Multi-creative optimization hero image

The New Standard of Creative Testing: Multi-Creative Optimization

By Rita Kaddoum | June 7, 2022

Creatives are the pillar of mobile ad performance. To drive the most successful UA campaigns, advertisers need to continuously refine winning concepts by testing many creative variations and exploring new and emerging trends.

Liftoff has historically supported our customers with A/B testing, setting explicit testing times and budgets to reach statistically significant results. However, A/B testing has its limits.

As advertisers increasingly compete for inventory and users, it’s imperative to utilize a testing solution that can iterate quickly and compare a high volume of creatives simultaneously to maximize conversion rates.

With multi-creative optimization, Liftoff does just that.

Multi-Creative Optimization: A New Approach to Creative Testing

Multi-creative optimization is designed to save our customers time and money by concurrently testing a higher volume of creatives more quickly. With the ability to test up to six creatives at a time, multi-creative optimization expedites creative wins and insights, reduces testing costs, and maximizes ad performance.

With multi-creative optimization, our advanced algorithm automatically shifts spend to your best-performing ads. By reallocating spend in real-time, multi-creative optimization reduces spend on poor-performing creatives and allows you to reinvest spend back into your top-performing ads. This dynamic approach is a significant change from traditional A/B testing, where spend is split evenly between creatives throughout the duration of the test.

Because multi-creative optimization runs tests in live settings and always optimizes spend on the best-performing ads, install costs fall and results remain consistent once the test concludes. Advertisers realize the best conversion rates for each campaign both during and after the creative test. This is not always the case with A/B testing, where creatives in a closed testing environment may behave differently once set live alongside other legacy creatives.

During our closed beta, customers who used multi-creative optimization doubled their testing cadence. By testing more creatives concurrently, some customers were able to test over 150% more creatives than usual.

Multi-creative optimization also produced faster test results. On average, our algorithm capitalized on wins 25% more quickly than A/B testing. It also moved spend away from poor-performing creatives earlier than we saw with A/B testing results.

Because multi-creative optimization reallocates spend away from under-performing ads into your top-performing creatives as soon as possible, the overall cost of creative testing decreases and campaign performance improves. In some cases, the cost of creative testing was reduced by up to 50% compared to A/B testing. These savings accrue with each creative iteration, contributing to better conversion rates over time. By cutting costs and optimizing spend on their best ads, customers saw their ITIs (impression-to-install rate) increase while CPIs and CPAs fell for their UA campaigns.

Peeking under the Hood of Multi-Creative Optimization

The key to these improvements is our new, advanced algorithm. The multi-creative optimization algorithm is fundamentally designed to improve performance and balance exploration with optimization. By shifting spend in real-time, it quickly reduces spend on under-performing creatives and reinvests it back into your top-performing ads. This delivers faster results and optimizes performance.

Multi-creative optimization visual representation

How does this actually work?

First, unlike A/B testing, multi-creative optimization does not attempt to serve the new and legacy creatives with equal impressions. Multi-creative optimization requires less data to validate a winning creative. As a result, less impressions are needed as more certainty is established by the alogrithm, and optimization occurs faster. This is especially true if the creatives being tested are very different from each other and early performance indicators—like install rate—show a clear winner.

How will creatives behave during and after testing, and what – if anything – gives new creatives a fighting chance at success?

New creatives are served alongside existing creatives under real-world conditions, so results during and after the test remain consistent. System controls guarantee that new creatives are not at an unfair disadvantage or overspending compared to their legacy counterparts. In fact, when customers kick off a creative iteration by introducing a new concept, ITIs improve by 40%, CPIs fall 15%, and CPAs decrease by 10% on average compared to A/B testing.

Ultimately, the system gathers data from all creatives until it’s confident that one performs better than others and allocates more spend to the better-performing creative.

So, What’s Next?

Optimizing your creatives is the first step to unlocking higher conversion rates and growing an engaged user base. While A/B testing provides data-driven insights and key learnings to improve UA campaigns, it lacks the efficiency advertisers need to make quick decisions and keep up with market trends. Multi-creative optimization is an agile solution that increases testing volume and speed while cutting costs, allowing advertisers to refine their creative strategies more accurately and set their campaigns up for success.

Interested in taking your creatives for a spin? Contact your account manager to get started, or sign up with Liftoff today!