main

Technology

Win Over World Cup Watchers with MediaMath

November 7, 2022 — by MediaMath

blog-cover-header-960x292.jpg

The 2022 FIFA World Cup and the holidays overlap this year, with the sporting event scheduled to air globally from Qatar between Nov. 20th and Dec. 18th. Up to 5 billion viewers are expected to tune in to the event.

While 77% of consumers plan to cut back on holiday spending this year, the World Cup happening in conjunction with big sales days like Black Friday and Cyber Monday means advertisers could see increased purchases in categories like food and drink, streaming subscriptions, TVs, football kits and merchandise and travel to Qatar. This also means there could be high demand for holiday ad space that could push up media costs.

Having the right full-funnel, omnichannel advertising strategy in place will be paramount to capturing consumer attention during both the World Cup and the holidays. MediaMath is your go-to partner to reach the right high-value users with high-quality media at scale.

Download our 2022 World Cup Campaign Guide to learn how to execute a winning campaign strategy with MediaMath.

In our guide, you’ll find:

  • Ways to reach your ideal prospects and re-engage your best customers using MediaMath’s audience and contextual suite of targeting solutions in partnership with Oracle, Eyeota and Lotame
  • Information on MediaMath Marketplaces with curated inventory that reaches key World Cup audiences, such as gambling, alcohol, performance, brand engagement and viewability
  • Tips for executing a full-funnel approach across DOOH, CTV, native, social display and more using our transparent supply ecosystem

Get the 2022 World Cup Campaign Guide here and, if you’re a current client, reach out to your MediaMath account manager to start planning your strategy today.

TechnologyUncategorized

Standing with Ukraine

March 9, 2022 — by MediaMath

Ukraine.png

MediaMath Standing with Ukraine

MediaMath is committed to making advertising better for brands and consumers worldwide. While it may seem like a small contribution, advertising should not help support Russia or Belarus’ actions in Ukraine.

MediaMath will not support any advertising in Russia or Belarus for the foreseeable future. No campaign material targeting Russia or Belarus will be accepted, ad traffic from Russia and Belarus will be removed from our ecosystem and geo-targeting blocks will be in place.

We have a long-standing policy against disinformation and work collaboratively with the Global Disinformation Index (GDI) to ensure our supply complies with their recommendations. GDI protection blocks approximately 1,700 domains that publish and promote a broad range of disinformation narratives, including approximately 120 domains that are publishing disinformation narratives and content related to Russia’s invasion of Ukraine with that number is being added to regularly as events unfold. Any site found to be spreading disinformation is added to our Universal Block List and no ads will be served to these sites. We also offer additional disinformation protections based on journalistic standards via NewsGuard from Peer39 and Comscore.

Our advertising clients can also use brand safety tools from our contextual partners like Oracle, Peer39 & DoubleVerify to avoid their advertising appearing on content featuring sensitive material. For example, custom negative keyword lists surrounding terms such as Russia/Ukraine, military, war etc. However, we caution our clients to avoid over-blocking via keyword targeting. Studies suggest quality news content is not generally a brand safety risk, regardless of topic. More importantly, journalism is essential and should continue to be funded through sensitive and appropriate advertising. If you have any questions, please reach out to your account representative.

Neil Nguyen, MediaMath CEO

DIGITAL MARKETINGTechnology

MediaMath and Lotame Prove Cookieless Targeting Can Drive Addressability in the Future

January 18, 2022 — by MediaMath

lotame-mediamath.jpg

In the lead-up to cookie deprecation in 2023, brands are grappling with which identity solutions to assess and test. That’s why MediaMath gives marketers choice when it comes to evaluating the identifier best suited to their business and the ability to easily activate it through our recently updated platform built for this moment and the next.

We recently worked with the Lotame Panorama ID™, the only people-based identity solution for the open Web and part of our Identity Marketplace, to see if third-party data could be used in cookieless targeting across all browser environments for one global financial brand. This first-of-its-kind cookieless test campaign in Latin America consisted of targeting personas showing continued interest in traveling locally or abroad: “Tulum-inatis,” single, young adults with no children, and “Lady and Lord Multitask,” married individuals ages 29-45 with children.

