Within three years, I can see the death of programmatic display advertising, following the steps of the miserable demise of ad networks.
You may be surprised that as a digital media person, I would make such a counterintuitive statement, just when everyone else has been talking about the rise of programmatic display and how it revolutionises and changes how marketers can reach their audience with precision, which in turn, improves media efficiency.
Unfortunately, with the push from agencies and vendors on programmatic display buying, the sudden surge of demand has not been properly addressed by experienced programmatic display practitioners and data scientists, who actually know how to make the best use of trading desks, DSPs, and DMPs. What we are good at is throwing out more jargon and acronyms in presentations. This in turn creates unattainable expectations for marketers that programmatic is something that they can just turn on the next day, without detailed planning—and that they should expect to see the performance flying through the roof compared with the traditional premium website placements.
Programmatic display is as much an art and science as search. While the technology platforms themselves can help take care of the ongoing bid adjustments and optimisation, it is the optimisers that work behind the scenes, managing the platforms, that matter the most. Data and analytics has become an inevitable part of each campaign's planning, rather than something that you retrospectively go back to leverage when you analyse the performance data after the campaign is launched.
How do you avoid falling into the trap of running programmatic display advertising blindly and ending up digging a hole for yourself? How do you avoid proclaiming in a year’s time that you tried and are disappointed by the performance, only to revert back to the old days of buying premium ad inventory?
To make it simple, here are the five handy questions that you should ask yourself before kickstarting any programmatic display campaign:
1) What is the objective of running programmatic display?
Are you leveraging it to drive acquisition through remarketing? Running predictive/lookalike modeling to reach out to potential audience based on the profile of existing customers? Or trying to reach out to an untapped audience pool that has not been exposed to your products? Programmatic display is not just for performance marketers for acquisition campaigns. It can be an effective tool to run your branding campaigns, re-engaging with converted customers by controlling your ad exposure to your target audience.
2) How does your service provider plan to collect, utilise, and optimise your campaigns based on the data collected on your website, as well as any other internal CRM data that you may have?
If the response that you get is simply an ad tagging solution based on placing the same DoubleClick or Sizmek tag throughout the entire website or placing a tag on the confirmation page to exclude the completed buyers, you are obviously working with the wrong partner.
The beauty of programmatic display is that you can craft your audience segments based on your own first-party data, as well as second- and third-party data that you can leverage from data providers such as Eyeota. Major publishers including Facebook have taken a very active approach in merging an advertiser’s internal CRM data through the push of custom audience adoption. Linkedin has launched a very similar solution in APAC recently, which allows you to identify audience in the Linkedin platform thorough the placement of a pixel on an advertiser’s website. Even from an agency perspective, Xaxis, which is part of GroupM, has invested hundreds of millions to build Turbine, a Data Management Platform (DMP) platform, to draw insights into the audience data collected on the website for clients.
3) What is the lead time needed to launch a programmatic display campaign?
If someone says that they can launch a programmatic campaign within a week, I would cast a doubt about the campaign effectiveness. Normally, it is recommended to place the pixels at least 30 days in advance on a site to start collecting the cookie pool, allowing you to analyse and understand the online behaviour and profile of your existing website visitors.
This allows you to uncover behaviour traits of your online customers that you normally don’t get if you simply look at your web analytics data, even if you take the extra step to modify your Google Analytics code to learn about the interest and demographic of your existing site visitors.
4) How do you ensure brand safety and viewability, and safeguard the campaign against click fraud?
If you look at the development of programmatic display in US and Europe, IAB, Google and other partners have taken a step to further tackle issues such as click fraud, viewability and brand safety. Technology providers Integral Ad Science and AdVerify are playing a much more important role in this landscape. However, similar to any other technology platform, would you be willing to pay the additional technology fees to protect your campaigns against undesired outcomes?
5) How should you evaluate the success of your programmatic display campaign?
If you are still evaluating the success of your campaign by CPC and CPM, then programmatic display is not for you. Remember, programmatic display allows you to zoom in and target a specific audience segment for your product, rather than trying to buy the cheapest inventory and run your ads randomly across all sites. The technology, the science and the brain power behind programmatic buying allows you to reduce wastage and effectively help you increase the reach of your targeted audience.
In short, I challenge you: Instead of jumping on the bandwagon to do programmatic display, take the time to work with your agency partners or technology providers. See how they develop their master plan and roadmap before launching your programmatic campaign, whether it’s for branding, performance or acquisition.
Only if you do that can you ensure that you have the full understanding of how programmatic fits into the grand scheme of your marketing plan and how you can take the learnings from data and translate them into your overall marketing strategy.
Antony Yiu is head of search and performance at MEC Asia Pacific