main

DataIntelligence

1995-like Media Buying Would Bring 1995-like Outcomes

December 6, 2017 — by Lewis Rothkopf0

LR_LinkedIn.jpg

This article, written by Lewis Rothkopf, General Manager of Supply at MediaMath, originally appeared on LinkedIn. 

Like many of you, I read with interest Augustine Fou’s piece, “What a Concept! Buy Media As If It’s 1995,” in which he shares some disturbing examples of middlemen subtracting disproportionate value from media transactions. Also like many of you, I consider Dr. Fou to represent the very best of our industry — he’s a much-needed independent voice in cybersecurity and fraud research and someone whose perspective I hold in the highest regard.

Our opinions differ, however, when it comes to some of the solutions advocated in his latest piece. To be sure, there is no legitimate place for vendors who add no value to the ecosystem — they are a drag on publisher RPMs and do not promote better business outcomes for marketers. Simply put, if you don’t add value to the media transaction, you don’t deserve to be a part of it.

But to paint all intermediaries — “from agencies to ad tech vendors” — with the same brush would be a mistake. When looking at those entities who play a role in the value chain, it is important to separate Black Holes from Shining Stars.

Black Holes:

• Roll up, or aggregate publisher supply without adding a layer of value or differentiation such as data enrichment, geolocation or fraud protection
• Misstate the provenance of the supply or their right to resell it
• Are highly discrepancy-prone
• Are often several layers separated from the underlying publisher inventory

Whereas Shining Stars:

• Offer a layer of protection for buyers by offering assurance — and sometimes a guarantee — that the supply is legitimate, safe and fraud-free
• Add value to otherwise-commoditized supply with the addition of data
• Help identify the most appropriate demand for particular inventory
• Innovate on formats with marketers and publishers
• Normalize disparate pools of supply to make them addressable across channels
• Improve measurement and accountability
• Work to keep costs low

Additionally, though it can be tempting to dispose of baby along with bath water and flee from programmatic and its defense mechanisms (and the associated costs of same), doing so is not without significant risk. Dr. Fou acknowledges this in his piece, in which he says, “Of course, the publishers have to be vetted…” and that marketers shouldn’t simply take their word for it.

To this point, I’ve spoken with several publishers in recent months who have a strong understanding about where invalid traffic comes from and who have taken the necessary steps to avoid it. But we’ve interacted with many others who do not yet have such an understanding. One media property whom most would consider to be of the highest quality recently told us that invalid traffic is not preventable by publishers. This is most certainly not an accurate statement, and it was asserted in response to a suspicious traffic inquiry that we raised on behalf of our customers. Of course, when you buy inventory you are also effectively transacting with everyone from whom that publisher has sourced traffic in the past!

Dr. Fou raises several important — and downright horrifying — examples of non-value adding middlemen sucking trust and credibility out of the ecosystem. But turning back the clock to a time when there were no intermediaries facilitating media transactions is not the answer. Lots of things must happen in order to make a transaction possible, safe and effective — including things that can’t be taken for granted, such as data-decorating inventory to make it more addressable, ensuring that the inventory is free of fraud and that the supply source has a right to sell it, and intelligently marrying the right supply with the right demand. These things cost money, and should be considered every bit as much a part of a marketer’s investment in “working media” as the final act of displaying the ad on the web page.

It is in the interest of every legitimate actor in digital media to eliminate Black Holes from the ecosystem. By working together, we can continue to reduce costs and remove points of friction from media transactions, and reach a near future state in which good actor intermediaries are able to compete on their technical capabilities and their ability to drive business outcomes, and not only on their existential legitimacy relative to bad actors.

DataIntelligence

How to Kick Start Your Holiday Campaign this Season

November 30, 2017 — by Laura Carrier0

holiday_2017.jpg

This article originally appeared in MarTech Advisor. 

MediaMath’s VP of strategy and measurement Laura Carrier explores how marketers can ensure their advertising campaigns are timed in accordance with the height of consumer holiday spending.

Holiday season is the largest retail season of the year and as the gift giving traditions get underway, now is the time for advertisers to start detailed planning of how they’ll effectively target and reach holiday shoppers. According to eMarketer, Holiday sales will total $923.15 billion, representing 18.4 percent of US retail sales for the year.

To help advertisers make the most of holiday shopping budgets, we looked for trends in the way our best brands and retailers made the most of this season, including the way they think about timing and key dates, budget, media, targeting and more. Consider the following best practices:

Market to Your Audience Based on Deep Understanding

Marketers know it’s important to understand how audiences demonstrate different shopping behaviors – that’s nothing new. But it’s what marketers do with these insights which matters. Ask yourself, does your ad spend correlate to consumer’s shopping behavior? According to our own analysis, 2016 ad spend lagged behind the time frames consumers expected to do most of their online shopping. Over half of consumers plan to start holiday shopping no later than Black Friday, yet marketers had only spent 25% of their campaign budget by that time last year. This year, make sure to pace your holiday budget before customers do their shopping (while they are researching & planning)!

Get Creative Right 

Knowing who you’re targeting on an individual level, as opposed to different segments of customers or audience groups, will help fine tune your creative and targeting strategies this season. Executing true customer-centric marketing with a single view of the customer will allow marketers to optimize against all marketing touch points. Using this approach, dynamic creative optimization, which updates creative elements on the fly without advertisers having to manually build or modify new assets, will allow for more relevancy in the conversations you have with consumers.

When it comes to optimizing campaigns, marketers should take into account differences consumers shopping habits on key holiday dates when deciding on content. For instance, if you’re marketing to someone who is shopping the weekend before Christmas, getting an item to them as quickly as possible is much more important than the price, e.g. offering free shipping or in-store pick up. On the other hand, if you’re marketing to somebody who is shopping on one of the major one-day sales, like Black Friday, Cyber Monday or any retailer’s one-time sales, content around price would take priority, e.g. Buy one, get one free. Knowing the different types of consumption patterns will help advertisers optimize their Holiday campaigns.

M-commerce Market Grows

On the busiest shopping days of the season, customers are reaching for their phones first. Site traffic is just as likely to come from cellular devices as it is desktop site visitors with 47% of mobile share occurring on Black Friday and 49% of mobile share on Cyber Monday out of all total site traffic by device.

Increasingly, customers are continuing to buy sale items on their phones and check out one-day sales. According to eMarketer, US m-commerce sales will rise by 38% this year, and sales via smart phones will increase by 57.8%. With that in mind, marketers should be adopting an ominchannel approach when making marketing channel decisions.  Consumers are influenced by all of the various different media & channels available to them, so understanding behaviors across devices is becoming even more paramount today than it has historically been.

Online vs. Offline Shopping

The share of eCommerce is growing as 55.6% of US consumers plan on doing most of their holiday shopping online . Marketers will make smarter decisions if they understand the influence of online marketing on offline purchases, without ignoring the fact that offline marketing also influences online purchases.

Online shopping is growing at a faster pace than anything else, now 16.6% in 2017, compared to 3.1% for in-store retail. With Holiday shopping beginning earlier, coupled with the growth of online shopping, it’s important to remember that consumer research and holiday purchase planning is happening a lot earlier, too.To fully market across the customer journey, marketers must speak to consumers online in efforts to influence offline store sales, and measure the impact of those marketing touch points on offline behaviors. This will allow for true customer-level understanding, and ultimately the closing of the loop-optimizing marketing to those consumer behaviors.  As a result, brands and retailers alike will benefit from building out a digital strategy that includes both online and offline presences as one strategy-not as two separate tactics.

Intelligence

51 Artificial Intelligence (AI) Predictions for 2018

November 28, 2017 — by Amarita Bansal0

AI_Forbes.jpg

This article originally appears in Forbes.

It is somewhat safe to predict that AI will continue to be at the top of the hype cycle in 2018. But the following 51 predictions also envision it becoming more practical and useful, automating some jobs and augmenting many others, combining machine learning and big data for fresh insights, with chatbots proliferating in the enterprise.

