Guide: How to optimize programmatic advertising campaigns
Tips and tricks to improve programmatic performance
With many levers to pull within your demand-side platform, the number of optimizations you can make to improve your programmatic campaign performance can feel overwhelming.
It’s vital to come up with a structured plan beforehand, so you don’t make too many simultaneous changes. This will limit your ability to determine which changes resulted in performance improvements.
Best practices for campaign optimization
After you’ve selected a DSP, determined the audience you want to target, and created your ads and landing page, it’s time to launch your campaigns. Once your campaigns are running, data will begin to flow, and you’ll have the opportunity to start optimizing to drive better performance. However, there are various factors to consider before you begin this process.
You have to collect enough data to make precise optimizations within your campaigns. The amount of data you receive depends on your budget and timeframe for each campaign. Starting optimization before a sufficient amount of data is collected will result in skewed insights and pushes you to make disadvantageous decisions, and may ultimately cause your campaigns to underperform and underdeliver. When making any optimization to the creative or placement, we recommend having at least 2,000 impressions served to that specific placement or creative before considering any changes.
Optimization frequencyYou’ll need to keep in mind how often you make optimizations within your campaigns. Making too many optimizations at once will make it near impossible to determine which change resulted in improving performance. Conversely, you won’t know which optimization caused the decrease in performance either. Our recommendation is to make more extensive changes one-by-one, such as modifying audiences, changing creatives, or altering targeted inventories and placements. Ensure enough data is collected after the change has been made to distinguish which change resulted in performance improvement.
MetricsThere is a multitude of metrics to review to determine where and what kind of optimization is required. These metrics are CTR, CPC, CPA, ROAS, reach, frequency, viewability, and conversion rate. The primary metric you focus your optimizations on should align with what your campaign and marketing goals are.
Optimizing budgets & audiences
As programmatic is auction-based and impression availability changes frequently, budgeting perfectly is nearly impossible. Programmatic campaigns are often granular and consist of multiple insertion orders and several line items. Allocating an optimal budget at any given time is difficult and time-consuming.
Finding a balance between efficient spending and reaching your marketing goals requires constant vigilance, in-depth performance analysis, and making regular changes to budget allocations.
Ambine does this on your behalf by using a proprietary reinforcement learning algorithm that determines optimal line items’ budgets. If Adform was the DSP of your choice, you’re just a few clicks away from transforming your programmatic campaigns to high-performance machines!
As part of your targeting strategy, you’ve most likely found different types of data to use for reaching your target audience. It can be either your 1st party data (website visits, for example), 2nd party data (data from partners), or 3rd party data (data exchanges and vendors). The variety and available options are often nearly overwhelming, and choosing the best possible data to use for targeting is difficult.
If some of your line items use several data sources as targeting, run reports on which of the audiences perform poorly and remove them. Removing audiences might affect the line item’s impression availability and reach, so make sure to amend budgets accordingly. If you’re using Ambine, let the algorithm do the work for you.
Inventory, placement & creative optimization
Optimizing inventories & domains
Inventory optimization should start from the broader picture (exchanges) and gradually go deeper (placements). Evaluating what type of ad exchanges are working helps drive better results in your campaigns. You can create line items targeting specific exchanges and to bid higher on their inventory. Setting up line items with top-performing exchanges will elevate your campaign performance and provide more insights into targeting.
On a more granular level, determining which sites or apps are performing well and cost-effectively is essential to your overall success. It’s vital to take a step back and evaluate placements based on data and performance and not on the type of website or app the ad appeared on.
When you’re running a programmatic campaign, your objective is to reach the user where they’re browsing online. Unless the placement is located within questionable content from a brand safety standpoint, you should only evaluate it based on its overall performance.
When determining what placements to remove, focus on overall cost, cost efficiency, engagement, and quality. Being too rigorous or inflexible may result in a severely diminished availability of impressions and reach. Aim to find a balance between high engagement, tolerable cost-efficiency, and sufficient reach to ensure you’re in the best possible situation to achieve our objectives.
When looking at performance-based ad creative, it’s essential to monitor the ad frequency. Showing the same ad to the same user will often result in ad fatigue and make your campaigns less impactful. Setting frequency caps on all applicable levels (campaign, insertion order & line item) will limit the number of times your audience sees your ads. Doing a relatively simple regression analysis to determine the link between frequency and your primary metric can help you find the saturation point (i.e., the maximum frequency per user before results begin to diminish).
Varying frequency caps between audiences is recommended to ensure the budget is allocated effectively. Remarketing audiences can benefit from a slightly higher frequency as the user has already engaged with your brand. In contrast, larger prospecting audiences should be capped a bit more rigorously to ensure the budget is not overallocated towards this audience.
Another approach is to do A/B testing to determine which approach resonates most with your audience. You can test this based on the message, visuals, or CTA within the ad. It’s vital to change only a single element during each test to determine what drives higher engagement. These tests help you decide how to amend your creatives to keep them fresh and engaging for your audience.
Before you start optimizing your campaigns, ensure you have collected enough data, set up a cadence for optimizations, and determine what metrics you want to optimize.
Once you have enough data, you can begin optimizing your campaigns’ budgets, audiences, creatives, ad exchanges, or placements. Using the techniques outlined above is a great place to start to drive better performance and for an extra boost, consider using Ambine to take your programmatic to the next level!
Liked this post? Share it with your friends