Andrew Tu joined The Trade Desk last month as managing director to build the DSP’s (demand-side platform) business in Asia Pacific based out of Singapore.
The Trade Desk went live in Asia one year ago and its role is to support media buyers at agencies and agency trading desks to buy real-time biddable (RTB) inventory in the region.
Its APAC head, Tu has more than 20 years experience in digital marketing from launching the first Ogilvy Interactive office in Hong Kong in the ’90s, as part of the team rolling out Microsoft MSN in Asia, to setting up Brandscreen, an Australian DSP startup in the region.
With so many acronyms yet loose definitions in the display ad tech ecosystem emerging in various markets of Asia Pacific, Tu explains the difference between The Trade Desk and an agency trading desk as well as the challenges for programmatic buying in the region.
Excerpts of the interview below:
ClickZ: How is a DSP different from an agency trading desk?
Andrew Tu: In the purest sense, a DSP is a “tech solutions provider” whose sole purpose is to provide a SaaS tech platform to agencies and their ATD (agency trading desk). More importantly, a DSP should really align themselves with the “demand” side of the equation and be laser-focused on providing the best “buy” tools.
The actual “Agency Trading Desk” is then a centralized internal team within the agencies (usually at the holding company level), which provides programmatic buying services to the holding companies’ “brand” agencies. E.g., Omnicom Media Group’s trading desk is called Accuen. GroupM’s trading desk is Xaxis. These teams, more often than not, use several DSPs in the process of delivering campaigns.
In the context of the Asian ecosystem, a DSP provides a platform that gives these ATD access to multiple exchanges/SSPs, data/audience targeting, algorithm-driven optimization, and attribution modeling capabilities.
In an ideal world, a DSP should be a tech platform whose sole purpose is to “power” the ATD, not to compete by going after advertisers directly. If they do, they are ad networks or using the latest buzz phrase – an independent trading desk (e.g., RocketFuel globally, AdzCentral in Singapore)
CZ: What value does a U.S. DSP bring to advertisers compared to local DSPs?
AT: In today’s global context, it’s all about scale and expertise. End of the day, we’re all working with the same base template. In other words, we all connect and provide inventory across the exact same global exchanges/SSPs, e.g., Google AdX, PubMatic, Rubicon, AppNexus, FBX (Facebook Exchange) (with the exception of Google’s DBM), etc. We all have algorithms and technology that provides the same base services (e.g., site optimization, creative optimization, white lists, black lists, geo-targeting, etc.)
However, having worked across both a local DSP (Brandscreen) and now with The Trade Desk (Global/U.S. DSP) what I noticed as significant differences are the scale and expertise that a global DSP brings to the equation.
1. Scale: Global/U.S. DSPs works on a completely different scale than local DSPs because they must be able to process trillions of requests at any time across multiple geographies, target audiences, etc. A local DSP also must process quite a bit but they’re limited to the region they are working within.
2. Competition drives innovation: At a local DSP, the competition is really only as fierce as yourself. For the global DSP, the competition in the U.S. alone forces the DSP to continue to innovate quickly in order to keep ahead of the fierce competition in the market. As the Asian arm of a global DSP we get to enjoy the benefits of these innovations which, in turn, lifts the performance for our ATD partners in this region because they end up working with best-of-breed solutions.
3. Team expertise: Irrespective of how advanced technology gets, it’s still the human element (e.g., traders) that help lift performance. A local DSP may have a strong team that understands how to manage a campaign in the Singapore context. However, this is working in isolation. A global DSP has teams running campaigns across the world. Internal knowledge sharing is something that we pride ourselves in doing very well. So imagine taking a trader that is an Asian expert and then augmenting that knowledge with learning from New York, London, and Frankfurt. Although it can be argued that each market is different, but the lessons learned can only enhance their expertise. A local DSP most likely will not have the benefit of this level of daily training/knowledge sharing.
A local DSP, at first, may have an advantage over a global DSP in their connections with “local” exchanges. For example, China DSP iPinYou was the first to connect with Sina Ad Exchange in China. A global DSP will get connected, but in this case, a local DSP had an advantage of being the first to connect. Why it is “at first” is because in the long run, the local exchange wants to connect to as many demand partners as they can so eventually they will connect with the U.S. DSP.
CZ: What are the challenges for RTB adoption in Asia?
AT: Some of the operational challenges I see are:
Access to premium inventory: Getting advertisers to expand their digital budgets so that publishers are more likely to release their full inventory instead of putting into RTB what their direct sales team hasn’t sold.
More and more publishers are doing this, so we need to get ahead of the perception that RTB is only about remnant and get to a place where publishers see RTB as their next biggest opportunity. A good example of a “premium” publisher leading the way is the News Corp global ad exchange that just launched recently.
Metrics: Encouraging advertisers/agencies to become more disciplined in terms of measurement frameworks. Advertisers in Australia or Singapore are very good at looking beyond the click as a metric but outside of these more sophisticated markets, CPC and CPM are still the key metrics many advertisers use.
The key here is to get advertisers and their agencies to start looking at CPA/CPL (cost-per-acquisition/cost-per-lead) as a key metric to measure success. Which means that although a campaign’s CPC or CPM is relatively higher, it still got them the best CPA/CPL.
On a more macro level:
Talent: Talent, as in all things digital, is hard to come by, especially now that media buying via RTB is no longer as simple as a PowerPoint plus insertion order. Media buyers are swamped. With the advent of RTB, they now need to find time to understand that buying RTB combines media/consumer consumption understanding as well as the science behind optimization and driving performance.
In short, buyers must now have the capacity to combine both art and science. Talent that’s good at both is hard to come by (on top of just getting digital talent in the first place).
Brand versus performance: Brand versus performance is and will continue to be a major question that advertisers/media buyers must face when looking at using ATD and DSP. At the moment and for the most part, performance is the key benefit for RTB campaigns. However, there will always be a point where performance spend stops.
Brand dollars will always be a larger proportion of the media split. At the moment, in the display world, RTB doesn’t provide an easy way to measure brand impact (e.g., Dynamic Logic), hence tapping into those brand dollars is tough. And, similar to search, RTB performance dollars will hit a point where it will be hard to grow.
The shining light in all of this is that the trading of video in RTB is breaking this barrier. This is especially true where companies like TubeMogul is making the investment to put brand metric tools into place that can help quantify brand impact (e.g., Tube’s BrandPoint planning tool). The Holy Grail behind all of this is that the brand dollars will shift to programmatic faster for video than display.
Tech-enhanced customer experiences have always had a very receptive audience in China, and as a result, VR technologies are being widely embraced by both brands and platforms.
Header bidding is a programmatic technique that allows publishers to offer their inventory through multiple ad exchanges before they serve up ads from their ad server.
All top Chinese retailers, banks and internet companies share mobile data in earning releases. None of the top 10 US retailers do, nor does Google. US banks and Facebook are better.
American Apparel's chief digital officer discussed the future of retail, the importance of delivering value to the consumer, and strategies for an IoT and omnichannel world.