The brand’s agency worked with Lotame to build audiences that reflected their finely crafted personas for their Latin American consumers. Leveraging the enrichment capabilities of the Lotame Panorama ID, Lotame provided the high-quality data necessary to transform the brand’s personas into targetable audiences. MediaMath then activated and optimized the segments, using the same age and intent travel parameters to target both audiences via cookie segments and Panorama ID cookieless segments. The latter delivered superior results. In addition to increased impressions, the Panorama ID campaign resulted in:

  • 2.5x more efficiency in frequency capping
  • 44% more unique individuals reached
  • A $7 cost savings in eCPM, with the cookie campaign 116% more expensive than Panorama ID

See the details of the test and the results – Read the full case study below.

TechnologyTrends

Marketers Must Master These 4 Things to Win This Record-Breaking Holidays Sales Season

December 1, 2021 — by MediaMath

aaron-burden-Mu_9w7l1koI-unsplash-960x720.jpg

The last 18+ months have accelerated certain digital trends that will impact this holiday shopping season in a major way. Marketers must be prepared with strategies that help them stay agile and adapt. Built for this moment and the next, the MediaMath platform can help you execute your holiday campaigns with stability and flexibility in mind.

The Shopping Landscape for Holiday 2021

After more than 18 months in and out of lockdowns and quarantines, US consumers are ready to spend like they’ve never spent before this holiday season.

Even with supply chain issues, shipping delays and an uncertain economic forecast, retail and e-commerce sales are expected to reach between $1.1 trillion and $1.3 trillion in the US this Q4 shopping period according to Deloitte. This trend will extend globally: more than a quarter of Australian shoppers and more than half of UK shoppers are planning on spending more this Christmas. Download detailed holiday guides to help with campaigns in each of these markets: North America here, EMEA here and Australia here.

Call it “revenge shopping” or the YOLO economy or whatever you’d like — taking advantage of this return to pre-pandemic shopping requires a data-driven, omnichannel approach to keep your brand top of mind in the month ahead.

Find the Right Shopper as They Search for the Right Gift

The path to capturing your share of this year’s massive holiday spending begins with your own data. Onboard whatever first-party CRM data you have, such as email addresses, to target seasonal shoppers that bought last year. To reach shoppers in the research phase, target audience segments based on what they have looked at on your site or the ads with which they have engaged. You can then model out those audiences to find shoppers with the same characteristics as those who buy from you most. Keeping your audience segments fluid is key as shoppers will browse, cart and purchase rapidly. As users move in and out of segments, reach customers based on what they intend to buy and drive awareness before targeting them with specific promotions.

Whether you are activating first- or third-party data, MediaMath’s identity graph can expand the reach of audience segments in privacy-first and probabilistic ways. Utilize our identity graph to ensure you can control ad reach and frequency at a person level and extend reach to mobile devices and high-quality video this holiday season.

Keep Your “Presents” Strong Throughout the Funnel

COVID-19 has accelerated two main trends in consumer behavior: use of mobile and adoption of CTV. Mcommerce will continue to gain share of total holiday e-commerce, jumping 18.8% to $97.15 billion and accounting for 8.5% of holiday retail sales. CTV households will increase to 115.2 million by 2025, from 106.4 million in 2021.

For holiday performance-based campaigns, you want to target at the user or device level to drive performance across all OTT channels — that is where you will see conversion. CTV is designed for awareness, so you can follow up from this channel with sequential messaging to drive engagement down the purchase funnel. As holiday travel ramps up, consider emerging channels as well, such as audio or DOOH. A unique approach not to be missed this holiday season is social display, in which you can use the same social media creative across quality inventory for engagement and creative at the scale of traditional display. Advertisers using social display see 5x improvement in brand lift and consideration on average.

Here is the best breakdown of channels to pursue for each buyer stage this holiday season:

  • Awareness: OTT formats, social display, audio, mobile video, DOOH
  • Action: OTT formats, social display, online video, native, contextual, behavioral, retargeting, audio
  • Interest: Display, contextual, behavioral, retargeting, audio, social display
  • Conversion: Retargeting, display, audio

Measure True Business Outcomes

It’s critical to have a measurement strategy in place that fits your business needs, whether it’s on-site engagement, in-person footfall traffic or CTV measurement. Leverage footfall traffic data to inform attribution and Brain optimization in our platform. Automate as much as possible by utilizing MediaMath’s data connections to import your data into our platform so that campaigns can automatically optimize towards your most valuable users. Increase your media efficiency and bid accuracy by ingesting your third-party attribution data to inform optimization in our platform and have it available in data export reports. Go deeper into your supply path optimization strategies through publisher revenue reporting and working media percentage calculations.

Select Supply that Sleighs

The content your ads run on is more important than ever in this age of dis- and misinformation. Our SOURCE ecosystem is the best way to guarantee brand-safe inventory and transparency into how holiday campaigns are performing. We see advertisers committing to programmatic guaranteed, private deals and marketplaces to secure inventory across all channels that reach holiday shoppers. Additionally, targeting options via custom contextual segments help you reach valuable audiences during the holiday crunch. Check with inventory partners, especially on PMP deals, to see if they will still have available inventory as you move through the quarter. Also consider diversifying supply vendors as well as possibly moving deals to programmatic guaranteed.

A final note: If possible, attempt to extend holiday campaigns through the New Year to take advantage of lower CPMs and increased scale. This past January, there was a record-breaking number of returns for 2020 holiday gifts during what has become known as “National Returns Week.” Be prepared to attract those shoppers looking to swap out their wares for something better to put the final bow on your holiday sales performance.

DIGITAL MARKETINGTechnology

IBM Watson Advertising Weather Targeting Now Available in MediaMath’s DSP

November 9, 2021 — by MediaMath

Rainy-Window-2-960x667.jpg

From water cooler small talk to heated debates on climate change, the phenomenon of weather has fully saturated our lives in recent years. And with good reason.  

Extreme weather events are becoming commonplace, increasingly impacting our lives: our commutes, our vacations, choices on where to live and our health. Weather remains a proven and privacy-friendly predictor of consumer intent and behavior — it accounts for nearly half a trillion dollars in economic impact in the U.S. alone each year. But just as it is becoming increasingly challenging for brands to effectively target and engage their audiences due to privacy regulations, weather itself has become harder to predict, making it difficult to leverage for advertising products and services to consumers.   

For all the reasons noted above, MediaMath is excited to announce the next phase of our longstanding partnership with IBM: the integration of IBM Watson Advertising Weather Targeting into our DSP, which will empower brands to future-proof their strategies with addressable, performant solutions for the open Internet that do not rely on traditional identifiers.  

IBM Watson Advertising Weather Targeting uses AI to recognize what the weather “feels like” and how consumers in a specific location are likely to react. It then aggregates and analyzes increasingly large and complex data sets to enable more effective, efficient targeting opportunities using only a contextual ZIP Code that is passed to an ad server or platform.  

Third-party cookie data is not required. Weather Targeting only serves relevant ads to consumers based on the projected impact of upcoming weather conditions in a defined area, helping marketers increase the precision and effectiveness of their campaigns, reduce wasted ad spend and meet new privacy standards.  

More than 300 triggers proven to perform 

IBM Watson Advertising Weather Targeting turns the relationship between weather, location and complex data sets like health conditions, product sales and consumer activity into actionable, proven solutions. Here is how this can break down across categories: 

  • Relative conditions can account for how weather in its various forms is felt differently across regions. For instance, 50 degrees might feel cold in Miami, but not Chicago. 
  • Product triggers can reflect the optimal weather conditions that drive demand. For instance, a huge snowstorm might drive consumers to stock up on essentials like toilet paper. 
  • Pharma triggers include Symptom, Prescription, OTC and Predictive Health and can be based on published medical research, IMS data and sales data for conditions such as allergies or asthma. 
  • Activity triggers reflect the mix of weather conditions that are likely to drive certain behaviors and experiences, such as hiking outdoors or a game night in the living room.  
  • Watson Health anonymized data can enhance the precision of IBM’s triggers, helping to reach the consumers most likely to be experiencing symptoms. 
  • COVID-19 infection rates are gathered from public sources, aggregated and analyzed using Watson AI to serve messaging tied to increasing or decreasing cases based on local conditions, driving the relevant consumer actions. 

