It’s Time for Publishers To Build Trading Desks of Their Own
Editor’s Note: We want to highlight this post from contributor Tony Mamone from last fall for two reasons: One, because it’s a visionary piece on why and how publishers should employ a “trading desk” approach to manage and optimize paid social distribution of content, emulating the way agencies have used data and technology to more effectively purchase audiences at scale across programmatic channels. The piecemeal approach to content distribution many publishers take doesn’t scale and is draining valuable resources from content, marketing, and traditional audience development teams. Two, because Tony will be speaking at the FUSE Media Summit in October on how his company Granite Media is combining marketing automation and machine learning to sustain a 100% programmatic ad revenue model. Learn more about attending the FUSE Media Summit here. – Denis Wilson, Editor-in-Chief, Publishing Executive
With social media dominating the way readers find and consume content, digital publishers are faced with daunting challenges. Sponsored posts and native promotions are gaining traction and fewer people are visiting website homepages or responding to organic updates and shares. As costly paid promotions gain traction, the long-term future of digital publishing hangs in the balance.
Even for pubs with millions of social media fans, getting new stories in front of readers is a daily challenge. Take Facebook for instance. The average user follows hundreds of friends and brands, all competing for limited space in the news feed. Each story a publisher posts must compete with photos from friends’ weddings, kid birthday parties, or a flood of stories from other news publications. Facebook also makes room for ‘sponsored posts’ – further reducing the reach of free updates. A typical post is only seen by a fraction of your followers, and only a fraction of those will click to your website. If you want to reach more people, then you need to pay to sponsor your story. But publishers aren’t exactly flush with excess profits to pour into content marketing – we need to be able to measure and manage detailed return on investment (ROI) to justify the spend.
Publishers need both expertise and an infrastructure for programmatic buying for audience reach and development. The industry has faced a similar challenge in the past, when agencies started to scale their real-time, bid-based campaigns. The solution was to create “agency trading desks,” or managed service teams that use data and technology to more effectively purchase audiences at scale across programmatic channels.
I believe that major publishers now need similar (but different) capabilities. It’s high time to introduce the idea of a “publisher trading desk” to manage and optimize paid social distribution of content.
So What’s a Publisher Trading Desk?
A Publisher Trading Desk (PTD) will manage content promotion and distribution to attract and engage target audiences on a large-scale. PTDs will become the centralized point of audience development expertise for media companies, combining human capital (a team with industry and analytical expertise) and powerful technology (to visualize the data needed to find audience insights, automate campaign management, predict outcomes, and learn how to optimize towards KPIs). Real-time bidding for social media promotions within a publisher trading desk will get increasingly sophisticated and eventually start to look like the transactions in a stock exchange.
To optimize results and sculpt a target audience, pubs often need to manage multiple campaigns for each individual piece of content. In addition, prices and KPIs fluctuate constantly. Real-time data, prediction, and automation tools are therefore critical to success (no matter how expert your team is). Today, pubs patchwork together various components, but no unified solution yet exists.
- Embrace programmatic best practices to buy social promotion to grow and sculpt audiences.
- Buy in real time, using data science to drive audience, financial, and engagement results
- Connect target audiences with premium content by coordinating editorial and marketing efforts.
Why Not Use an Agency?
I believe that PTDs will start as in-house operations in order to experiment and stay nimble. Trading activity will be in lock step with editorial strategy, workflow, and the content calendar. The editor and the trader will need to see eye-to-eye if they hope to maintain the brand’s voice and style. Also, big publishers will want to limit the number of RevTech and MarTech vendors who are taking a bite out of margin by limiting the number of hops between audience and content. Eventually, managed services may emerge, but initial experimentation and innovation will spark from within.
What’s This ‘Audience Development’ You Speak Of?
We don’t need to reinvent the wheel. Publishers can apply best practices from brand marketing to find and attract readers and viewers to media. We’ll start with three core segments.
- Test and learn how to grow our audiences by promoting content to new readers with many interests and demographic profiles.
- HOW: For each story, video, or new series, automatically try broad test campaigns - then study economic and engagement stats to cull the losers and double down on the winners.
