Overview
This VC-backed electronics company sells consumer privacy hardware. The device connects to your internet modem and scrambles your internet identity and blocks intrusive data collection.
Challenges
After their initial funding round and product launch, the original ad campaigns were generating 300 sales per month. Previous attempts to increase sales beyond this mark failed to maintain a profitable customer acquisition cost.
Our initial analysis proved that brand, organic, and direct traffic took a dive after Q2 2020. Our hypothesis was that the lack of brand awareness efforts was limiting our reach and exposure to new quality prospecting audiences.
Our goal was to use YouTube campaigns to increase reach and drive incremental revenue, all within our target profit margins.
However, given the niche nature of this product, traditional YouTube targeting presented too much risk. We needed to find a creative solution to our audience targeting: one that allowed us to hyper-target our ads and ensure that our placements were in-line with the client’s brand image and guidelines.
Strategy
We partnered with AdSkate, an AI-powered technology platform that helps advertisers improve the effectiveness and transparency of their ad targeting.
AdSkate allowed us to leverage data signals from multiple sources (Google, Facebook, Instagram, and other customer data), to determine an ideal customer profile. Then, they used their AI technology to discover 300+ YouTube channels for optimum targeting.
Learn more about AdSkate here.
Results
The YouTube campaign ran for 45 days. We measured its success against an average historic benchmark of traffic and revenue.
The impact of such brand awareness efforts are not easily measured in the Google Ads dashboard. We therefore measured direct traffic, organic traffic, and overall site revenue as the benchmark. We also used a regression-based incremental lift study to measure the correlation.
After just 45 days, the YouTube campaign generated an incremental revenue increase of 44.1%, compared to our benchmark.
The regression model proved a statistically significant p-value below 0.05.