Working together to make advertising better 

Back in 2017, MediaMath Founder and CEO Joe Zawadzki and IBM Chief Digital Officer Bob Lord stood on stage at DMEXCO and painted a picture of the future of marketing that paved the way for our accountable, addressable and aligned SOURCE ecosystem launched in October of 2019. The vision they laid out would solve many long-standing challenges that have prevented marketers from unlocking the full potential of brand engagement, and help them deliver advertising with transparency and trust. Just two years later, IBM as a client was our first advertiser to run a campaign in SOURCE, seeing outstanding results including 60% of the ad dollar making it to the publisher and viewability increasing by 40%. A little over a year ago, we integrated IBM’s upgraded suite of AI advertising tools — including IBM Watson Advertising Accelerator and IBM Watson Advertising Predictive Audiences — into our DSP.  

 We are thrilled for this next phase of our continually evolving partnership that we know will help more brands and their agencies deliver the right ads to the right consumers with relevance and respect.  

TechnologyTrends

MediaMath in Conversation with Smart Adserver on Driving Advertiser Outcomes and Achieving Better Monetization for Publishers

December 7, 2020 — by MediaMath

MM_Smart_Scorecard.jpeg

In October 2020, we released the new SOURCE Ecosystem Scorecard to continue our quest to offer unmatched access to data and transparency in the digital supply chain. The Scorecard is a configurable setting that provides brands and agencies confidence in where their advertising dollars go regardless of the channel.

Demands for greater transparency are everywhere. In May 2020, the Incorporated Society of British Advertisers (ISBA) released the findings of its two-year-long research into the workings and value of the current programmatic supply chain. There are several callouts and recommendations in ISBA’s report for all players in the industry to strive toward that more transparent, efficient supply chain—and they are things we are already doing with our partners.

As part of the Winter Ad Tech Virtual Event, Vignesh Narayanan, Senior Director, Media Partnerships, MediaMath recently spoke virtually with our partner David Pironon, Chief Programmatic Officer, Smart Adserver about how the industry has been working to solve some of these challenges since ISBA released its report and how through our new Scorecard they are helping drive advertiser outcomes and achieving better monetization for publishers. Watch a recording of the panel here.

Watch a recording of the panel here:

Smart is well aligned with MediaMath on the vision of a high-quality, transparent and healthier digital advertising industry. Smart has been an early adopter of new market standards and best practices in ensuring both publisher and ad quality as well as taking industry-leading roles in early proofs of concept such as Trust.ID.

Among these, Smart was an early leader in the adoption of ads.txt, app ads.txt and, more recently, sellers.json and SCO, all of which are meant to benefit both sides of the business:

  • For brands, quality audience in premium publisher environments on every channel, format and device with the assurance of brand safety and efficient investments
  • For premium publishers, not only a better valuation of their quality inventory, which will ultimately lead to an increase in revenues, but also a better understanding of the value added by each tech partner and the opportunity to optimize their monetisation path.

The alignment of core values and operational excellence provides a strong foundation for the integration between MediaMath and Smart. Together, we take a shared interest business approach, bringing advertisers and publishers the best-in-class, highest-quality advertising ecosystem.

Technology

MediaMath, SuperAwesome and Penguin Random House Discuss How Brands can Safely Engage with Kids Through Digital Marketing

November 23, 2020 — by MediaMath

PPS-MM-PRH-SA.jpeg

Back in October, we announced an exciting new partnership with SuperAwesome to enable child-safety and privacy in programmatic advertising. Through SuperAwesome’s Kid-Safe Filter, MediaMath ensures all ad impressions presented to children through our platform are safe and free from children’s personal identifiers. Additionally, all ad creatives are reviewed by SuperAwesome’s moderation process to ensure they are suitable for the audience in question.

As part of this year’s virtual Programmatic Pioneers Summit, MediaMath’s VP of Client Success Noemi McKee sat down with SuperAwesome’s COO of North America Kate O’Loughlin and client Penguin Random House’s Ad Ops Manager Rick Garcia to chat about how brands can safely engage with kids through digital marketing as we enter the critical holiday shopping period. Watch the recording of the chat below.

 

DataMediaTechnology

MediaMath Continues to Bring Unmatched Transparency to Digital Advertising with New SOURCE Ecosystem Scorecard

October 20, 2020 — by MediaMath

Scorecard-1.jpeg

In 2018, we released an open letter to 50 SSPs saying we’d stop buying from them if they played games that contributed to unfair, murky programmatic auctions. Almost a year to the day later in 2019, we took another huge step towards greater transparency with the unveiling of SOURCE, our effort to build a programmatic ecosystem that is accountable, addressable and aligned. And today, two weeks after celebrating SOURCE’s first birthday, we are releasing a new SOURCE Ecosystem Scorecard to continue our quest to offer unmatched transparency in the supply chain.

