Blog

This post is continuously updated to reflect how programmatic creative and dynamic creative optimization (DCO) are evolving.

When people think of the relationship of programmatic creative vs. DCO, a common misunderstanding is that dynamic creative optimization and programmatic creative are different technologies.

One term is actually a subset of the other. DCO is a form of programmatic creative.

While DCO falls under the programmatic creative umbrella, but it’s not the only way to build creatives that activate the 6+ billion dollars of data in contained in programmatic media.

To illustrate this idea, let’s start with a definition of programmatic creative.

Programmatic Creative Definition

Programmatic creative refers to the set of advertising technologies that add speed, scale, and automation to the creative process. This covers ad production, dynamic ads, and creative optimization.

Programmatic media has unlocked tremendous potential in how we tell stories in paid media online. The purpose of programmatic creative is to enable data and creativity to come together to tell brand stories in a more resonant and effective way than ever before.

The most often cited use of programmatic creative is to personalize ads to be more relevant to their audiences. Instead of showing one generalized creative to all, it is now possible to segment audiences into groups and show each group a custom creative that is more likely to impact them specifically.

Programmatic creative also enables better message testing and optimization, by both enabling the creation of multiple ads for the purpose of testing, and potentially automating the test to optimize on its own.

So at its essence, programmatic creative allows designers to quickly turn around a high volume of quality ads, match them to specific audiences, and take action based on the results. This is possible largely through computer-enabled workflows, software automation, and big data integration.

Programmatic creative encompasses a broad spectrum of technologies. This includes fully automated dynamic ads. It also covers ad production tech that is purpose-built for programmatic media but where the operator has final control over each creative or optimization.

Now, DCO is quite specialized and very useful for certain situations, like product retargeting, as well as optimizing creatives against audience segments in a data management platform (DMP). However, while every brand needs a programmatic creative technology and strategy to do data-driven marketing, not every brand or campaign needs DCO and truly dynamic ad tactics specifically.

What Technologies Fall Under the Category of Programmatic Creative?

As we explore the technologies that solve creative problems for programmatic media, the terms and approaches that stand out fall into two main categories:

1. Dynamic creative optimization (DCO)
2. Creative management platforms (CMPs)

What is Dynamic Creative?

DCO is defined as a highly automated and rules-driven approach to advertising that actually encompasses two technologies: dynamic creative, and dynamic creative optimization.

A good metaphor to call to mind is that DCO is like an ad factory. This is where dynamic creative comes in. DCO ad units use feeds of data and a set of rules to generate hundreds, thousands or even millions of unique creatives on the fly.

Dynamic creative ads, as a technology, emerged before programmatic media took hold. The original dynamic creative advertisers were dynamically retargeting shoppers with products they had viewed or left abandoned in their digital shopping carts.

As stated earlier, dynamic ads can sync with DMP audience segments to serve a specific creative to each segment. They also typically allow the creative to relate dynamically to the following types of data:

1. Time
2. Device
3. Weather
4. Ad placement / contextual targeting
5. Location
6. Behavioral targeting
7. Demographics
8. Retargeting information.

While this dynamic data is readily available, it also means that the manner in which the creative is made relevant must align with one of these inputs. Additional code is required to process the data in real time. This means DCO ads tend to be slightly heavier overall, slowing load speeds, which can have a negative impact on viewability. The slight drop in viewability is hopefully offset by an even larger jump in overall performance.

Defining The O in Dynamic Creative Optimization

To state the obvious: the O in DCO stands for “optimization”.

Most DCO technologies offer automatic multivariate testing. This testing can relate to traits like button color, ad copy, images, and offers—although these elements must be able to replace each other without requiring any adjustment to the formatting of the ad.

More sophisticated marketers will use automated testing as an opportunity to learn about their audiences, testing different creative messages with each audience to see what resonates.

Think of a hotel brand for example. They might test two ads: one that speaks to the luxury of their accommodates, and another about the value for travelers on a budget. Knowing what message resonates with an audience—luxury vs. value—can be a powerful insight.

These automatically optimizing ad units were traditionally used for large direct response campaigns, where the number of impressions that will be served justify the opportunity cost of testing so many variants, and where ROI can be clearly tracked all the way to the sale. More recently many dynamic advertisers are using DCO to split test a handful of creates against all of their audience segments.

Dynamic Advertiser Setups

Setting up dynamic creative to optimize against DMP segments can be a complex process. How do you choose one segment over another in the decisioning tree? Some organizations answer the question based on their best guess.

However, the best DCOs can use machine learning over the course of their operation for three types of optimization:

1. prioritize high performing audience segments
2. composite segments together for better reach
3. pick the best creatives to pair with each individual segment

Depending on your vendor, setting up DCO ad units can be a complex and technical process. Careful planning is required. So is a reasonable buffer of time.

It’s important to pick a vendor with high marks on service. It is very common for vendors to offer managed services to assist advertisers in getting their campaigns launched. However, the technology is becoming increasing usable in a self-service manner.

