Blog | Relevant Digital

How can a publisher do multivariate testing in advertising sales?

Written by Thuy Ho | Feb 2, 2023 1:53:22 PM

"Start testing, stop guessing" is Relevant Digital's motto for 2023 as our experts continue to encourage multivariate testing among publishers and programmatic networks. Although it's not a new concept, it doesn't seem to receive as much attention as it should.

 

So what is multivariate testing?

By definition, multivariate testing is a technique for testing a hypothesis in which multiple variables are modified. The goal of multivariate testing is to determine which combination of variations performs the best out of all of the possible combinations (Optimizely).

Multivariate testing is a critical component of advertising operations that helps publishers make data-driven decisions on investment and strategy. In the challenging economy of 2023, testing is for not only optimisation but also cost saving. 

 

What happens when publishers don't test?

The lack of testing can have significant consequences for businesses, including missed opportunities for growth, ineffective campaigns, waste of resources, low-quality user experiences, and an inability to adapt. Programmatic is a fast-paced market with constant changes, publishers need a proactive approach to testing in order to not fall behind.

 

What can multivariate testing help publishers improve?

The possibility of multivariate testing is limitless. Besides growing ad revenue sustainably, it also minimises loss for publishers by providing accurate data. Below are some key aspects where publishers will see improvement by conducting proper testing:

  • Ad revenue: by testing various metrics like geo, formats, floor prices, and different ad placements, etc, publishers can determine the optimal configurations for their inventory to maximise CTR and obtain high CPMs.
  • User engagement & experience: testing can help publishers optimise the layout, content, and design of their website to ensure user satisfaction and engagement. This results in increased page views, higher viewability, and reduced bounce rates, which drives more ad revenue and reduces the cost of acquisition.
  • Data-driven decision-making: testing provides publishers with data-driven insights into what works and what doesn't, allowing them to make informed decisions with calculated risks.
  • Competitiveness: by continuously improving their website with data from tests, publishers can stay ahead of the competition and maintain their position as a leader in their industry.

 

How to conduct effective multivariate testing?

To make the most out of multivariate testing and avoid wasting resources, these practices should be followed:

  • Define clear goals: define the goals and objectives to determine what elements to test, and what metrics to track. 
  • Choose the right sample size: select an appropriate sample size for the tests, taking into account the website traffic and target audience.
  • Monitor and organise the data: to avoid confusion and accurately interpret results, publishers should have the performance analytics in one place but with different segments/categories.
  • Analyse data and make decisions based on it: analyse the data based on the dimensions and metrics that have been set out to make decisions.
  • Continuously test and optimise: more tests equals more data to securely better the stack's profitability and optimise operations.

 

What challenges do publishers face in multivariate testing and how to overcome them?

Many publishers and programmatic networks face obstacles that prevent them from experimenting. For some, limited resources, technical barriers, lack of knowledge, and data privacy concerns are hard to overcome. 

Understanding this, Relevant Yield's multivariate testing capabilities have been upgraded to eliminate those obstacles so publishers and programmatic networks of all sizes can achieve high ad revenue with a proactive approach.

Using Relevant Yield's straightforward UI, publishers are able to set up multiple tests with different configurations with no coding involved. This minimises the traditional dependence on engineers, which saves time and gives yield specialists the tech freedom to actively test their hypotheses. 

All performance analytics are logged in real-time and can be sliced and diced with various dimensions. HB Analytics builds a holistic picture of how each configuration impacts the earnings. From there, specialists can draw conclusions and make decisions based on solid proof. 

Sometimes, mistakes happen in testing, that's normal! However, to diminish the negative effects, specialists can take advantage of Relevant Yield's customisable alarm systems. If anything goes wrong, the damage will be minimal.

 

While there are valid concerns around resource constraints, technical barriers, and data privacy issues, the benefits of multivariate testing make it a critical focus for the sell-side in 2023. By making informed, data-driven decisions, improving the user experience, increasing ad revenue, and staying ahead of the competition, publishers and networks can position themselves for long-term growth and success.

 

If you wish, you can find out more about HB Manager here or if you have any questions or would like to find out more about the solution book a meeting here.