“With SOURCE, we set out to address some of the biggest issues facing the industry—to create an environment that is accountable and addressable, and to align the interests of brands, agencies, publishers and technology providers,” said MediaMath Founder and CEO Joe Zawadzki. “The SOURCE Ecosystem Scorecard continues to make good on that promise and is truly a Holy Grail for the industry— a configurable setting that provides brands and agencies with unparalleled data and transparency, giving them confidence in where their advertising dollars go regardless of the channel.”

The Scorecard has nine, equally weighted data signals that are split across Accountability, Addressability and Alignment and tally up to a final score across all ecosystem partners which directly leads to increased buyer value. Currently, the average score for the 82 participating SSPs is a 7 out of 10. We wouldn’t expect a perfect score of any SSP—that’s not the purpose of the Scorecard. Instead, we want the Scorecard to provide an incentive and roadmap for our partners to pursue a more  transparent ecosystem in which we provide measurable value and collectively empower advertisers to make better, more informed decisions.

ACCOUNTABLE

  • Fee Transparency Method: To achieve a top score, partners must provide complete transparency into supply chain costs so customers can understand the true value of every impression and create data-driven optimization strategies.
  • Supply Authorization: Partners with the top scores provide impression-level seller data and publisher path transparency to inform data-driven enterprise and campaign-level SPO strategies.
  • Inventory Type: Top scoring partners enable granular inventory and placement targeting, target instream instead of outstream video and define skippable vs. non-skippable video access.

ADDRESSABLE

  • Fraud & Invalid Traffic: Top scoring partners for this signal have 100% of their supply filtered by White Ops for pre-bid IVT, provide free fraud protection for buyers, have no post-campaign IVT claw backs necessary and no undelivered marketing budgets due to IVT.
  • Preferred Identity: Partners with top scores here will provide an increased find rate to customers by transacting on multiple common IDs and measuring consistently so they can attribute at the person and household-levels.
  • Unique Channel Identifiers: Partners who exhibit this signal use appropriate identifiers for channels such as CTV and ensure omnichannel reach to audiences across devices.

ALIGNED

  • Bid Feedback: This signal is focused on informing effective bid shading, ensuring advertisers pay the fair price for impressions and providing future support for auction insights reporting.
  • Brain Sync: Partners who exhibit this signal at the highest level better inform seller bidding algorithms and improve supply performance toward advertiser goals by ingesting MediaMath algorithm data.
  • Impression Counting Method: This signal is scored on how well partners ensure marketers pay only for impressions that have the potential to be seen, align counting methodology across the supply chain and reduce counting discrepancies.

A sample of the Scorecard:

The SOURCE Ecosystem Scorecard is available globally. If you’re interested in learning more, please reach out to your Account lead.

Technology

Streaming Data at MediaMath

September 30, 2020 — by MediaMath

Streaming-Data.jpeg

Welcome to MediaMath’s Tech Series, which features voices from our Engineering team, responsible for developing and maintaining our industry-leading tech stack. Check back every few weeks for technical deep dives from these expert engineers.

At MediaMath, an entity is defined as an object that can be uniquely identified and has a state which can change over time. The change in entity state drives business decisions for downstream consumers. The entity state is shared across multiple services and a consistent view for all consumer services is required.

MediaMath is a leading omnichannel demand-side platform that offers a rich set of capabilities for marketers to manage and execute their campaigns. Our omnichannel platform supports execution, management, targeting and reporting all in a centralized location. Under the hood, multiple services are at play that share entity state to achieve that uniform experience. Any change in one part of the flow can affect other aspects of the workflow handled by a different service.

To ensure that state is consistent across all services, we use an -compliant database to store entities. ACID guarantees the database avoids the problem of inconsistent views. All services read and write to a database that provides transactional guarantees. The records in the database can be safely considered the source of truth.

As MediaMath scaled, this source of truth potentially became a single point of failure. During heavy load, the database would experience delays serving requests to multiple clients. Adding read replicas and vertically scaling the database mitigated some bottlenecks, but we realized it was not the ideal solution.

