men’s and Women’s tournaments to be released on Tuesday, April 5th following the Men’s Championship Game, sign up for our media insights newsletter here.

The following insights were generated using Hive’s AI-powered media intelligence platform, Mensio, which provides always-on measurement of traditional advertising and in-content brand exposure for more than 7,000 brands across 24/7 programming from national TV channels and regional sports networks. Mensio is trusted by a diverse set of leading brands, rights holders, and agencies to measure the value of and share of voice from sponsorship activations, product placement, and other in-content exposures as well as traditional advertising.

Official NCAA partners lead in-game brand exposure

March Madness is an exclusive environment for brands, featuring fewer opportunities for in-game brand exposure compared to other sporting events. During the first week of games (through Sunday, March 20th, excluding the final 8 games of the Women’s Second Round played on Monday, March 21st which concluded after the press deadline), March Madness generated an average of 82 minutes of televised in-game brand exposure per game, excluding conference, team, and network brands. That average was roughly 15% less than the average total minutes of brand exposure across all nationally-televised Men’s and Women’s conference tournaments this year (96 minutes).

With significant NCAA branding and deliberate assets for sponsorships, in-game brand exposure was concentrated within different tiers of official NCAA sponsors and uniform sponsors (see Figure 1).

While the media value generated by exposure in Men’s games is significantly higher than that in Women’s games due to relative viewership levels and the resulting commercial spot costs, the average minutes of total brand exposure during games in the Women’s tournament, 89 minutes, was almost 20% greater than that of games in the Men’s tournament. The primary drivers of this difference were the additional sponsorship assets available on the basket stanchion in the Women’s tournament. In addition to Spalding, which was present on the stanchion across both tournaments, State Farm was a mainstay on the stanchion arm in the Women’s tournament along with a collection of other brands which were visible on the base of the stanchion and varied by arena.

Figure 1

The most prominent asset in both tournaments to date was the front of the basket stanchion arm, which featured Spalding’s logo in the Men’s tournament and State Farm’s in the Women’s tournament. Spalding, which was also visible on the stanchion in the Women’s tournament, was the most exposed brand during the first week of games, accumulating more than 25 hours of screen time within games through Sunday, March 20th. Other highly exposed brands included the Official NCAA Corporate Champions – AT&T, Capital One, and Coca-Cola – and uniform sponsors including Nike, Under Armour, and Adidas (see Figure 2).

Figure 2

Nike leads exposure among apparel brands to date; Air Jordan expected to outperform going forward

While most NCAA sponsors enjoy exclusivity within their category during the tournament, apparel brands are unique in that uniform sponsorships are contracted with teams – resulting in a competition for exposure among Nike, Under Armour, Adidas, and Nike’s Air Jordan brand.

Entering the tournament, 35 of the 68 Men’s teams (51%) wore Nike uniforms, 14 (21%) wore Under Armour, 13 (19%) wore Adidas, and 6 (9%) wore Air Jordan. 34 of the 68 Women’s teams (50%) wore Nike uniforms, 14 (21%) wore Under Armour, 15 (22%) wore Adidas, and 5 (7%) wore Air Jordan.

Adidas’ opportunities for exposure in the Men’s Tournament were quickly lessened with four of its teams losing in the First Four matchups, and five of the remaining nine losing in the First Round. Under Armour lost nine of its 14 sponsored teams in the Men’s First Round. Teams outfitted by Nike vaulted to almost 60% of active teams in the Second Round to earn the largest share of voice among apparel brands. Nike’s Air Jordan brand, however, is poised to show the greatest outperformance going forward. While Air Jordan outfitted only six of the 68 teams in the Men’s tournament, the brand’s four remaining teams make up 25% of the Men’s Sweet Sixteen field (see Figure 3).

