(I’m republishing this post from my Tumblr that wrote last month regarding some of the basics of Ad Networks and Display. I felt that’s a good blog post to kick off the new blog, a blog I’ll use to write about the internet space: social media, online advertising, and lead generation. This will free my Tumblr up for more general posts, usually around sports, stock market, or interesting articles. Feel free to leave me a note!)
Nearly two weeks ago, at the very last minute, I made a trip to Las Vegas to attend LeadsCon, a highly successful conference held by founder and CEO, Jay Weintraub. It’s a conference dedicated to those operating in the lead generation industry. This is my narrative of the knowledge I picked up in the various sessions I attended. I will try to write on a few sessions, one post at a time.
Background on me: I work in the lead generation/online advertising space as a product manager focused on helping my company expand in its product offerings. I have been fortunate enough to also learn the media buying side (distribution) which allows me to further understand the root issues in building out our product offerings.
Ad Networks and Display
Moderator – Rob Leathern (CEO, CPM Advisors)
Panel – Rob Deichert (SVP Publisher Operations & Yield Management, Platform A — AOL), Dilip Dasilva (CEO, Exponential aka Tribal Fusion), Dave Zinman (VP of Network Management, Yahoo!), Murthy Nukala (CEO, Adchemy)
I was particularly interested in this topic. The purpose of this discussion is a simple. Displayed advertising is really hard to make it consistently work. So, some of the biggest display buyers came to share their knowledge. When you’re in the directing marketing & lead generation space where the only thing that matters is backing out the CPM buy to a CPA/CPL (100% performance, 0% brand), the difficulty of making display advertising work consistently is no exception. Some of the many external factors include:
Traffic quality (who you’re buying), traffic segmentation (who you’re targetting), banner performance (banner creative/design), landing page performance (website destination for clicks from the banner), banner to landing page flow (user expectations), conversions/CPA/CPL (clients stay happy with you). On top of that, a media buyer sometimes face resistance from a large network to complete their requests (ie. pausing campaigns) in a timely/immediate manner. This is the case when the buyer does not have an approved 3rd party ad server (extreme inefficiencies).
In the display space, varying and competing eCPM rates rule and very widly from vertical to vertical and suffice to say, “Display is a heavily operational problem.”
Dilip mentioned that for his company to take on a new advertiser, the minimum price/budget would have to do $10K-$20K month tests at a minimum and goal is to reach $100K/month. He emphasized a shared risk approach. While I don’t disagree with him, most ad networks is seemingly always “testing” new placements and with each new tests run, it could take just one bad test to ruin your margin for the month. The next challenge becomes that even if a media buyer do crack the right combination of banner and landing page effectiveness, it’s only a matter of time where you repeat the process of creating the next new banner design — and the routine of testing and managing the success/margin of the campaign happens all over again.
Murthy gave some great pointers (and was aggressive in stating his systematic approach is the way to crack the display advertising problem): he repeatedly mentioned that an ad network consist of four points: inventory, data, algorithms, and sales. He insists it’s extremely difficult for an ad network to be great at more than one or two items and considers no one much superior than the other. Knowing this, he strongly suggests:
- Strategy in buying display media must include buying a portion in exchanges (where it’s based on a competitive bidding system and price points are at the buyer’s control ie. Yahoo!’s Right Media).
- Understand and must consider the end-to-end economics before buying.
- Develop your conversion funnel
- Segment your inventory, understand the price points you are paying, determine what segmentation converts and what doesn’t, and quickly pull what doesn’t work.
Take Away:
Display advertising is extremely difficult and heavily operational. There are plenty of varying factors, so as a media buyer (or product manager), the first step is eliminate any of the varying factors that can be automated. The obvious first step is to used an approved 3rd party display ad server so that it gives the media buyer control the stats in real time and ability to swap out poor performing display creative. In the day to day operations, a conversion funnel must be developed where data needs to be as close to real time as possible is accessable. This helps measure user experience satisfaction. At the higher level, in order to have a chance at longer term success, a reporting system needs to be able to bucket data (behaviorial — segmentation) and display in a meainingful way to the vertical you are buying in. By being able to collect and analyze this data, a media buyer will be able to continually get smarter with the traffic they buy.