“Making smart marketing decisions across all customer touchpoints, using all available data, to drive complex business outcomes is a herculean task — and artificial intelligence is an absolute requirement for making it all work. In 2018, we’ll finally start to see AI deliver on the omnichannel promise to make marketing that consumers — and others in the value chain — love. The technology is there — from players like IBM Watson and others — and now is the time to rally the right processes and people to put it in action.” Dan Rosenberg, Chief Strategy Officer, MediaMath

For the full article, click here!

IntelligenceMedia

Bolstering Brand Safety with Contextual Pre-Bid Segments

October 25, 2017 — by John Van Antwerp0

ML_resize.jpg

Seventy-eight percent of marketers report their brand reputation has been harmed in the past by “unintended” ad placement adjacent to inappropriate content, according to a CMO Council survey. In the contextual pre-bid landscape, our partner Grapeshot has focused on brand safety and their unique take on keyword segments — Predicts, which dynamically adapt to the relevant conversation happening on web pages, social and elsewhere. At a time when one of the largest perils of digital advertising is having your ad appear next to offensive or controversial content, Grapeshot’s brand safety segments are sought after by some of the largest advertisers in the industry. Their Predicts segments are a creative adaptation of keyword segments, along with an interesting application of technology, to create a differentiated and useful product in the market.

Today, we’re proud to announce our newest integration with Grapeshot, unlocking the entire Grapeshot portfolio of contextual pre-bid products in our platform to provide more choice and flexibility to our clients. The launch is the culmination of one of the largest integrations the Grapeshot team has performed in two years. Contextual pre-bid segments are available for both web and in-app including:

  • Brand Safety — make sure your ads only run alongside content that is appropriate for your brand
  • Standard Segments — target content based on a static set of keywords defined by the experts at Grapeshot
  • Standard Predicts — target content based on a dynamic set of keywords determined algorithmically and by following the social conversation across the web, defined by the experts at Grapeshot
  • Language — target based on page languages
  • Custom Segments and Custom Predicts — target content based on a static or dynamic set of keywords (respectively) created by you, tailored to your unique marketing needs

We’re excited to have Grapeshot available within our platform and look forward to hearing how it’s improved your targeting experience.

Intelligence

The Crawl, Walk, Run Guide To Audience Suppression

October 19, 2017 — by Amarita Bansal0

header.jpg

This article originally appeared in AdExchanger. 

Brands often have a long list of people they don’t want to show ads to: customers who just bought a product, current subscribers, non-subscribers or a group of people in their CRM database they know won’t qualify or be interested in a product.

To avoid showing ads to people – a technique called audience suppression – brands need tech, talent and a strategy to back up their choices. Audience suppression done right cuts down on media costs and makes customers happy. Done wrong, it can annoy them and send KPIs plummeting.

Here’s what agency and tech experts have learned along the way about how to do audience suppression right.

How has this tactic been used historically?

Direct mail marketers use audience suppression all the time to make sure the same person doesn’t receive too many offers, or to avoid exposing them to a better offer right after they convert to an okay offer. Email marketers have continued this tradition, withholding certain messages from subscribers unlikely to be interested in them. But in other mediums, like TV, marketers haven’t been able to use the tactic – although that’s changing.

My brand is just getting started. What’s the basic way to use audience suppression?

Make sure you aren’t showing ads to users who just bought something online.

Before advancing on to anything else, brands need to use this simple, high-value technique. “Suppressing converted customers from your retargeting campaign is critical to optimizing,” said Serge Del Grosso, SVP of media services at Merkle.

Still, many brands have a hard time even doing that right, continuing to show ads long after an online purchase. “That is the most common holiday complaint from family and friends,” said MediaMath Chief Product Officer Jacob Ross.

I’m crawling. How do I walk?

Brands can also suppress customers who don’t have the right purchase profile to qualify for the prospect pool, advised Del Grosso. And they can suppress prospects that have already been exposed to an ad multiple times – though figuring out just when to give up requires detailed analysis.

