Unlocking Insights with Paid Media Testing
We dive into how to use split-testing on performance marketing to uncover insights on your audience and validate acquisition initiatives.
While many marketing leads and growth PMs are familiar with using paid media to drive immediate results, few of them understand its true potential as a tool for testing and learning. Consider the following examples:
On Google Ads, you can select what keyword / search terms you’d like to display ads on. This enables you to measure the demand for those keywords (with search volume) and their value (with conversion rate + ARPU).
On Meta, you can select who (existing or new users) to reach and what to say to them (with your creative asset and copy). This enables you to understand how different messaging resonates with different audiences.
In this guide, we'll dive into the power of paid media testing beyond surface-level campaign improvements, uncovering its potential for audience insight and acquisition validation. This is particularly relevant for growth PMs and heads of marketing looking for new data points that can inform their marketing roadmap.
Subscribe to receive guides and articles on marketing + data + analytics:
The Different Benefits of Testing on Paid Media
You might find yourself wondering, "But why test on paid media and not just do CRO on my website?" The answer lies in the unique advantages it brings to your overall marketing strategy.
1. Paid media testing offers unparalleled control over your audience targeting
Unlike organic reach, where you have less control over who encounters your message, with paid media, you can finely tune who gets served with your ads. This includes targeting specific demographics or even excluding current users to focus solely on reaching new audiences.
2. You can test on paid media before you have the volume to test on your website
This reach-on-demand capability of paid media can enable you to start A/B testing even before your product has enough demand to reach statistical significance. Mike Fiorillo, founder of CRO agency flightdeck, recommends only start testing on your landing page once you have 60k monthly visitors.
3. Paid media provides a pathway to test and validate your messaging independent of the product or product marketing team.
With the ability to develop landing pages separate from the actual product, you can bypass potential bottlenecks and maintain a focused approach to testing.
Once you have validated hypotheses with paid media, you can feed those valuable insights to the product and product marketing team.
4. Paid media offers a unique environment to test significant hypotheses with relatively low-effort assets.
You can experiment with variations in static creative imagery or copy to identify what resonates best with your target audience. Multiple versions can be designed and tested in a short time.
The different insights that can be achieved with paid media testing
1. Messaging: Communicate your product's value effectively to your target audience.
Understanding how to communicate your product's value to new users can drastically improve conversion rates. Using paid media, you can test various messages with a new audience, providing insights into what resonates most effectively.
At HER, a dating app for LGBTQ women, we discovered with Meta ads that copy related to making friends and networking converted at a much higher rate than talking about dating or relationships. For a pre-revenue company with WAU as their north star, this shaped how we talked about the product across different channels.
2. Onboarding: Start customer onboarding from impression #1
Similar to messaging, understanding what needs to be communicated for users to appreciate your product's value is crucial. We ask questions like: What information drives new users further down the funnel? How can we motivate trial users to upgrade, or app downloaders to create an account?
Pariti was a personal finance management app, acquired by Tandem Bank, which enabled users to connect to their different bank accounts to centralise their earnings and expenses. Connecting to accounts was a big drop-off in the product. When we didn’t mention this need in the ads, CPI (cost of install) would react positively, but activation rates would plummet.
3. Pricing: Find the optimal pricing by balancing cost of acquisition and customer lifetime value.
By using paid media to test different pricing strategies, such as discounts, trials, and bundles, we can make informed decisions about how conversion rates and LTV interact under different scenarios.
A Black Friday campaign, for example, can help you understand how cheaper pricing affects conversion rate, cost of acquisition and customer churn. The goal is to find the optimal ROAS (return on ad spend).
Validation: Test resource-intensive marketing activities before doubling down
Paid media can also be a powerful tool for validating strategies that are resource-intensive to execute. The best example: before investing in content writing, programmatic pages or building tools for SEO, keywords can be validated with paid search.
Keyword research alone won’t tell you what conversion rate or LTV to expect from different search terms. And, although tools like Ahrefs or SEMRush are incredibly valuable, but they can’t accurately predict search volume or ad competition.
At The DataFace, we saw that even though Ahrefs predicted less than 10 monthly impressions for high intent keywords like “data visualisation agency”, this wasn’t the reality. In the end, targeting these high intent keywords with organic and paid activities became a core source of leads.
Thanks, and future articles!
Thanks for reading this. I hope I’ve helped showcase how the power of paid media extends beyond immediate results. If not, then you will probably not care about my future articles either.
BUT — if I did — I have other two guides in the oven: Measuring incrementality of different marketing activities & How to Audit perfomance marketing campaigns, from tracking to creatives.
If you’d like to receive them once they’re published, subscribe below:
Enter your email to receive my future articles on marketing and data: