Neglecting the startup mentality. Increasingly, we’ve seen small ad tech startups struggle to garner the wealth of data necessary to move the needle for their customers. This is one of the reasons we’re seeing fewer ad tech companies get funding today than in the past. The barriers to entry are now focused around scale, and it’s difficult for companies to get that right out of the gate. As more of our industry moves to buying data-driven media, it’ll be more and more difficult to ignore the prospect of partnering with some of the bigger players in our industry.
Taking enough data into account. A common mistake many companies in the RTB and programmatic space make is to buy a “segment” from a data provider and then use a platform (DSP) to buy media for that segment. Segmentation doesn’t take into account some of the other actions a customer has made in the past or is making in real time, which hinders your chances of reaching the target with the right messaging.
For example, if you’re looking for “smartphone intenders” and someone searched for “new iPhone 6+ for sale,” while another searched for “free iPhone,” it’s very likely both people would be included in a segment called “smartphone intenders.” However, the value of each customer is likely to be very different. So while you may have the right amount of scalable data, it won’t be helpful if you lose sight of individual data points.
Weighing the benefits of private marketplaces (PMPs). We’ve seen a lot of PMPs spring up over the last year or so — but they still suffer from lack of scale and high pricing. It makes sense for publishers to charge a premium for a PMP, but too high a price will greatly limit scale. In the end, most platforms are plugged into both private and public marketplaces with algorithms that decide what works best in real time. While PMPs do offer known, high-quality inventory, they won’t be effective until they scale up and price down.
Customizing at scale. One of my favorite oxymorons is “custom advertising at scale.” Essentially, this is the promise of native advertising. All jokes aside, we are super-excited about the performance of native ad units. As ad resizing has become more automated, we’ve been able to take standard units and make them run in non-standard locations. Additionally, there are specialists who can aggregate all the different native opportunities into one location. This is a good sign that soon, native advertising will also be able to work at scale.
Adjusting your sample size.This may be obvious, but it’s easy to make anything work if you reduce it to a small enough sample size. Not only is it important to be able to scale your data, but you need to be testing at a scalable rate too. If your customer base is measured in the tens or hundreds of thousands, than your scale is going to be different from someone whose customer base is measured in the tens of millions or more. Your test campaigns need to run at the same size as your comparison campaigns, or they will look artificially better than they are.
Taking these factors into consideration will provide better, and more effective, results for marketing efforts as we move into the back half of 2015.