IMAGE

While the brands that appeared within the games themselves across the Men’s and Women’s tournaments were generally consistent, there was more variance in the brands which chose to buy traditional commercials alongside the respective tournaments. 40 brands aired national commercials within both Men’s and Women’s games (see Figure 4; brands with 25 or more nationally televised commercials across tournaments through Sunday, March 20th). These included the Official NCAA Corporate Champions and Partners, as well as brands such as GEICO, Gatorade, State Farm, and Progressive, which aired the most commercials during the first week of the tournaments.

Different brands advertise in Men’s and Women’s tournaments

22 brands had national airings unique to the Men’s tournament, including brands such as Lowe’s, GMC (although parent company General Motors advertised across tournaments with other brands), Corona, and Samsung.

12 brands had national airings unique to the Women’s tournament, including brands such as McDonald’s, Dodge, Skittles, USAA, and Walmart.

The differences in advertiser mix likely reflect varied consumer targeting strategies, as well as brands’ broader relationships with the respective broadcasters (WarnerMedia and CBS for the Men’s tournament; Disney for the Women’s tournament).

IMAGE

The remaining two weeks of both tournaments will not only determine the National Champions, but also the final score for all of the brands who have become a part of the tournament. A second part of this analysis of brand exposure within and around the Men’s and Women’s tournaments will be released on Tuesday, April 5th following the Men’s Championship Game; sign up for our media insights newsletter here to be alerted when the piece is published.

About Hive

Hive is the leading provider of cloud-based AI solutions that unlock an increased understanding of video, image, audio, and text content. The company empowers developers with a portfolio of best-in-class, pre-trained AI models, serving billions of customer API requests every month. Hive also offers turnkey software powered by proprietary AI models and datasets, enabling industry-leading applications for critical business needs. Collectively, Hive’s solutions are transforming legacy approaches to content moderation, brand protection, sponsorship measurement, context-based ad targeting, and more. For more information, visit thehive.ai or follow on LinkedIn.

About Elevate Sports Ventures

BACK TO ALL BLOGS

Elevate Sports Ventures and Hive: Super Bowl LVI Telecast Generates $170 Million in Equivalent Media Value for In-Game Sponsors

Top sponsors are expected to receive an additional 3.5 to 4.5 minutes of televised screen time from Super Bowl-related news and highlights per minute of in-game exposure earned

Hive

February 14, 2022July 4, 2024

At a glance

While commercials typically dominate water cooler conversation following the big game, brand exposure within the Super Bowl telecast can earn league, broadcast, and stadium naming rights sponsors as much, and in some cases more, visibility.
According to analysis by Elevate Sports Ventures and Hive, in-game exposure translated to $170 million in Equivalent Media Value earned by brand sponsors.
Excluding commercials, more than 75 minutes of cumulative in-game brand exposure was earned by brands during the Super Bowl LVI telecast, and 19 brands earned more than 10 seconds of screen time.
The cumulative screen time of in-game brand exposures was down 28% compared to last year’s Super Bowl. The reduced exposure was primarily driven by less camera-visible in-stadium signage, most notably including branded tarps covering the lower seats of the stadium during last year’s game which had limited attendance due to the pandemic.
According to analysis using Hive’s logo detection and brand mentions models, Nike was the most visually exposed brand with more than 46 minutes of time on screen, while Pepsi received the most verbal mentions during the telecast with 11.
The value from in-game exposure will be amplified across TV from Super Bowl-related coverage in news and highlights; based on analysis of last year’s Super Bowl, top sponsors should expect to receive an additional 3.5 to 4.5 minutes of televised logo exposure for every 1 minute of in-game exposure earned.
Historical analysis suggests that SoFi, which holds the host stadium’s naming rights, will likely receive the most televised brand amplification relative to the brand’s in-game exposure, led by an outsized share of coverage on news and entertainment programming likely to film outside of the stadium.

As is the case every year, the Super Bowl is not just the pinnacle of the NFL season but also the tentpole event for brands looking to capture the attention of fans in and around the game. On the field there was only one winner on Sunday but, off the field, a host of brands will claim victory from their roles within TV’s biggest night.

While commercials typically dominate water cooler conversations among viewers, brands know not to overlook the value earned from brand exposure generated within the telecast itself. With 30-second spot costs for Super Bowl LVI reported to be as high as $7 million, the value generated from in-game brand exposure can be massive. Elevate Sports Ventures, a best-in-class sports and entertainment consulting firm, and Hive, a leading provider of cloud-based AI solutions, teamed up to analyze in-content brand exposure within and around Super Bowl LVI.

The following next-day insights were generated using Hive’s AI-powered media intelligence platform, Mensio, which provides always-on measurement of in-content brand exposure for more than 7,000 brands across 24/7 programming from national TV channels and regional sports networks. Mensio is trusted by a diverse set of leading brands, rights holders, and agencies to measure the value of and share of voice from sponsorship activations, product placement, and other in-content exposures.

Brands earn $170 million in equivalent media value from in-game exposure…

Excluding commercials as well as the official pre-game and post-game shows, more than 75 minutes of cumulative in-game brand exposure was earned by brands during the Super Bowl LVI telecast, and 19 brands earned more than 10 seconds of identifiable screen time. Coupled with the value from verbal mentions within the telecast, this translated to $170 million in equivalent media value, according to Mensio’s proprietary valuation methodology.

The total value earned by brands was roughly in-line with the $169 million earned from in-game brand exposure in last year’s Super Bowl but was generated with 28% less cumulative in-game screen time for brands compared to last year’s Super Bowl. The reduced exposure was primarily driven by less camera-visible in-stadium signage, most notably including branded tarps covering the lower seats of the stadium during last year’s game which had limited attendance due to the pandemic.

Predictably, a subset of top league, broadcast, and stadium naming rights sponsors dominated the in-game share of voice (see Figure 1).

Figure 1 – Cumulative Time on Screen Within Super Bowl LVI Telecast (Excluding Commercials)

Nike, the league’s on-field apparel sponsor, led the pack with a staggering 46 minutes and 37 seconds of cumulative screen time from TV-visible brand exposure from swooshes on jerseys and cleats.

Two of the NFL’s official sideline sponsors – Gatorade and Bose – were the next most exposed brands in the telecast, earning more than 8 and 5-and-a-half minutes of in-game brand exposure, respectively.

Pepsi again headlined the star-studded Super Bowl LVI Halftime Show, of which related in-game references contributed most of the brand’s 3 minutes and 49 seconds of visual exposure within the game, along with some in-stadium signage on the stadium’s second level.

Broadcaster NBC provided the most opportunities for in-game exposure, with 15 brands being exposed through digital billboards and set signage – in addition to a broader set of brands featured in the official pre-game and post-game shows. Toyota, which sponsored the network’s halftime report, led the group with almost 2 minutes of in-content exposure within the game.

Stadium naming rights sponsor, SoFi, made headlines for reportedly paying more than $30 million in fees annually as part of a 20-year naming rights deal. The brand ranked 8th in total visual exposure and 3rd in verbal mentions during Super Bowl LVI, with whistle-to-whistle exposure within last night’s telecast alone worth more than $3.5 million in equivalent media value, based on Mensio’s valuation methodology. However, the brand received noticeably less identifiable exposure than last year’s stadium naming rights holder, Raymond James, which earned roughly three times as much exposure in the 2021 game.

While season-long league sponsors led the pack in visual exposure in-game, the leaderboard for verbal mentions told a different story (see Figure 2). Halftime show sponsor Pepsi, NBC’s halftime report sponsor Toyota, and stadium naming rights holder SoFi captured more than half of all brand mentions, with 11, 6, and 5 whistle-to-whistle mentions during the telecast, respectively, excluding commercials and promotional units for the halftime show.

Figure 2 – Visual vs. Verbal In-Game Brand Exposures Within Super Bowl LVI Telecast (Feb 13, 2022; Excluding Commercials)

While we can close the book on brand exposure within last night’s official telecast, the value of media exposure earned by featured brands will continue to accumulate in the days ahead as Super Bowl LVI remains topical in content across television and social media.

Using Hive’s Mensio platform, which provides always-on measurement of brand exposure across 24/7 television programming, a comprehensive view of the incremental value from televised brand exposure can be understood. Based on analysis of last year’s Super Bowl LV, top sponsors should expect to earn an additional 3.5 to 4.5 minutes of televised logo exposure next-day for every 1 minute of in-game exposure (see Figure 3). Although those equivalent ad units are not valued at the same spot cost as the game itself, they produce a meaningful amplification of brand exposure beyond the whistle-to-whistle measurement.

“The amplification from in-game exposure in highlights and news coverage has long been notoriously undermeasured, namely due to limitations from legacy measurement solutions that have relied on largely manual processes,” said Dan Calpin, President, Hive – Enterprise AI. “The ability to measure the amplification of sponsorship placements accurately and at scale provides brands and rights holders alike the opportunity to more fully value their placements.”

Figure 3 – 2021 Case Study: Increase in Total Time on Screen from Super Bowl-Related Exposure

As measurement capabilities have further developed in recent years, marketers have jockeyed not just for which sports and programming to align their brands with but also for how to increase both the impact and efficiency of their investments.

“There are many ways for brands to deliver content, ranging from official league partnerships, team partnerships, broadcast partnerships, athlete endorsements, to name a few. Across sports, brands are looking for what will deliver the most connectivity and relevancy to its target audience. The data now allows us to help brand’s decide where to invest to yield the greatest return.” said Cameron Wagner, Chief Client Officer at Elevate Sports Ventures who leads the company’s brand-specific consulting services.

In Super Bowl LVI, apparel brands commanded two-thirds of total screen time among brands within the telecast. This group was led by ubiquitous Nike swooshes, but also included exposure by Oakley-branded helmet visors and New Era-branded hats as well as a handful of native exposures for Adidas, Under Armour, and Air Jordan.

Among the other asset types, league sideline sponsorships earned the greatest exposure with 57% of the remaining screen time, followed by promotion for the Super Bowl LVI Halftime Show (15%), and broadcaster-controlled assets (13%).

Interestingly, the amplification of sponsorship assets appears to be non-linear, according to data from Mensio on Super Bowl-related exposures across all nationally televised programming following 2021’s Super Bowl LV (see Figure 4). These trends are expected to be representative of expected brand exposure following 2022’s Super Bowl LVI.

While exposure for 2021’s stadium naming rights holder Raymond James earned only 9% of all visual brand exposure within the game (excluding apparel sponsors like Nike), the brand earned a staggering 45% of the amplification within non-sports programming across national television, led by numerous news and entertainment programs across TV networks filming coverage onsite on the night of and the day following 2021’s Super Bowl LV. This year’s stadium naming rights holder, SoFi, is likely to earn a similar boost in exposure in the days to follow.

Within sports programming, including the official post-game show as well as SportsCenter and other sports highlights shows on the night of and day following the game, post-game amplification favored the sponsorship assets on and nearest to the field. The league’s three official sideline sponsors (Gatorade, Bose, and Microsoft) and a handful of other in-stadium sponsors amassed 89% of the Super Bowl-related exposure among non-Apparel brands within sports programming through the Monday after the 2021 game, compared to 68% in-game.

While attention paid to broadcast sponsors – which typically include paired visual exposures and verbal mentions – may well be higher within game, those types of sponsorship placements typically earn less post-game amplification since those placements predominantly occur in between plays.

Figure 4 – 2021 Case Study: Cumulative Time on Screen from Super Bowl-Related Exposures, Mix by Asset Type (Feb 7-12, 2021; Excluding Commercials and Apparel Brands)

In-game signage and product placement may not make you laugh or cry in the same way that Super Bowl commercials do, but it’s hard to argue the volume of in-game exposure earned by top sponsors doesn’t help brands break through.

