Marketers are always expected to improve on results from previous years, but we are not often given the additional resources to make that happen. Not to sound cliched, but we are being asked to do more with less.
Marketers are resourceful, and we usually find a way to meet the c-suites demands. But not with sacrificing in other areas. This is especially true for paid media. Our budgets are being cut, or if they remain the same, the rising costs of advertising combined with increased competition means that we are spending more but getting less.
But what if we could reverse this trend?
Through experimentation, we can apply continuous improvements to our paid media, resulting in a higher return on investment. Doing so not only enables us to deliver on what the c-suite wants but also ensures our methodology remains strong, compliant and competitive year on year.
How do we turn experiments into continuous improvement?
The process of experimentation that we learned in grade school can be applied to marketing. By testing specific ideas and theories, we can better understand how to best reach our target audience.
The purpose of an experiment is to test a hypotheses and draw a conclusion from it. With a bit of pre-planning, we can set up deliberate experiments that test a specific hypothesis. These results allow us to gain actionable insights into our audience and the best way to reach them. Continued experimentation, and the application of actionable insights, builds off of each other and compounds our results leading to ongoing, continuous improvement in our marketing efforts.
Let’s say a software company want to know what type of message best resonates with their target audience. By testing two different versions of direct response creative – with each featuring a different message – it is possible to determine which is more effective. Once the experiment has returned statistically significant results – meaning they were reproducible and not just a fluke – the winning creative can be updated leading to an immediate increase in conversions. In addition to this uptick in conversions, the marketing department can take the information gathered on the audience’s preferred message, and can update the persona and customer journey accordingly. Insights gleaned from this one experiment can lead to improvements across multiple mediums and targeting methods, as well as future creative and content development.
The following sections explore where, what and how to set up an experiment that will produce actionable insights, leading to continuous improvement that can be applied not only to ads but also across the organization.
Where can we test?
There are two types of media placements: awareness and direct response.
Awareness placements are designed to create a familiarity with the brand, building trust and affinity through ad frequency. Because these ads are focused on a single message and do not include a call to action, they are not ideal for testing.
Direct response placements, on the other hand, include a specific call to action that is specifically designed to elicit a response. It is this ability to measure response that makes direct response placements ideal for experimentation. If the call to action is strong, it will work for digital ads and traditional advertising avenues.
This doesn’t mean that we cannot improve brand awareness ads. While it is nearly impossible to generate results without a clear call to action, we can apply what we learn in the direct response experiment to brand ads. Based on what we tested in the direct response experiment, we can make improvements to the efficiency of placements and targeting, and engagement with creative.
What can we test?
Here are 7 ways that direct response ads can be used to set up an experiment to determine the most cost-effective path to our target audience.
1. Placement
Setting up an experiment that tests the performance of individual publications, social networks or different types of placements, (e.g. direct vs. programmatic vs. an ad network), can help determine which type of buy is best for reaching a particular audience. Remember, the cheapest audience is not always the right audience, so pay close attention to engagement rates and cost per engagement.
2. Data Source
Experiments that pit third-party data providers against each other or first-party data against third-party data can be helpful in determining not only the quality of an audience, but also the most effective ways to reach that audience.
3. Demographics
Demographic data, like age, gender, sex, education, and income offer a unique and telling way to learn more about our target audience. By setting up an experiment that measures the results of one demographic against another may result in insights that have a significant and lasting impact on our overall marketing strategy.
4. Targeting Method
Another fruitful way to segment our audience is to set up an experiment that tests different targeting methods, such as contextual vs. behavioral vs. lookalike vs. direct placement. Comparing these methods via an experiment helps us to identify the most cost-effective way to reach our audience.
5. Ad Sizes
Experimenting with ad sizes is another way to improve the performance of our media spend. Different ad sizes will perform in different ways, based on the individual targeting and creative. The targeting method and even the industry we are targeting can have an impact on different ad sizes reaching the target audience.
6. Ad Medium
The type of ad we place, static, animated, video, or interactive, could also impact our ability to reach our target audience. There is no clear-cut answer as to which is better, as it will always depend on the individual program’s objectives, messaging, budget… and of course the design chops of the person creating the ad.
7. Messaging
Last but certainly not least is the message. Experimenting with messaging allows us to use quantitative data to determine what message resonates best with our target audience. As with many of the experimentation ideas outlined above, messaging can also be combined to result in even greater insights into understanding how to reach our target audience.
How do we test?
While how we test depends on what we are testing, the end goal for both is the same: to determine if a small change can can impact the overall outcome.
When talking about creative, A/B (or split) tests are ideal as they quickly provide insights about our target audience. By focusing on two, sometimes three, different versions, a split test allows us to evenly split traffic between each version. Many email marketing platforms have A/B testing built in, making it easy to set up and run: for landing pages and other creative tests, platforms like Convert are easy to use and fairly inexpensive.
A multivariate test is a better option when you want to test more than one variable. While you will lose some of the control you have with a more straightforward split test, multivariate testing can be helpful for testing multiple variations. For example, if we had 10 third-party audience lists and wanted to know which one was best, multivariate testing would be the most efficient way to find an answer as we can test the lists at the same time .
In developing experiments, and applying the insights gathered, we can make continuous improvement to our paid media. This improvement naturally results in achieving a higher ROI, keeping the c-suite happy and thereby validating and reinforcing the value that marketing adds to the organization.
Leave a Reply