Or, instead of not showing ads to users who converted, brands can mix things up by showing users a different offer – like a camping backpack to go with their hiking boots.

Good audience suppression also needs to work across devices.

“We have one customer that sells subscriptions for a TV service,” said LiveRamp CMO Jeff Smith. “Similar to a direct mail marketer, they are always tweaking their offers to acquire the customer. The last thing they want to do is offer that customer they just hit on a web browser with the basic offer [who converted] and give them the advanced offer on the mobile device.”

How do I run?

Smart advertisers will understand the customer life cycle for their products and use that knowledge to inform audience suppression strategies.

“You should not be suppressing audiences in perpetuity until you have built up a healthy picture of how your customer historically interacts with your brands,” said Essence’s Jon Taylor, global director of data strategy. “I certainly wouldn’t suggest suppressing audiences from day one if you are uncertain about when you should bring them back into the targeting mix.”

Brands make audience suppression more effective by bringing together offline and online purchase data. “Offline transactions are as important as ecommerce transactions in looking at what customers to suppress, ” Del Grosso said.

A bank, for example, needs to bring together online and offline databases in order to smartly suppress audiences.

“If you just use online information about your customers, you aren’t taking into consideration all of your customers that already have credit cards, for example, in an offline database,” said Bryan Simkins, technology solutions partner at Transparent Media Partners. “You may be wasting acquisition media dollars on customers who already have a credit card because your systems aren’t connected.”

What will mess up my campaign?

“One of the most common mistakes is the time lag between the conversion and removing them from the prospecting pool and retargeting pool,” Del Grosso said. Campaigns require active management to ensure customers are moving in and out of the right suppression buckets.

Second, brands shouldn’t start suppressing audiences right out of the gate. It’s better to let a campaign run across a broad audience and then start making decisions. That way, traders can look at the average ad frequency before conversion, for example. Otherwise, brands risk removing people who were still on the path to purchase or incorrectly excluding audiences.

Data quality is also a biggie. If brands want to suppress ads to women, and the data is only 50% accurate, they’re wasting money on media and on data, Ross said. And brands might be making incorrect assumptions about their audiences. For example, men may use the product, but maybe women are buying the product as a gift. 

How do I measure the ROI?

Audience suppression can move the needle on almost every brand metric, from CTR to sales. Focusing on the proxy metrics, like clicks, can be an easy way to get a feel for efficacy. But it’s even better if the brand or agency can connect new audience suppression strategies to sales or a cost per acquisition.

“We have a number of clients looking at a CPA [cost per acquisition] on the financial services side,” Del Grosso said. “When the operator starts seeing the cost per order creeping up, that’s a trigger that tells us our costs are going up and we need to refresh our audiences.” The agency will then suppress users that haven’t converted after a high frequency or bring back others that haven’t been served ads in a while.

Brands should also keep qualitative metrics in mind. “The softer aspect is not wanting to make your brand seem creepy because you are following them around with an irrelevant message. That’s part of the ROI too,” said LiveRamp’s Smith.

“You have to think about what the true lifetime value is of exposing users to ads over and over again,” MediaMath’s Ross said, himself a veteran of retargeter Criteo. But making determinations about lifetime value is difficult, and one he thinks is best solved by machine learning algorithms. 

What technology do I need?

Brands definitely need a DSP to optimize quickly. If brands need to activate in multiple places, they may need a separate DMP. If they’re bringing in CRM lists or offline data, a data onboarder may also be in order.

Do I need to hire anyone with special training?

Mastery of DSPs and DMPs is a must, said Essence’s Taylor. “Historically, segmentation skills have existed further back in the campaign planning process, in strategic planning. We need to bring more of that thinking forward into campaign management and data management.” That said, most media planners are adept at doing audience suppression – but analytics experts will help refine strategies and provide more insight.

What’s the future of audience suppression?

Better identification will enable omnichannel audience suppression. Suppressing audiences across multiple channels, like television, out-of-home and digital, is the oasis that agencies and tech companies are plodding toward. There is steady progress in cross-device and cross-channel identification of audiences, which will pave the way for omnichannel audience suppression.