About Hive

Hive is the leading provider of cloud-based enterprise AI solutions, helping companies use AI to interpret video, image, audio, and text. The company offers end-to-end AI services, including pre-trained AI models served via API and a suite of enterprise applications. Hive’s technology enables use cases including automated content moderation, brand protection and platform integrity, content-based ad targeting, advertising and sponsorship measurement, and more. Hive processes billions of API requests per month, with industry-leading accuracy enabled by Hive’s full-stack approach. For more information, visit thehive.ai or follow on LinkedIn.

About Elevate Sports Ventures

BACK TO ALL BLOGS

Ads More Expensive Than Ever on Superbowl Sunday

Hive President Dan Calpin joins Bloomberg Business of Sports Podcast hosts Michael Barr, Scarlet Fu, and Mike Lynch. They break down the big winners of Superbowl Sunday, where ads were more expensive than ever.

Hive

February 14, 2022July 4, 2024

Listen to the episode at the Bloomberg Business of Sports Podcast.

[TRANSCRIPT]

This is Bloomberg Business of Sports with Michael Barr, Scarlet Fu, and Mike Lynch from Bloomberg Radio

MB: This is the Bloomberg Business of Sports show, where we explore the big money issues in the world of sports. I’m Michael Barr,

SF: I’m Scarlet Fu,

ML: and I’m Mike Lynch,

MB: And today we are talking big money. I mean big, big, big, big money, with Super Bowl ads more expensive than ever. Let’s break down the big winners of the night with Hive Enterprise AI President, Dan Calpin. Dan, welcome to the podcast.

MB: First of all, there were a lot of ads out there, but it seems to me the one that really dominated, and caught my eye because I’m an old man, looked like an old pong game. And I didn’t get it at first, but I guess it was like for Coinbase or something?

DC: Hi and thanks for having me. And yeah, that will probably go down in history as maybe the best direct response ad ever put on television. That was definitely a fun one.

SF: You’re referring to the one with the QR code, where people kind of stared at it and couldn’t believe that it was still on 15 seconds later and finally took a picture of it and decided they needed to do something about it. What does that say about advertisers’ need to rely on star power? Because when I was watching the ads, which I enjoyed, I saw a lot of big names – Gwyneth Paltrow, Scarlet Johansson, Lebron James, the Joneses, the various Joneses – and I couldn’t remember what company they were advertising for. I just remember seeing them and thought that it was really funny, but I couldn’t tell you which company made the pickup truck that Leslie Jones, Rashida Jones, and Tommy Lee Jones were driving for.

DC: It’s a great point, Scarlet. And I think different brands have different approaches for both what they’re trying to achieve with their ads, but also who they’re trying to connect with, and I think so much of the value of the commercial isn’t, anymore, just the 30 or 60 seconds that you’re in the program itself, but all of the amplification before and after on social media, cetera. I think the other opportunity though that you bring up is, if I were to say “What is the headphone sponsor of the NFL?” or “Which company produces the jerseys?,” I bet all three of you or the folks out in the audience would know that, and that’s the angle that we at Hive and our partners at Elevate Sports Ventures covered looking at yesterday’s game, which moved beyond the traditional ad – the 15 and 30 seconds – and actually looked at the brands that were exposed in the content. And that’s a massive amount of value. So we ended up estimating that there’s north of 170 million dollars of value generated inside the game from the brands that were exposed during yesterday’s telecast, excluding commercials.

ML: Hey Dan, this is Mike Lynch in Boston. I’m fascinated at all this data that you people have accumulated over here. So are you saying that maybe going forward that the best return might not be that 30 or 50 second spot?

DC: I think the interesting thing with marketing is, unlike the game where they’re one winner on the field, there’s lots of brands that can claim victory, and depending on what you’re looking for, there’s different ways to connect with your audience and get value from the game. But with traditional commercials, those have been understood and well-measured for decades, and the opportunity that we see with sponsorship and branded content is that it represents billions of dollars of investments, but historically there’s never been a consistent or scalable way to measure that. Most brands have looked at, kind of, whistle-to-whistle measurement, from the time a game starts to the time a game ends, but no one really has put all those pieces together to truly value how much that exposure is worth, both across brands, but also across every second of every program of television or other types of media.

MB: We’re talking with Dan Calvin with Hive Enterprise. The prices of a Super Bowl ad on TV are just going up and up and up and up and it is not going to stop. The NFL obviously is king, and as long as they have a product – and a great product that they have delivered so far this season that has just ended – the ads are going to continue to climb.

DC: Yeah, I think that’s right. In general, brands want to be associated with both where audiences are and where they find value with the content, and I think the NFL and live sports in general will always be a place that brands find heavily-engaged audiences and value in associating. And I think in the same way that, if you were to go to an electronics store and wanted to find a good set of headphones, very likely if you’ve watched NFL games Bose would be in your consideration set, because every Sunday and, you know, every Thursday and Monday, you’re exposed to that brand within the game that you love, and if it’s good enough for your coaches to hear, there’s probably that presumption that it’s good enough for you.

SF: So now that you collect all this data from companies that have brand exposure in the actual game rather than during the commercial breaks, what is the takeaway for some of the smaller companies who may not have the budget to do their traditional advertising, but have an opportunity to be part of the game itself, the content itself? What is the takeaway for them in coming Super Bowls or live sports events?

DC: It’s a great question Scarlet, and I think that our friends that Elevate Sports Ventures who we published the report along with today work with brands and right holders on this question every day. And I think if you think of the question which is as simple as, “I want to be a part of the NFL,” there’s a lot of different ways you can do that. You can buy an ad in an NFL game. You could be an official league sponsor like a Microsoft or a Gatorade or a Bose. You could be a team sponsor at any of the local stadiums. Or you could be a broadcast sponsor with NBC and CBS and Fox. And so I think increasingly, with those options, you actually need data to make those choices. And so for us at Hive, being able to produce the data that isn’t just whistle-to-whistle measurement on where your brand is, but actually the landscape of sponsorship and branded content more broadly – being able to put a value on how much that is worth helps companies like Elevate work with their clients to able to make better decisions that are more informed with data.

ML: So Dan, some of these exposures are pretty much by accident. Let’s say there’s a bumper shot. We’re going to commercial break and they go outside and there’s a drone or a blimp that’s showing and you see “SoFi, SoFi, SoFi” all over the place. That’s just bonus on top of what they pay for the naming rights, correct?

DC: It’s a good question. I don’t know if it’s necessarily – or said differently, it’s not necessarily committed in contracts, but you can bet that SoFi I think paid 625 million dollars for 30 year naming rights and a big piece of that was knowing that that brand would be front and center on screen with obviously a game like yesterday being the largest exposure. SoFi is actually an interesting story. Given the profile of both the beautiful stadium in Los Angeles, but also all of the funds that went into naming it, if you look at the exposure that SoFi got during the game yesterday whistle-to-whistle, it was actually only about a third of what last year’s stadium naming rights sponsor Raymond James received in Tampa. And to your point Mike, I think that you can’t control that perfectly, but what’s actually really interesting, and I think probably the most compelling part of the data set and platform that we built, is that there’s no doubt that SoFi is going to be a winner from this week. We’ve already seen it with all the pre-game game coverage where not just sports, but news and entertainment are broadcasting outside of the stadium and essentially presenting a billboard for SoFi. And so that ability, even if it’s not specifically committed, but to be able to understand and value how much equivalent media value you’re getting from sponsorship placement – for brands, can help them essentially both measure your return on investment and return on objectives, but also pay the right amount of money and help make those decisions. And then for the rights holder, if you know how much value is being created from your asset, even if it’s not specifically what you’ve committed to delivering, that can actually help inform how you price your assets.

[ad]

MB: Is it a mistake for advertisers to show off their ad online before you see it on the Super Bowl? Because part of the excitement of seeing the ads is that hey, I want to see it debuted, but sometimes it sneaks over onto certain websites and it kind of takes the luster away from it. Am I right or am I wrong?

DC: I think the answer is probably in the respective brand. The challenge with the Super Bowl, and especially with commercials, even though they’re kind of a second game in and of themselves, there’s still lots of brands competing for attention, and you’re counting on that one 30-second moment in time to be able to capture the attention of the world. And so the benefit of pre-releases and post-releases is that it creates reach and exposure so that you have more opportunities to meet folks, but I think there’s a broader, interesting point of, if you’re kind of placing your bet on exposure on those 30 seconds, that does key up again the value in relative terms of sponsorship placement. So if you look at yesterday’s game, the top 8 brands by duration, if you take commercials and/or time in the game, actually didn’t air a commercial in the game. So if you take a brand like Nike, they had a cumulative amount of minutes that were more than 46 minutes of duration, that there were swooshes on jerseys, on sleeves, on gloves, and if you go down the list, I think in total we had 8 or 9 brands had more than a minute of exposure. And obviously you don’t have the sight and sound in a theater of telling a commercial, but if you think of opportunities to associate your brand with the NFL and just be top of mind with the world’s audience, there’s a really compelling opportunity with sponsorships and branded content.

SF: Do companies reach out to you to say or to ask, “How can I be the next Nike to affiliate myself, align myself more closely with the NFL?“

DC: Yeah, so at Hive, our business is really focused. So one step back – we’re predominantly a technology company in the AI space. Hive helps companies use AI to interpret video, image, text, and audio, and historically clients have built their own solutions around our technology for a diverse set of use cases, whether that’s out of domain, like content moderation for social networks, or content-based ad targeting for video publishers. And part of our focus over the last couple of years has been selectively building an application for ourselves for select use cases. And in this space specifically, we saw a market where the adoption of quote unquote AI was early, and the demand for data and insights AI produced was very underserved. And so the media and advertising industry broadly was an area that we saw opportunity in, and within it, sports sponsorship was an area that was ripe for disruption – historically characterized by very ad hoc and mostly manual measurement, which was frustrating clients who would have to wait weeks long for data and still only have part of the picture. And so we took our core technologies of world-class logo detection, object detection, speech-to-text models as building blocks, and then actually set up our own content ingestion pipeline so that we could feed always-on data into those models, and then built out a point-and-click platform to put more comprehensive and granular data in the hands of brands, agencies, and right holders. So today we work with a broad and diverse set of rights holders and TV networks like Disney, some of the world’s largest and most active brands and sponsorships like Anheuser Busch or Walmart, and a whole host agencies and consultancy partners, including Elevate sports Ventures who we published last night’s report with, who can kind of take our data, but then use that with their clients to turn it into really actionable insights.

ML: Dan, can you put a value on the promotional spots run by NBC last night promoting Peacock and primetime shows –

SF: Oh, good question.

ML: – their newscasts I saw were promoted, the Olympics, and anything else that’s going to run – and Telemundo? You know, they’re giving up 7 million dollars every 30 seconds to run a promotional spot for which they receive no monetary value immediately, but is there a long term uptick for them?

DC: It’s a great question, and that’s always the opportunity cost of promotional units is, the opportunity to advantage your own platform versus to realize revenue. I think historically – and it has probably only become even more the case in a streaming world with Peacock and Disney Plus and HBO Max – but many shows launched through the Super Bowl. And historically – you know, last night NBC let us see the Olympics – but for many decades, networks have taken their biggest bets on essentially the show that follows the Super Bowl. So it’s a delicate balance between what you sell versus what you use, but definitely something that creates value either way.

MB: Dan Calpin who is with Hive Enterprise. Dan, you are just full of knowledge when it comes to advertising, and I appreciate you for just taking the time out and giving me a good education. We appreciate it. Thank you so much, Dan.

DC: Thank you so much. Appreciate talking with you all and enjoyed the game last night and fun to get to extend it into Monday morning.

