Connexity’s Findings from the Front Lines…

Connexity’s Findings from the Front Lines of Dynamic Creative

dco-targetingDynamic Creative Optimization, or DCO, is a key way that marketers can quickly serve highly customized ads based on consumers’ visits and interest. If you’ve ever looked at a specific product, then later while browsing the web seen that same product or a few similar alternatives in a display advertisement, then you’ve seen DCO in action.

When executed correctly, dynamic creative optimization allows you to put the right message in front of the right customer at the right time. DCO has traditionally been associated with retargeting campaigns. In its simplest form, you need to know at least one product page the user has visited and then use that to fill the ad creative with similar items. For example, if a shopper was looking at an athletic shoe, you can retarget the user with that same shoe and a few other similarly priced and styled shoes as they travel across the web, in order to bring them back to purchase.

When marketers get caught up in the details of DCO, or “best practices”, such as deciding which colors, buttons and fonts will raise conversions, this can cause them to miss the overall picture of why shoppers didn’t convert. What works well for some product categories, such as fashion or electronics, or for certain demographics, may not work well for others. Instead, we’ve found that focusing on segmentation, personalization and optimization based on a depth of data signals will enable more effective targeting over time.

How Segmentation Taught us to Avoid Assumptions

Dynamic Creative Ad ExampleHere’s a fun example of how Connexity stumbled upon a realization that we would never have considered, had it not been for careful segmentation.

For a retail client, we tested retargeted display ads that promoted 3 products against ads that only displayed 1 product. We found that although the 3-product ads were converting well overall, since they provide additional options, the 1-product ads converted better with users ages 65 and up, because they were easier to read—the font and images in 3-product ads were too small for older demographics to see. We may not have considered vision issues when creating our campaign, but listening to the data uncovered that issue for us.

Personalized Scoring is the Future

Beyond demographics like age and gender, there are many data signals marketers can use to classify users. We use thousands of data signals to give personalized scores to individuals, which allow us to better segment and drill down. These scores are based on thousands of data points and very specific for each individual depending on their shopping habits.

Connexity scores individuals on a personal level, based on many factors:

  • Demographic data (gender, age etc.)
  • Over 2,000 different retail categories (purchase, browsing, interest in different types of products)
  • Clickers vs. buyers (are they the type of people who just browse, or do they actually buy?)

Unique Signals Lead to Better Segmentation and Results

In a recent campaign for an electronics brand, we segmented users based on their browsing category (i.e. computers, cell phones, cameras) and search keywords, rather than simply by traditional demographics. We also tracked performance metrics beyond conversions, such as the Average Order Value, Cost of Sale, Click Rate etc. Segmenting by such unique data points uncovered some interesting findings:

  • Search Page retargeting had a much higher Average Order Value with a low Cost of Sale – very promising!
  • Category retargeting was not as effective as search retargeting. Meaning that shoppers who are browsing in a category are not as valuable as those
  • Cart Abandoners have the best Click Rate & lowest Cost of Sale – but also a lower Average Order Value.

Electronic Retailer Campaign DataAnonymized electronic retailer data

The takeaway here is that the future of good dynamic creative involves advanced segmentation based on a rich variety and depth of signals, categories and individual behavior data.

Tracking performance for various segments can also reveal other issues in the purchase process; Connexity tracks campaign performance by device, and in one case we found that although visitor numbers across mobile, tablet and desktop were normal, the conversion rate on tablets was much lower than expected. This helped the product team of the retailer identify a bug in their tablet checkout process.

The Future of Dynamic Creative Optimization

The most common type of retargeting is a method that targets shoppers who visited a product on a retailer’s website, and then tries to bring them back to the same retailer.

We believe that the next generation of dynamic creative optimization is to take a person who has never visited a retailer before, and dynamically serve them an ad for the product they need. By analyzing their personalization scores and which product category they’re browsing at that moment, we can target them with products they are currently interested in. This scoring is based on their visits to other related websites, which are scored and help to categorize them.

Due to our depth of data, Connexity can find the top 5 items a shopper is interested in and serve a dynamic ad to them based on a combination of these items. We can also look at more complex attributes, like whether a person is a bargain or luxury shopper, and show them products that fit their shopper profile.

Let’s say we’re showing an advertisement for an exciting expensive sports car—this might garner clicks from teenagers and people who can’t purchase such an item. Since we have demographic information and data on the type of shopper a person is, such as an older, luxury brand buyer, we can target ads to people who are not just “clickers” and “browsers” but are actually “buyers.”

The advancements in online advertising and data science have opened a new world of options for digital marketers. Simple retargeting and dynamic creative tactics are giving way to more personalized and targeted campaigns, where technology can instantly gauge thousands of data points which we’ve collected or modeled and decide the best type of ad for each specific user.

As the complexity increases and more data is being measured, it becomes paramount to stay competitive by working with a partner who is on the cutting edge of targeting and personalization.