CQRS: Command Query Responsibility Segregation

CQRS is a well-known pattern first described by Greg Young and well-explained in this article by Martin Fowler. It means having separate models to read information and update information, also called CommandQuerySeparation.

The CQRS pattern tends towards a micro-services architecture having separate services for reading and writing data. It decouples the responsibility between read and write service so that each service can be built independently to maximize efficiency. It is ideal for event-based architecture where decisions are taken only when an event occurs. Each service can be scaled independently. This can be huge when there is a disparity between reads and writes. Along with the benefits, there are some caveats especially related to the consistency of data. Most of the systems built on CQRS guarantee eventual consistency with some accepted latency.

After a period of assessment on our use-case, we decided a CQRS-based service would meet our needs to build the entity platform.

Enter `Changes`

Changes is MediaMath’s distributed entity platform. Based on the read model of CQRS, Changes facilitates the exchange of entity data across internal teams. All entity updates on the database are served to the consumers within an accepted latency through the platform. All updates are emitted as events that are available for consumption for any services subscribing to them.

Changes provides several guarantees:

  1. Durability: Messages are stored to non-volatile storage and can be accessed repeatedly.
  2. At least once delivery: If an event has been acknowledged, it will be delivered to all consumers who ask for it.
  3. Sequential consistency: Event order is maintained for all consumers.
  4. Globally distributed: Events will be available across the globe.

Changes was developed to solve multiple problems:

  1. Reducing the load on our legacy database deployment: As consumers of entity data move away from it, it reduces significant load on the database, thus reducing the risk of choking our single point of failure due to the read clients.
  2. Support multiple subscribers: Changes is a centralized, highly available platform that can support many subscribers reading concurrently to consume entity data.
  3. Scalability: Due to its distributed nature, as the number of subscribers or the entity data scales up, Changes is built to scale along by simply adding more nodes to the system.

Architecture

Changes is built on Kafka, a distributed log service optimized for low latency, high throughput systems and provides scalability and fault-tolerance right out of the box. The set of services that makes Changes are written in Go and heavily optimized to reduce latency in delivering entity data.

Changes offers two kinds of data:

  1. Event data or history: All events in the order they were committed to the database.
  2. Entity data or snapshot: Current state of the entity as on the database.

Both history and snapshot data are available as messages in Kafka topic. All consumers of entities are supposed to have their own consumer client to read data from those topics. We leverage the Confluent Schema Registry to manage schemas for all messages published to the topics.

There is also a separate daemon service that constantly monitors events on both topics to ensure consistency and accepted latency.

Event sourcing: history

The event data is stored in what we call history topics. A history topic contains the ordered stream of events. Kafka guarantees ordering only within a single partition. To ensure ordering across all entity types, we publish all updates to a single partitioned history topic. The history topic has a retention period of 30 days; hence, the ordered events go back as far as 30 days. Any update that occurred more than a month (30 days) back will be removed from the history topic.

To get the state of an entity at a certain point in time (within the last 30 days in our case), we can replay all messages from the beginning of the topic up to that point in time. Messages re-played from the beginning of a history topic cover a 30-day-old snapshot and should provide you the current state of any entity. The time concept in Kafka is record/message offset. This is useful for auditing events that happened in the past.

Entity state: snapshot

Entity snapshot is defined as the current state of the entity. Every entity is uniquely identified by a 64-bit integer entity ID. Think of it as a primary key of a record in a database. These are stored in snapshot topics, which will eventually keep only the last record for every entity ID.

The snapshot topics are defined per entity type, have multiple partitions and are log compacted. Every Kafka record has a key and a value. The key is the entity ID of the record. Log compaction deletes all stale entries for a particular key and keeps only the latest one. However, it is possible to have multiple records of the same key which have not been compacted yet. Any cache sourced from snapshot topics should only consider the last record for every key. The cache should be updated for the same key.

At any point, a snapshot topic contains the entire state of an entity type. This helps bootstrap new services or old ones that need a cache refresh. Simply consume all messages in a snapshot topic and build the cache.

Changes in production

Changes has been running in production at MediaMath since 2016 and has evolved to support large-scale updates. Changes observes 3.5 million updates per day on our database alone and 2.1 million updates pass through as captured events to be published downstream per day. The numbers are current over a seven-day rolling average.