Above: DCO ad units can be configured to pull in retargeting data like product catalogs, special offers, and store information. This requires consistency in the formatting of text and image assets for best results. Ad units for high impression campaigns can use machine learning to present multiple offers and optimize for the best combination over time based on attributes of the person viewing the ad.

What are Creative Management Platforms?

Creative management platforms (CMPs) represent an all-encompassing approach to programmatic creative that includes dynamic ads and other static formats.

If DCO is like an ad factory, CMPs are more like an ad factory—plus design power tools. CMPs allow designers to produce large sets of ads at the same time, but also dive into making small changes to each individual creative as needed.

In contrast to the complex technical nature of traditional DCO, the power tools elements of CMPs provide a scalable creative environment that is conducive to creativity. CMPs are focused more on enabling a designer to create and manipulate a lot of creatives all at once, one at a time, or somewhere in between, versus focusing purely on automation. When creating many versions of a creative, there are monotonous or repetitive tasks. CMPs seek to remove as many of those as possible.

Full Programmatic Creative Coverage: CMPs Include DCO

By offering dynamic ads and DCO solutions, as well as design tools that do not require data feed integrations or pre-programmed logic, CMPs enable creative for any purpose, regardless of what data or audiences are being used in the targeting.

File load sizes are lighter than with the non-DCO units because the ad doesn’t need additional code to build itself on the fly. This process also removes the need for heavy upfront planning or use of managed services because the toolset is closely aligned with existing design software in many respects.

Above: CMPs allow a high volume of ads to be created using a base layout. Then individual elements can be customized to suit the design as needed. This is includes modifying traits like image position, text position, and text size. Entire objects can be added or removed from individual creatives as well. This example illustrates offers created from a template with a CMP. The creatives are timely (holiday, winter) as well as tailored (by a specific designer or product type). Slight changes to each variation provide the needed polish to complete the design.

Programmatic Creative Case Studies

The potential uses for programmatic creative are incredibly diverse. To root these technologies firmly in reality, it may be helpful to see some examples and results.

• See how Anheuser-Busch knocked out an ambitious campaign in record time and at a 46% cost savings versus their usual process. Download Anheuser-Busch case study
• Agency Anagram increased sales for their client by 565% while using Thunder to power their programmatic creatives. Download Anagram case study
• HomeAway used programmatic creative to double ad production in Asia. Download HomeAway Asia case study

For more resources like these case studies, plus eBooks and helpful videos, see the Thunder Resources page.

Programmatic Creative: Comparing CMPs and DCO

If you are evaluating programmatic creative technologies, the most important thing is to understand the problem you are trying to solve and how you want to work.

We’ve touched on some of the comparison points between CMPs and DCO already, and in this section we will go into more detail.

It’s worth reiterating the dynamic creative optimization is one of the multiple capabilities in creative management platforms. Other capabilities in a CMP may include:

1. Dynamic ads
2. Mass HTML5 banner versioning
3. Mass image versioning
4. Rich media ad building
5. Video ads
6. Facebook ad production

With this array of possibilities, every brand can use a CMP to improve marketing efficiency across channels, formats, and types of campaigns. Not every campaign needs just DCO specifically.

What Skills are Required?

These emerging programmatic creative technologies can require new combinations of disciplines to operate. To incorporate data into digital advertising requires adding a bit of technical complexity into a creative production workflow that hasn’t changed fundamentally in many years.

However, it’s necessary to maintain a level of creativity and design polish in the process. Advertising, even when assisted by programmatic technologies, is a creative and visual trade after all.

Likewise, just about all programmatic creative tech will benefit from a strong grasp of how the dynamic advertiser’s target market can be segmented and what data is available for those segments.

Typically demand side platforms (DSPs) and data management platforms (DMPs) will be used for programmatic targeting. Programmatic creative can be helpful with non-programmatic media as well.

That combination of creative and audience skills is enough to make use of a creative management platform. If dynamic creative optimization is going to be used, additional technical and coding skills are generally required. You must link the data feeds into the ads and set up the logic that will drive the campaign. It can be smart to use managed services for the initial setup of a DCO ad unit, and then to transition to a self-service model to continue to operate the unit over time.

DCO Trafficking and Logic Setup for Dynamic Creative Advertisers

To configure the automatic decisioning in DCO ad units, the dynamic advertiser must configure two places:

1. the data source (frequently a DSP or DMP)
2. the logic in the ad unit itself.

Each input maps to an output, with rules created about what to do with the information.

Here is a plain English example of the setup of a hypothetical DCO campaign:

The DSP is configured to target audiences for the campaign that include 16 targeting line items. For each impression, the DSP will use the viewer’s real time geolocation to determine their local temperature. It will then pass that information to the ad unit.

In the DCO unit, the following logic is configured: When the temperature is below 64 degrees, show the creative with a picture of a hot coffee. Otherwise, show it with a picture of a cold smoothie.