- Identify attributes of your best/target readers and viewers and use those as targeting criteria to find similar audiences
- HOW: Drop pixels to gather data from prospecting campaigns or organic audiences in order to identify interests and demographic profiles, then use those to create targeted campaigns.
- Retargeting Segments
- Now it’s time to bring repeat visitors back to your sites.
- HOW: Those same retargeting pixels can be used to segment your readers within a data management platform (DMP) or directly on Facebook. You can use those segments to target promotions to draw the audience back.
Where Will the Audience Come From?
By charging on a cost-per-click basis for content distribution, big social networks are catalyzing publishers’ initial need for trading desks. Eventually pubs will be able to tap audiences from a broader set of channels. Trading desks can seamlessly connect with multiple supply vendors, SSPs, and social media platforms. Smart use of data and integrated brand-safety and fraud detection vendors, such as WhiteOps, ForensIQ, or Pixelate will also become critical to maintaining quality and avoiding bots and fraud as pubs explore new channels.
For instance, PTDs will buy from:
- Social Media Platforms
- Boosted or promoted posts across social networks such as Facebook, Twitter, and Pinterest can be micro-targeted using pixels, interests, demographics, or other traits.
- Native Networks
- We’ve all experimented with the content recommendation widgets offered by companies like Taboola and Outbrain. But did you know that you can now customize your bids and campaigns for specific downstream supply sources within their networks? Trust me, not all of the websites in these networks are created equally -- a robust whitelist/blocklist is critical for brand-safety. Publishers can also explore programmatic native offerings from services like TripleLift, Sharethrough, or Yahoo’s Gemini.
- Native Display Media
- Open standards already exist to help format headlines, thumbnails, and deks into native ads. It’s early days, but this will grow. At it’s Think Publishing conference, Google recently announced a big push for “AdExchange Native.” AppNexus and other SSPs are also ramping up their native offerings.
“Sculpting” An Audience
Every big site has many audience segments, whether broken out by demographic profile (gender, age, location, HHI, etc.) or behavioral characteristics (interests, context, etc.). Sometimes advertiser demand will spike for a certain segment. Wouldn’t it be nice if you had a trader who could shift gears and try to aggressively grow that segment during peak demand by feeding the information from your DMP into a private programmatic deal to help it fill?
Example: Let’s say that one of your major advertisers wants to set up a programmatic buy targeting young mothers in New York. The CPMs are high and you have a DMP with a custom segment that you can pass to the advertiser as first-party data programmatically. But the segment is too small to fill the demand during the limited flight dates. Your trading desk could jack up the intensity of promotions to help profitably grow that segment during the campaign.
Trading desks succeed when they have a competitive advantage achieved through superior technology and sharp data science. As publishers, we each may have some of the component parts, but it’ll take a concerted effort to pull everything together.
Some of the initial puzzle pieces include:
- A unified data platform with costs, revenues, traffic, and engagement data all in one place
- Real time revenue and incremental revenue per session data
- An integrated DMP or audience segmentation solution connected to programmatic sales
- Campaign creation tools to help launch promotions across multiple supply vendors
- Bid and budget management (ideally with API connections to the big supply vendors like Facebook and Pinterest)
- Algorithms, heuristics, and machine learning tactics to predict outcomes and optimize campaigns
- A user-friendly interface for the traders to set goals and make adjustments
We’ve got our work cut out for us!
Technology alone won’t be enough. Remember: even if you build the fastest fighter jets on the planet, you still need amazing pilots. New roles will emerge that blend editorial intuition with quantitative and technical chops.
For some publishers, the very idea of “traders” promoting content to draw in audiences will feel foreign. But for those organizations where technology and data are already at the heart of what they do, a publisher trading desk will help them achieve measurable business results in an increasingly competitive environment.
Tony Mamone is a co-founder of Granite Media, a start-up focused on using data science and technology to make digital journalism profitable and sustainable. Granite launched its first title, workandmoney.com, earlier this year. Tony is an experienced media entrepreneur. He was co-founder and CEO of Livingly Media (sold to the Axel Springer's aufeminin in 2015) and launched FindArticles.com (sold to CNet in 2007). Tony holds engineering degrees from Dartmouth College and Thayer School of Engineering and a business degree from Stanford Business School.