Also, brands want to start doing audience suppression using people-based identifiers, not cookies, which will make it easier for brands to tie their decisions to business outcomes like sales.

Intelligence

Social Advertising is Getting Less Antisocial

October 16, 2017 — by Sara Skrmetti0

iStock-805119394.jpg

Marketers have woken up to the fact that their communication with consumers needs to be consistent across every touch point. There’s just one problem: Some touch points remain resistant to becoming part of an omnichannel solution.

Take social networks. While video and mobile advertising have mostly folded into a 360-degree view of the consumer, social as a channel remains siloed. Mainly, that’s because the largest social networks are walled gardens, meaning they don’t make their data available or their auctions open. This makes everything about campaign execution hard – separate platforms, discontinuous and manual optimization and incomplete pictures of performance and attribution.

This situation won’t last. The market demands that social becomes more of a standard component of an omnichannel solution. In time, the market will win.

Forrester’s view

The latest proof that social is not exceptional came in the Forrester Wave Social Technology Q3 2017 report in August. The report portrays a market divided by small players with DSPs and marketing cloud solution providers waiting in the wings.

Forrester counts 144 vendors, which range from agencies to pure-play tech companies. As the report notes, “most are small fish managing microscopic media budgets.” Most reported less than $80 million in social ad spending running through their platforms in 2016.

Forrester predicts that social adtech won’t be its own category for long. There is already consolidation and although U.S. social ad spending will increase from $4.1 billion in 2016 to a projected $21 billion this year, the amount of ad spend managed by the top players in the space didn’t even double. Where’s the rest of that money going? To the social networks themselves.

Social ad buying has remained immature because the tech involved has remained immature for a long time, aside from the tech provided by the social networks directly. Many social ad companies are managing campaigns on behalf of advertisers, while the rest of the adtech ecosystem has leaned toward self-service. That indicates that many marketers are still figuring social advertising out. As the report notes, DSPs including MediaMath are beginning to offer social ad inventory alongside their RTB inventory. Eventually, social ad buying will be standard in DSPs, and ultimately absorbed into marketing cloud solutions, Forrester predicts.

My take

I couldn’t agree with Forrester more on this subject: Social can’t remain its own category for much longer. Currently, consumers spend about 50 minutes a day on Facebook properties alone out of a total of roughly 12 hours of daily media exposure. Marketers now realize that they need to coordinate their messaging during all of those 12 hours.

That said, the social networks won’t make it easy. It is in their best interest to remain walled gardens for a number of reasons, but this desire to stay separate won’t withstand the market’s demand for an omnichannel solution that includes social from activation to optimization, all the way through to measurement.

Market demand caused Facebook to partner with Visual IQ and the Neustar-owned MarketShare last year to provide third-party attribution. Similarly, they have partnered with a number of firms on third-party viewability measurement. I predict we will see more of this – the market will push walled gardens further toward transparency and independent verification.

I don’t expect the social networks to abandon their walled garden status any time soon. The siloed nature of social advertising will remain a challenge for marketers a bit longer, but I am confident the market will work out a solution that allows social to be one component of a 360-degree view of the customer.

IntelligencePeople

Taking Digital Marketing to the Next Level with AI and Cloud

September 14, 2017 — by MediaMath1

IBM-and-MM-announcement.jpg

By Bob Lord, IBM Chief Digital Officer, and Joe Zawadzki, MediaMath Founder and CEO

Let’s talk about how far digital marketing has come and be proud of that for a moment.

Twenty years ago, it was just an idea. Then, 10 years later, programmatic marketing heralded a necessary move toward a better marketing future fueled by data, powered by technology and driven by math.

Today, advertising can be found across connected screens, all controllable with a touch of a button. Ads get billions of impressions a day, touchable through APIs and UIs, which is something that was impossible just five years ago.

Marketers are breaking down organizational silos where collaboration across brand, agency, tech, media and data is finally seen as not simply necessary, but right. Real-time machine learning is used for more than half of every dollar spent in digital, where a 100-percent programmatic future is on our doorstep. Data-driven marketing is moving from one way to do marketing to the way marketing is done.

Still, the reality is we need to go much further.

If you listened only to the press, you’d hear a cacophonous cry of “fake news, fake traffic, fake metrics.” You hear that the infrastructure that manages the now billions of dollars flowing through digital marketing pipes isn’t up to the task anymore. Pixels, redirects, JavaScript and headers are the stuff of a startup industry, not the foundation for mature marketing at scale. Perhaps worst of all, experiencing that moment when kids see an ad and exclaim, “Ugh, I hate advertising.” This, above all, is a daily reminder that we can and must do more.

If marketers really want to pay off the promise of marketing as an engine of business, the connection of thought and deed—the 3 percent of the gross domestic product that powers the other 97 percent, that enables the free internet, that consumers don’t hate and could even learn to love—to move from rendering banner ads to driving business, they know they need to change.

MediaMath and IBM saw in each other something important: a shared worldview, a desire to do better and the will and capability to make it happen. So we’ve partnered to take the next evolutionary steps together. What does this mean? It means we’ll work to:

  • Develop infrastructure that connects brands, consumers and all of the companies in between in a way that is enterprise-class, open and smart.
  • Infuse AI into real-time marketing decisions across all channels, arming the marketer to do her job better with insights as opposed to reports.
  • Delight the human behind the screen with advertising people don’t just tolerate, but appreciate as entertaining, informative and meaningful.

By providing marketers with a neutral, security-rich computing environment and giving them the ability to maintain ownership of their data through the IBM Cloud, marketers will have the insights they need to deliver the campaigns consumers want.

MediaMath and IBM are building the foundation that makes great marketing that moves at the speed of human beings possible, and we are incredibly excited to see what you make of it.

 

IntelligenceMediaTrends

Market Forces Drive Addressable TV To Become Reality

August 30, 2017 — by John DeFilippis0

TV_addressable.jpg

This article originally appears on MediaPost. 

YouTube, Facebook, Amazon and Netflix all have mountains of data on their users. They are also digital natives that operate on the backbone of the Internet.

For these companies, content delivery is automated, data-driven, smart, fast and on-demand. It is of course no surprise to see those that are ad-supported set their sights on TV ad dollars. For all of them, digital video and originals are at the top of the priority list and this is what will finally drive the industry to broad addressable TV solutions.

For years, addressable TV advertising has been an ideal that was often theory more than reality. Rather than place a television buy that reached tens of millions of households, but would only appeal to a small minority of those households, the addressable buy promises to largely remove the waste.

On its face, the appeal of such a targeting mechanism is clear: efficient marketing meant to drive higher ROI.  In the digital world, if a  campaign was targeted to people in the market for a new car, the TV buy could follow suit instead of relying on broader targeting aimed at everyone who was watching the same TV program at the same time. (Of course, addressable targets by the household meaning if a family of four resides there, no one will be sure if the target consumer will be watching at any given time).

What has made the prospect of addressable TV advertising frustrating for many marketers is that while it is technically achievable, the economics are sometimes difficult to balance.  Meanwhile, tapping data from third-party sources requires extra work for  marketers and it may not always be worth the effort.

But there may be a real shift in the market coming soon. The highly targeted audiences that are and will be available from the media giants mentioned earlier will (eventually) attract marketer budgets that are earmarked for traditional TV — cable, networks, local, etc. Traditional TV, the incumbents, will not passively stand by as the market shifts. The stakes are too high. So this is one of the primary market forces having an impact on change in the traditional TV industry. The other is a change in consumer behavior. Consumers want top quality programming, this is nothing new, but they want it on-demand and on all devices, which is an inherent characteristic of digitally delivered content today. Evolution is inevitable.

As a result, I expect to see a  rise in addressable TV advertising, but it will remain a relatively small slice of the market for the next few years as the industry sorts itself out.

