The Ad Ops Podcast

On the podcast this I guess this week we had Daniel Kornblit, who is the GM of Baron’s Media. Baron's is an interesting part of the ad tech world because they are smaller shop, in terms of headcount, but have decided to take a lot of the technology that they would normally pay vendors and to instead build it in-house. This has both dramatically increased their ability to provide services to their clients and cut their internal costs.

Barron's started by building an ad server to allow them more flexibility and lower serving costs than anything they could find in the market. They then blanched often built in SSP and now in the process of building a DSB lab with her clients and advertising partners a full range of internal options.

As anti-tax has become a larger topic industry, with some reports of it exceeding 55% of overall earnings, it was very ancient to talk to company was built a lot of their own technology in-house. it doesn’t sound easy to both hire the right talent to build this type of technology or to invest the time and energy in rolling out, but it sounds like this is a worthwhile use of money if us if your company can afford.

On this podcast Marc Goldberg was kind enough be our guest and answer several fraud related questions on the industry. He is the CEO of a company called Trust Metrics, which builds whitelists for advertiser and help them buy safer programmatic inventory.

Marc raised a very interesting theory, that there is not nearly as much programmatic inventory online as everyone assumes. He is main point is that when you remove the walled gardens, such as Facebook, Youtube, Instagram, then take away porn, then take away fraudulent sites and copyright infringers, what you have left isn’t the inventory levels that everyone assumes.

This certainly holds true for video and extends across many classes of advertising, such as Mobile Web and In-App display. It was really interesting to hear Marc’s throughs on these formats and how he thinks each inventory class functions differently.

Marc talks us through some of the methods his company uses to analyze a publisher site and determine if he things that traffic there is real. One of the great tips that he included was making sure that the domains you’ve set a campaign to run on actually matches the list of domains that your ads served on. Domain spoofing is becoming a larger problem and methods such as this allow you to better understand the inventory that you’re buying.

This week on the podcast we talk with Roy Peleng, who is one of the founders of a company called First Impression. Roy has build the technology that allows a publisher to dynamically test 19 different ad units, along with the monetization partners behind them, to determine what the best format for each page/device type is.

I was initially interested in talking with Roy because his comapny is sitting on a huge data set about how users respond to different ad formats and he is in a place to definitively say what formats work the best in what scenarios.

It turns out, as with many parts of ad tech, general rules of thumb are rare and almost ever ad format is dependent on the content around it. There is no single format that works the best everywhere, but with enough diligence and testing you can determine the format that works best for you.

First Impression, similar to Sortable, is a company that has brining large scale infrastructure and machine learning capabilities to a field that has traditionally be run by humans. I think this new style of ad tech companies are extremely interesting and will make major waves in the next few years as the market changes.

This week on the podcast Shailin Dhar returns to discuss the study that he recently released on advertising fraud. He just released a report that shows how ridiculously easy and profitable it is to make money with bot traffic.

Prior to reading this report, I thought that defrauding an advertiser is at least complicated or took some time. It turns out neither of those are true.

In under a week, Shailin was able to create a fake publisher site, fill it with bot traffic, and make a profit. He did all of this while having the industry standard anti fraud filters installed on the site.

The vast majority of the traffic that he purchased was able to beat MOAT, Double Verify and IAS’s detection platforms, as it was engineered to pass their screening processes.

This week on the podcast Victor Lopez returns to discuss the current trends in programmatic media buying and ad tech. As the industry changes, client needs are also constantly evolving and the media plans that best server brands need to change with them .

Victor has extensive experience running programmatic media buying, and is the Programmatic Media Director at Affiperf, which is a subsidiary of Havas. He has recently been promoted to take over buying across all channels of online media.

One of the interesting trends that we talked about in this episode is that more dollars flowing into programmatic media and coming from all other channels, not just TV or traditional print. Programmatic spending is increasing at the expense of all other media. Clients initially took some time to warm up to the concept of programmatic, but now are adopting it in larger numbers.

This week on the podcast we talked publisher side latency and header bidding with Gabriel DeWitt, who is the VP of Technical Operations at Index Exchange. Gabe was highly recommended to us as the person to break down the complexities of latency and what most publishers can do to speed their site up.

The most frequently talked about concept within latency is the affect that header bidding is having on publisher site. Gabe has an interesting perspective on this and doesn’t think that header bidding has to slow a site down. In many instances it can actually make ads load faster.