Jeremy Hlavacek heads up programmatic advertising sales at The Weather Company, where he is VP of Global Automated Monetization.
I recently spoke to him about innovation in programmatic ad sales, how major weather events like Hurricane Matthew affect business and the challenges posed by ad fraud.
Econsultancy: What proportion of The Weather Company’s ad sales are done programmatically?
Jeremy Hlavacek: There are two sales channels at The Weather Company. We have our direct sales organization and then I run the side of the business where we put our inventory up for sale through online platforms and exchanges.
It’s about half and half at the moment and the trend is towards more programmatic.
E: How do you make sure that your programmatic sales don’t cannibalize direct sales?
JH: We’ve spent a great deal of time looking at yield and pricing across the two channels. From our point of view it doesn’t matter if an advertiser wants to buy through a human being and a handshake or a technology platform as long as the inventory is priced appropriately.
The second part is that both me and the gentleman who runs direct sales report to the Chief Revenue Officer, who can see the performance of both channels and make strategic decisions.
He may want one client to buy through automated platforms and another to buy through direct sales channels, maybe because of the client relationship or the type of buy they want to execute, but both divisions are on equal footing.
E: Some media companies think of using programmatic for unsold inventory. What do you make of that?
JH: That’s an outdated way of looking at the business. At this point we see lots of premium advertisers who want to use automation technology to execute their buys.
As long as the price works with the yield model we have no problem with that.
We are finding that inventory can be worth even more when we sell it through automated platforms. When a buyer uses a DSP [demand side platform] they can get very precise with the impressions they want to buy so they get good return on investment.
E: What data points can you use to target consumers more accurately?
JH: As a weather company we are unique in that we have many first-party data sets, primarily weather and location data.
If a beverage advertiser wants to advertise they can use typical programmatic tools to identify the audience of, say, young men between 25 and 35 on the weekend when they are more likely to drink beer.
We can then tell them when it is hot and sunny in New York City that’s a good time to run Budweiser ads for drinking a beer on the beach or whatever.
Or if it’s cold and stormy in Chicago but it’s football season, it might be a good time to run a beer ad with a different message like “stay inside and watch the game with your friends and enjoy a Budweiser.”
E: Earlier this year the east coast of the United States had to deal with Hurricane Matthew, how does such a major weather event affect your job?
JH: To really understand how publisher yield management works for digital properties a hurricane is a great unusual use case.
In media traditionally lots of planning goes into forecasting inventory, understanding how many impressions you’ll have and selling those out in advance. That works well for our direct sales business.
However, when a hurricane, snow storm, heatwave, tornado or other major weather event hits, it’s not unusual for us to see anywhere from a 200% to a 400% increase in traffic.
It’s almost impossible to forecast that with enough precision to sell directly, so it’s really beneficial to have programmatic as we can immediately put that inventory up for sale in an exchange and monetize it.
E: What is the biggest challenge facing the programmatic industry?
JH: One problem that has arisen in this space is ad fraud. Companies running the exchanges have perhaps been a little bit liberal in terms of who they let into that exchange.
This means advertisers are spending good money on properties that are either very long tail, have non-human traffic or might not have highly viewable ads.
E: How does this affect your own business?
JH: The good news is that advertisers are getting a lot more serious about that issue to the benefit of companies like Weather, which can offer premium inventory at scale without any of the worries of fraud, viewability and non-human traffic.
We can give advertisers what they want: efficiency through automation and targeting through data. We want people to buy real legitimate impressions.
E: How can the problem of ad fraud be solved?
JH: It’s gone on in media for a long time if we’re honest – from the early days of TV and print. That’s why third-party verification companies like Nielsen were invented, to track exactly what was served.
This is just a case of the programmatic industry growing up and recognizing it needs to be held to the same standard as other media.
When it was a new and disruptive emerging industry maybe it didn’t matter as much, but now $20bn+ is being spent in programmatic ad technology it’s not okay to say, “We’re going to let this slide.” The stakes are too high.
E: What are some of the innovations in programmatic?
JH: It really grew out of web display inventory and the targeting technology has been pushed to the limit, but there is a huge opportunity in mobile.
On the web side we have cookie targeting, but that doesn’t exist for mobile in-app inventory.
We know there are huge audiences there, but advertisers haven’t figured out how to use data to target yet, so they are dumping most of their dollars into Facebook or Google.
E: How do you solve the mobile targeting problem?
JH: To me the key is data. Look at web. People used to buy impressions blindly and would use the site as a proxy.
So to reach young men they would probably buy ESPN, but then you would also reach women who may not be in your target so would be wasting your dollars.
Weather.com, the New York Times and other brands also have large audiences of young men. So we need to move away from contextual targeting where companies’ brands represent the audience towards truly defined audience targeting.
E: What’s the future of programmatic?
JH: In 2017 we’re looking at ways to apply our data across all media. Display is great but social advertising, outdoor advertising and TV advertising are very powerful.
There’s no reason why our data shouldn’t be effective across all channels, but it’s early days. The operational systems behind all these media channels are very different from what we do to put an ad on a website.
That’s an area where the industry needs to get smarter. Once you master the ad serving technology, you can then think about targeting programmatically.
For more information on this topic, check out these resources: