Gemini hits 1 Billion and counting: what this means for performance marketers.
Gemini has hit 1 billion users - but beyond the numbers, what does this mean for marketing leaders? Mike Sharp…
By Amy Hind, Experimentation Consultant, Launch
Back in 2022, fresh out of designing kitchens and stepping into my very first agency job (bright-eyed, bushy-tailed, and rocking a mullet), hearing that my new career path was already dying did not fill me with confidence. The narrative at industry conferences then, as it is now, was that AI was taking over and conversion rate optimisation was finished.
Four years on, experimentation is very much alive. Marketers have simply trapped it on website checkouts and lead forms.
Meanwhile, performance marketing teams across scaling DTC brands and high-volume lead generation businesses are hitting the exact same wall. Ad spend keeps climbing, traffic is flowing, but revenue or lead volume flatlines. We call this the Performance Plateau.
Reframing Experimentation Across the Funnel
Google defines CRO as the systematic process of increasing the percentage of website or app visitors who complete a desired action. That definition is fine, but it holds you to an expectation that testing belongs strictly on-site.
Nothing limits testing logic to web pages. Applying structured experimentation to Meta ad creative, Google Search copy, video, and post-click landing pages transforms testing into an engine for commercial growth.
Within Launch’s Connected Performance model, full-funnel experimentation bridges two vital drivers: Creative Effectiveness and Website Conversion. Fixing friction early in the buyer journey improves top-of-funnel efficiency and protects bottom-of-funnel conversion rates.
The 4-Step Testing Framework
To scale experimentation across DTC and lead generation channels, teams need a repeatable process that turns audience friction into commercial gains:
- Research and Insights: Gather qualitative and quantitative data through review mining, sales call recordings, usability testing, heatmaps, and analytics.
- Hypothesis Formulation: Define the specific point of friction, propose a direct solution, and select target metrics such as Cost Per Acquisition (CPA) or Cost Per Qualified Lead (CPQL).
- Controlled Testing: Run A/B variations across Meta Ads Manager, Google Search campaigns, or dedicated landing page variants.
Results and Iteration: Implement winning variations immediately and record losses to inform the next hypothesis.
Watch the Full Discussion
Catch the complete recording of our LinkedIn Live talk on catch-up:
- YouTube Replay: Grow with Launch Channel
- LinkedIn Replay: Launch Online Company Page
Case Study: Tackling Ad Scepticism with Icelandair
Usability testing is frequently reserved for site navigation but testing paid ad creative reveals why users hesitate before clicking.
When testing ad assets for Icelandair, audience feedback revealed unexpected friction. Viewers assumed real landscape photographs of Iceland were CGI or AI-generated, which damaged trust.
Instead of scrapping the photos, we formed a direct hypothesis: address the audience’s scepticism head-on within the ad copy. We added explicit reassurances (“This is not AI. It’s just the Northern Lights in Iceland”) and rolled out an out-of-home video campaign featuring local residents pointing out real geysers.
Addressing customer doubt before the click delivered major commercial performance gains compared to the prior period:
- Cost Per Acquisition (CPA): 23% lower
- Cost Per Search (CPS): 11% lower
- Return on Ad Spend (ROAS): 40% higher
This principle applies directly to lead generation. If prospective leads doubt an offer’s claim on social media or search ads, form completion rates downstream will suffer regardless of how clean your landing page looks.
Mining Customer Feedback for Copy Gold
Effective testing does not always require high video production budgets. Brands can extract high-converting messaging directly from data they already own:
- DTC and E-Commerce: Mining Trustpilot reviews for apparel brand Weird Fish highlighted recurring customer quotes. Placing verbatim customer phrases directly into paid social creative significantly boosted click-through and purchase conversion rates.
- Lead Generation and Niche Categories: In pet care testing, pulling customer review phrases into Google Search copy (“Sensitive Stomach Approved”) yielded an immediate conversion lift. For B2B and lead gen campaigns, reviewing support logs and sales call notes reveals the exact language prospects use to describe their pain points. Using those precise phrases in search copy and lead form headlines removes hesitation before prospects reach the form.
Why Losing Tests Protect Acquisition Margins
Roughly 1 in 10 experiments yields a clear winning variation. A negative or inconclusive result is still a commercial win because it saves your wallet and prevents wasted ad spend.
In one ad experiment, we tested replacing a static product image with a feature video, hypothesising that movement would increase engagement. Click-through rates dropped from 3.03% to 2.36%. The variation lost, but running a controlled experiment saved the brand from committing its entire media budget to an underperforming creative asset.
Businesses running up to 15 structured experiments a year see up to a 30% boost in ad performance in year one compared to non-testing brands. By year two, that figure jumps to 45%.
Centralising Insights in a Learning Library
Running tests across multiple channels generates dozens of hypotheses that quickly become difficult to manage. At Launch, we maintain a centralised Learning Library in Airtable, recording every test title, hypothesis, status, and outcome.
When a team member suggests an ad tweak, lead form alteration, or landing page change two years later, you avoid guessing. Consulting the library shows previous test results, why a variation failed, and what to try next.
Unlocking Scalable Growth
Growth does not happen by accident, nor does it happen simply by ramping up your ad spend. It happens in that messy middle between the first ad impression and the checkout basket or lead submission form.
Treating experimentation as a full-funnel discipline protects acquisition margins and builds a compound competitive advantage.
To see where your business is losing growth across creative, conversion, and media structure, benchmark your marketing engine using our self-serve Seven Levers Scorecard at https://www.launchonline.co.uk/grow/
ABOUT THE AUTHOR
Amy Hind is a Conversion Consultant at Launch. With four years of industry experience across experimentation, user research, and performance media, she helps brands build data-driven testing frameworks that unlock performance plateaus across the entire customer funnel.