This type of setup requires that there is a way to pass the needed information into the ad unit. This can potentially prohibit or restrict what is possible depending on what another ad tech is involved. In the event that conflicting information is passed, or data is not available for a particular impression, a “default” occurs in the DCO unit and a generic creative is served.

CMP Campaign Setup

CMPs also have the option to create and deploy many individual ads. Ad server integrations, like with DoubleClick Campaign Manager (DCM) make this process fairly painless. A discrete creative is produced in the CMP and provided for each placement or segment.

Here’s a similar example for setting up a campaign with creatives from a CMP:

The DSP is configured to target 16 segments. Segments include audience differences like age 18-25, age 26-35, loyalty program members, indie rock listeners, pop music listeners, hip-hop music listeners, etc.

Using the CMP’s streamlined workflow, 16 similar but unique creatives are produced. Each has messaging and imagery personalized to matter more to that segment. If elements need to be added or removed, or if layouts need to change to accommodate a particular creative, those changes are made. Ad tags or HTML ZIPs are trafficked via the DSP in each segment.

This process removes the possibility of data and decisioning conflicts, so CMPs are not capable of defaulting in the way that DCO ad units are vulnerable.

A Comparison of Programmatic Mindsets

The nature of a technology tends to shape how we use it, and programmatic creative is no different.

When using traditional DCO, the amount of know-how and setup required makes it a better fit for organizations that carefully plan campaigns months in advance.

These kinds of optimized creatives require the entire experiment to be set up beforehand. While it is theoretically possible to update an in-market DCO ad, because of the complexity of the technology, these updates tend to happen only when absolutely necessary.

In contrast, the flexible nature of working with a CMP allows it to be used in a wide range of work styles. The speed at which creatives are built in a CMP makes it ideal for teams that want to adopt a more agile approach to test concepts and learn about their audience more easily.

Updating in-market ads without re-trafficking is very simple. Creative optimization with a CMP is a more human-driven process. Inquisitive creatives can bring their ideas to life, see how consumers interact with them in the marketplace, and then adapt without delay.

Example Programmatic Creative Use Cases

To bring the comparison together, here are some of the more common considerations that advertisers will take into account when adopting programmatic creative or using dynamic ads.

Creative management platforms can help you if…

…You want to optimize creative and media together

Example: You are using a single DSP or DSP/DMP hybrid that can optimize both the creative and the media (e.g. MediaMath)

Why? By loading creative objects separately into the DSP, the media system can optimize creatives individually against placements it is buying. A DCO ad, on the other hand, isn’t privy to all of the information in the DSP and is thus optimizing independently. Allowing one system to optimize all variables together provides the greatest possible efficiency.

…You care about editing each ad version

Example: You have creative in many ad sizes as well as multiple Asian and Roman languages. Because the translations result in phrases and words of such different size and length.

Why? A good CMP lets you mass version and edit each version. Some manual adjustment is necessary to protect the quality of the creative. With the CMP you can preview and adjust images and text to your exact specifications.

…You want to build more creative concepts to test and personalize

Example: You want to try a new layout or template

Why? It’s far easier in a CMP to manipulate both the template and assets to produce new versions. Traditional DCO ads are more tedious to edit or build anew.

Or,

…You want to sustain a high bar of creativity in your programmatic creative executions

…You want DCO, but also have other creative needs

…You are looking for a way to build ads in sets of multiple sizes (300×250, 300×600, 320×50, etc) all at the same time. You would even enjoy being able to work on multiples of these sets at the same time.

…If you do want to localize your advertising, you want more customization than changing an address or city name

Dynamic creative optimization (DCO) can help you if…

…You want to optimize the creative independent of media variables.

Example: You have multiple demand-side platforms but want to set up the creative once.

Why? DCO runs its own rules separate from media. Because of this, it is a (mostly) single setup that can be run across all DSPs. Individually mass-versioned ads from a CMP are trafficked into each DSP separately.

…You don’t need to edit each version

Example: Product retargeting ads that use a catalog of interchangeable SKUs

Wh? DCO templates that render on-the-fly are rigid. You can’t insert vastly different types of content because you can’t adjust the layout for each version. You can only swap in and out elements like text and images. Retargeting ads use very standardized data, so inserting it into a dynamic template isn’t a problem. To support thousands of products, it is very efficient.

…You want to run a multivariate test of every creative permutation

Example: Test multiple background colors with multiple featured images and multiple button colors and calls to action (e.g. 4 x 4 x 4 x 4 = 256 total versions)

Why? Once you have set up your rules (just like your assembly line in a factory) the ad unit can cover all possibilities that a human may not want to build

Research: A Programmatic Creative Report [Video]

We worked with Digiday to survey the industry on important questions on programmatic creative. This quick video summarizes the results.

Taking the Next Step

If you found this article useful, you can find additional information in the eBook The Essential Guide to Programmatic Creative Technologies. The book includes example campaigns, workflow explanations, and a section for frequently asked questions.