Quality of data

The very first factor is the Quality of data. It is really important for us as marketers to identify the persona of our consumers. We need to have a clear understanding of the products and/or services that our customers are interested in, their intentions when they go online, and different behaviors that they conduct when looking to purchase a product and/or a service. This type of information goes above and beyond the demographic details such as age or gender. To explain this even further I’ll give you an example – if an advertiser wants to reach mothers living in Toronto and they define their targeting audience to be women age 24-35, there is a huge chance that a big chunk of their marketing efforts is going to fail. Why? Because all my girl-friends, including myself, fall within the same age range and none of us have kids. The more effective way is to describe a mother’s typical online behavior – for example, someone who searches for diapers, baby clothing and/or kids products.

Mediative has conducted extensive research and discovered that by targeting qualified users, advertisers can generate 75% better results.

Quantity of Data

The second pillar is the Quantity of data that we use. Once we identify our users’ interests, we have to add as many of those similar minded users as possible to the targeting pool. Because of the dynamic nature of RTB platforms and the competitive market of ad exchanges, the more inventory you have access to, the more chances you have of winning that inventory. To explain this further, imagine you have a dart board, but only one dart – your chances of hitting the bull’s-eye with that one dart is low. However, if you have more darts, the possibility of hitting the bull’s-eye is much higher. This also speaks to the same notion of the Probability of Large Numbers which explains that the larger set of numbers you have, the more probable for that event to occur again.

A comprehensive research study by Mediative shows that targeting capabilities increase by 141% when utilizing four different data providers versus only one. Furthermore, based on the objective of the campaign, different data providers could offer a more diverse set of behaviors.

Execution

Last but certainly not least, comes the Execution. Given the numerous options that are available through the exchange – such as frequency and/or recency of the cookie, bidding strategy etc. – there are many different tactics that can be applied to optimize the campaign. Despite content algorithms, we should not discount the human interaction when it comes to optimization. Moreover, frequent monitoring enables us to remove the non-performing data points and replace them with the better performing segments and/or find new related users.

Mediative implements a strategic optimization technique by frequent monitoring. Once the campaign is launched,  based on the advertisers’ goal,  marketers adjust the recency of the cookies, which is set to 30 days by default. For those campaigns that are meant to reach in-market users, we reduce the life of the cookie, and for those that are time agnostic, we increase the time frame to be able to reach as many potential customers as possible. Campaign experts also realized that there is an optimal number of impressions which drives users’ attention. While this number could be higher for certain campaigns, it could also be reduced to three impressions per user for others. The bidding strategy is also highly correlated to the success of the delivery, and the optimal bid is greatly dependent on external variables such as seasonality and the competitive marketing landscape. Consequently, the finest bidding methods can be calculated after carefully reviewing campaigns’ results after two weeks. All in all, the performance of the campaign can be enhanced by 84% through continuous observation and optimization.

It is worth mentioning that this should be an on-going process. Once the campaign is launched, the process would be identifying the qualified users, and then adding a large amount of those audiences. In the execution step, if we notify something needs to be replaced or shut down, we would go back to the first step and classify another set of qualified users.


About the Author
Niloufar Alaghehband is Mediative’s Audience Targeting, Product Manager. With experience in digital advertising, Niloufar seeks to provide insightful solutions for advertisers across the Canadian market with best in class audience targeting approaches. Niloufar is a strong believer in data advertising and is constantly studying various tools, platforms and technologies to ensure advertisers’ message is directed towards the right audience, at the right time in the brand-safe environment. Having completed MSc degrees in Engineering as well as Management, Niloufar is equipped with extensive analytical skills and a good understanding of business trends.