MB: Thank you. This is the Bloomberg Business of Sports Podcast, catch us here each and every Monday, Wednesday, and Thursday, exploring the world of money and sports.

BACK TO ALL BLOGS

Hive Completes SOC 2 Type 1 Audit

Hive

February 7, 2022July 4, 2024

At Hive, we understand that our customers continually put their trust in us to provide the best quality of service possible. We take this trust seriously and we work hard to ensure that we’re able to provide the highest level of security we can. As a first step to showing our commitment to security, we’re proud to announce that Hive has successfully completed a SOC 2 Type 1 audit with the Trust Service Criteria of Security, Availability, and Confidentiality. SOC 2 is the most accepted information security audit for North America, and we believe that passing this audit reinforces our commitment to maintaining best-in-class internal controls for safeguarding our customers’ data.
We recognize that data security is a critical concern for our customers. This is why we have ingrained security into all of our engineering processes at Hive. With a host of preventative, detective, and restorative measures, we believe we have enabled 360 degrees of security around our infrastructure and critical customer data.

What is a SOC 2 Audit?

The SOC 2 Audit is designed for organizations that provide services to other entities while interacting with their data. It provides a consistent set of criteria by which to measure the security, confidentiality, availability, processing integrity, and/or privacy practices of an organization. An independent third party CPA firm must conduct the audit, after which it issues an audit report with the findings. There are two types of SOC 2 audits conducted:

Type 1 – Report on management’s description of a service organization’s system and the suitability of the design of controls.

Type 2 – Report on management’s description of a service organization’s system and the suitability of the design and operating effectiveness of controls.

The SOC 2 audit is conducted on an annual basis to measure continued success in the defined criteria.

What’s Next?

We believe that continuous innovation is key to providing the best service possible. As Hive grows in size and complexity, we recognize that it is critical for our security practices to grow as well. On top of continuously monitoring and adapting our security practices, we will move forward with a SOC 2 Type 2 in 2022, and have ISO 27001 on the compliance roadmap.
For more information on our security practices and plans, please contact our security team at security@thehive.ai.

Understand

Search

Generate

Technology & Digital Platforms

Sports, Media, & Marketing

Careers

About Us

Contents

Building a Smarter Way to Search

Image Similarity Models: A Two-Pronged Approach

Why Use Similarity Models for Image Comparison?

Example Image Comparisons and Model Responses

Hive’s Custom Search: API Overview

Final Thoughts

Easily Connect Moderation Dashboard to Your Application

Implement Custom Content Moderation Rules

Intuitive Adjustment of Model Classification Thresholds

Streamline Human Review

Setting Human Review Thresholds

Review Feed Interface for Human Moderators

Access Clear Intel on Your Content and Users

Effortless Moderation of Spam and Promotions

Final Thoughts: The Future of Content Moderation

Presentation: How AI-Powered Measurement Can Increase the Value of Your Sponsorships by 30% or More

OCR Model for Text Recognition

OCR Moderation: Language Support

Moderation of Detected Text

Emoji Classification for Text Moderation

OCR Moderation: Final Thoughts

How Hive Can Help: Speech Moderation

Speech Transcription

Audio Moderation

Language Support

Beyond Words: Sound Classification

Final Thoughts: Audio Moderation

At a glance:

Official NCAA partners lead in-game brand exposure

AT&T, Capital One, Coca-Cola, Nike, and Spalding among March Madness exposure winners

Nike leads exposure among apparel brands; Air Jordan outperforms in both tournaments

Official NCAA Sponsors earned higher search engagement during telecasts

About Hive

About Elevate Sports Ventures

At a glance

Official NCAA partners lead in-game brand exposure

Nike leads exposure among apparel brands to date; Air Jordan expected to outperform going forward

Different brands advertise in Men’s and Women’s tournaments

About Hive

About Elevate Sports Ventures

At a glance

Brands earn $170 million in equivalent media value from in-game exposure…

About Hive

About Elevate Sports Ventures

What is a SOC 2 Audit?

What’s Next?