Improving Segmentation & Scale – E-Commerce $$Case Study - Audit
For a growing e-commerce client with an above average purchase price and a purchase path with a time lag also a little longer than normal, we were tasked with a YouTube video audit to see if their current targeting, creative, and bid optimization was yielding the best possible performance.
Variable Media was able to break down many variables based on their current campaign and cross compared against key metrics ROI & CPA.
This data evaluation yielded insight that could project an improvement in campaign results almost instantly requiring a mandatory testing budget & strategy for implementation
The primary challenge was to gather, analyze, and future project results based on campaign data totaling less than 5M views 850 conversions and 30K clicks. Although the client had tested very few targeted campaigns many creatives and first level variables were able to provide rich results that could be leveraged to increase performance results through optimization.
- Age & Gender & Parental Status
Understanding demographic & life stage was key to our future strategy. Ultimately these first level variables help set our bid strategy. Also, environment information can also be understood helping us target more effectively.
Not only must we know which creative performs the best, but what creative performs against a particular segment vs. another. With over 5M views against these video pieces, we should learn and optimize to efficiency.
- Device & Geo-Location
Where and how we reach consumers is vastly important to understand the user behavior and best conversion device opportunity. In many cases, the device is a key indicator as to where this type of advertising should be placed on the path to purchase or purchase funnel.
As we analyzed the data, we were able to find key areas of improvement as well as the additional insight that will help our target audience positioning.
- Age & Gender & Parental Status
Demographically we were able to isolate the best-performing age groups and bid adjust to maximize exposure to those users most likely to convert. Moreover, life stage played a significant role with non-parents converting twice as much as parents resulting in a better ROI & CPA.
Multiple creatives were testing and interesting to find they all performed very differently based on the age and gender variables. These important data points proved useful in future campaigns as we were able to dial in our audience to best converters combined with the creative that is most likely to produce a conversion.
- Device &oeo-Location
In testing our e-commerce product, our user behavior assumptions held true proving that most users are more likely to engage on a mobile device but convert on a desktop or laptop.
Multi-Variable Views Clicks Conversions CPA CTC Not a parent 2,407,166 15,482 500 $207.6 3.2% Laptops / Comps 729,148 1,540 333 $124.0 21.6% Mobile Devices 1,233,609 10,862 113 $387.6 1.0% Tablets 444,409 3,080 54 $346.4 1.8% Parent 1,775,688 16,410 242 $412.8 1.5% Laptops / Comps 451,262 624 145 $202.5 23.2% Mobile Devices 608,854 12,274 58 $544.7 0.5% Tablets 715,572 3,512 39 $998.2 1.1% Grand Total 4,674,226 33,740 777 $286.5 2.3%
After reviewing the data we were able to build a plan that would improve results by 25-40% dependent on user targeting at the top and bottom of the purchase funnel. In addition we were able to cross a few variables to better understand user behavior and demographics which increase confidence in our projection.
Variable Media’s free audit proved valuable to our leading e-commerce client partner resulting in a number of key findings.
By the numbers, the effort:
- With even spending and over 150K views associated creative one performed 50% better than creative 2. While creative one was less efficient, parents contributed 2x better performance than non-parents resulting in an optimization recommendation that would improve performance. Creative 2 produced the opposite results.
- Desktop users converted at a 20% click to conversion ratio while Mobile hovered around 1%. Also, females converted at 1% vs. males at 3% primarily due to their difference in desktop performance
- Age analysis helped shape our bid strategy applying a pyramid approach placing a higher adjustment to middle aged 25-44 with less emphasis on the book ends 18-24 and 65+