Changes has strong latency guarantees for when the data would be available for consumption from the time it is committed on the database. Over the previous week, we are currently at 100% delivery under one hour, 89% under one minute and 81.8% within five seconds. This considers all factors including change data capture, network latency and synchronization in Kafka.

Technology

MediaMath a Leader for the Third Year in a Row in 2020 Gartner Magic Quadrant for Ad Tech

September 24, 2020 — by Anudit Vikram

Gartner2020-Blog.jpg

Almost a year after we launched our digital supply chain clean-up initiative SOURCE by MediaMath, we’ve been recognized as a Gartner Magic Quadrant Leader for the third year in a row, based on our completeness of vision and ability to execute. Download the report here.

At MediaMath, we do not try to be the best at everything—we try to do what’s best for our customers. Our philosophy of bringing a product to market focuses on supporting brands and agencies by fully delivering on things we can do best ourselves and partnering on other capabilities that are not core to MediaMath. We know what our capabilities are—they include:

  • Focus on curated ecosystem: We were first to market with supply path optimization and the only DSP that takes a robust approach to active supply chain management and stands committed to building a 100% accountable, addressable and aligned supply chain. This supply chain includes complete, granular publisher-level transparency into all fees and costs that inform real-time optimizations or quick supply path optimization modifications. We’ve moved several major clients over to SOURCE and have hundreds of publishers making trusted, quality, inventory available, including news and minority-owned sites to help brands support healthy, diverse content in an age of misinformation, hate and propaganda.
  • Data science-driven innovation: Over the past year, we’ve overhauled our Brain machine learning algorithm to be more scalable and spend less time learning and more time optimizing. The latest iteration looks at both media variables as well as user variables to predict auction rates more accurately. Over time, Brain learns predictive features such as publisher, exchange or creative size, yielding more media efficiency for brands and their agencies.
  • Big brands choose MediaMath: We service both large, enterprise brands and agencies, from major holding companies to independents. We also are inventory-agnostic as far as from whom we source quality supply and have a breadth of expertise via consultative experts who maximize client value from campaigns, platform features and functionalities. We custom build for our clients, execute for our clients and immerse ourselves in their businesses.

The areas in which we have been working to grow our expertise and offerings, both through building internally and partnering externally, in addition to how we’ve pivoted our business to account for market forces these last few months, include:

  • Omnichannel addressability with CTV: We are committed to providing transparency, integrated, omnichannel addressability and direct connection between clients and publishers through our CTV offering. We’ve announced key partnerships with TVSquared, with whom we are evolving our measurement capabilities, Hulu, who provides differentiated inventory, and Tru Optik (watch the Beet.TV interview with their CEO and Co-Founder and Andre Swanston talking about our partnership here). Our identity work, including our partnership with LiveRamp to integrate their IdentityLink into our bidder, will associate CTVs with users and their other devices to enable true omnichannel marketing and measurement and add a household-level association to people and devices to enable more powerful attribution and features that appeal to linear buyers.
  • Partnership during uncertain times: COVID was an event none of us could have expected. Our MediaMath teams quickly adjusted to ensure we could provide top-notch service and support to clients and their agencies while working remotely across the globe. We have created custom dashboards for clients and continue to monitor the landscape to identify any change that could impact performance, CPMs and more (some of it is tracked on this public-facing microsite here). We have chosen to meet the transformation needs of the industry and our business in terms of scale and proven operational expertise in several key new hires in sales, product, finance and HR.
  • Clients and partners: MediaMath has focused on fewer, bigger direct brand relationships and relied on a reinvigorated “with, though, and to” channel strategy with our agency partners. We have also whittled down the number of direct partners with whom we work, driven by GDPR and CCPA on one hand where quality partners tend to attract quality media, and the need for explicit roadmap alignment to drive SOURCE which demanded more focused attention.

Clients—across industry verticals—acknowledge the importance and need of an accountable, addressable and aligned supply chain to drive performance, especially during tumultuous times in which budgets are being even more scrutinized. We commit to quality and focus as we build for what our clients need.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

Gartner 2020 Magic Quadrant for Ad Tech, Andrew Frank, Lizzy Foo Kune and Eric Schmitt [Sept. 21, 2020]