The current state of addressable

At this writing, tens of millions of homes in the U.S. are able to receive addressable TV advertising and some 70 million will be able to get it by 2020.  But just a fraction of homes are actually being served ads this way. According to eMarketer, addressable TV accounted for just 1.3% of the total market in 2016, a figure that the researcher predicts will jump to 3% by 2018.

As those numbers indicate, addressable has been used more for experimentation than a  regular line item on media plans. So far, carmakers have shown the most interest in addressable. For example, Toyota promoted its Prius Prime to 18-49 year-old tech savvy consumers with annual incomes of $75,000-plus last year. Hyundai placed a similar addressable buy to promote its Genesis model, targeting consumers with annual salaries of $100,000 or more.

Marketers, looking to gain more of a 360-degree view of their customers, are also clamoring for more data so they can connect all of their audience buys across all screens. Ultimately, driving outcomes that build a business is the goal and knowing what budget to spend and where to spend requires a holistic planning, execution, measurement, and optimization effort. That said, I don’t expect the addressable TV market to crack wide open too soon. There is much invested in the history of the industry; cable operators, for instance, are likely to keep experimentation on the edges until they see a clear benefit.

But the momentum is moving toward an addressable world. That means that even though addressable TV has been on the table for years and the pace of progress, for some,  has been a source of frustration, marketers may soon start achieving what until now has merely been an idea.

IntelligenceMedia

Solving Ad Fraud This Year

August 25, 2017 — by Kyle Turner0

identity_framework.jpg

Ad fraud has been menacing the digital ad industry for years and has grown beyond critical mass…Though one study by the Bot Baseline report predicts that fraud will fall 10% this year, not everyone agrees. An estimate by Adloox claims that ad fraud will cost the industry $16.4 billion this year, up from $12.4 billion last year.

The report details that nearly 20% of digital ad spending was wasted in 2016 and up to 29% of programmatic spending ran on invalid traffic.

From my vantage point, as head of Inventory Quality at MediaMath, I would agree that this industry problem is bad. In part this is because fraudsters are ingenious about inventing new ways to rob advertisers and steal from the ecosystem. Here are the three biggest trends I’ve seen this year:

1. Counterfeit/”spoofed” domains. In a sleight of hand, fraudsters show premium domains during the bid request (like www.abc.com) but when the impression is served it actually arises from something like www.joesfraudsite.com, which is phony. This may lead to putting the legit domain (www.abc.com) on an advertiser blacklist. One of the tools available to mitigate this issue is ads.txt, a tech tool developed by IAB Labs that helps to identify fake or unauthorized sellers. MediaMath supports ads.txt and is eager to see publisher adoption grow.

2. Invalid traffic. Also known as “non-human traffic,” this is a bit of a misnomer as it can actually be created by humans. Whether generated by bots or people, malicious invalid traffic wastes marketers’ money and undermines confidence in the digital media ecosystem. The scope of such activity is huge; security firm Imperva found that 28.9% of all web traffic last year came from “bad bots.” MediaMath leverages a mix of proprietary techniques and certified third party vendors to combat invalid traffic, and does not tolerate invalid traffic being sent to our platform

3. Brand safety. This isn’t a new issue, but it garnered new attention this year as journalists highlighted instances of brands advertising on brand-unsafe video content, along with an increased awareness of the challenge of “fake news.” Many advertisers have taken steps to limit their exposure to user-generated content and rely on ecosystem partners to enforce brand safety across their campaigns. MediaMath’s Curated Market guarantees marketers that their messaging won’t run on brand-unsafe properties.

What will fraudsters come up with next year? Unfortunately, it’s likely that the cat-and-mouse game between ad tech providers and bad actors will continue. But we hope that this year will represent a turning point in the industry’s efforts to combat fraud. Leveraging the right tools and partners, marketers can help ensure that their ads run in clean, safe, authorized, authentic and well-lit environments.

If you’re attending IAB AdTech & Data Brazil, head down to see our General Manager of Supply, Lewis Rothkopf, speak on Inventory Quality and Brand Safety Tuesday, August 29.