On April 24th, we launched a native integration to Facebook allowing retailers to create Custom and Lookalike Audiences using product, purchase, customer data, along with predictive data sets, such as predicted gender and predicted next order date. We’ve been proud to stand behind the success our clients, such as NewYorkDress and RightStufAnime.com are seeing with this tool; however, we were curious how our Custom Audiences would stand up in a test against the exact same Facebook Audience.
One online retailer, working with our partner, ROI Revolution, was willing to do such a test. This retailer, whose estimated online sales in 2016 topped $15 million, worked with ROI Revolution to set up the campaign.
Test 1
AUDIENCE: People who purchased once in the last 30 days and had an average order value (AOV) of at least $50
LENGTH OF TEST: 14 Days
AD SETUP: Facebook split-test with identical budget and targeting opportunities
RESULTS:
| Metrics | WhatCounts Custom Audience | Facebook Custom Audience |
| Spend | $321.16 | $352.08 |
| Link Clicks | 86 | 89 |
| CPC | $3.73 | $3.96 |
| Impressions | 11,349 | 11,763 |
| CTR | .76% | .76% |
| Metrics | WhatCounts Custom Audience | Facebook Custom Audience | Outcome |
| Facebook Conversions | 20 | 7 | 186% more conversions |
| Facebook Revenue | $2,416.02 | $533.40 | 4.5x more revenue |
| Facebook CPA | $16.06 | $50.30 | 68% lower CPA |
| Facebook Conversion Rate | 23.26% | 7.87% | 195% higher conversion rate |
| Facebook AOV | $120.80 | $79.06 | 1.5x higher AOV (2.6x higher than retailer’s 2017 AOV) |
| Facebook ROI | 652.28% | 57.18% | 1040% higher ROI |
Across the board Facebook’s Audience fell short. The ad itself wasn’t a promotion, but, rather, an evergreen ad. Everything from copy, creative, CTAs, etc. was identical for purposes of the split test.
So, the big question to answer: was this success repeatable?
In the interest of fully fleshing this question out, ROI Revolution did a second 2-week campaign. The results? Similar to the initial test with WhatCounts’ Custom Audience emerging the winner.
TEST 2
SETUP: Same as Test 1
RESULTS:
| Metrics | WhatCounts Custom Audience | Facebook Custom Audience |
| Spend | $386.66 | $388.00 |
| Link Clicks | 59 | 110 |
| CPC | $6.55 | $3.53 |
| Impressions | 16,945 | 16,636 |
| CTR | .35% | .66% |
| Metrics | WhatCounts Custom Audience | Facebook Custom Audience | Outcome |
| Facebook Conversions | 26 | 15 | 73% more conversions |
| Facebook Revenue | $1,758.13 | $1,066.45 | 1.6x more revenue |
| Facebook CPA | $14.87 | $25.87 | 42% lower CPA |
| Facebook Conversion Rate | 44.07% | 13.64% | 223% higher conversion rate |
| Facebook AOV | $67.62 | $71.10 | AOV $3 lower |
| Facebook ROI | 354.7% | 174.86% | 102% higher ROI |
Interestingly, in this second test, you can see that the WhatCounts Custom Audience had 51 fewer link clicks and a lower CTR; however, when the ad was clicked it converted more often (223% higher conversion rate), generated more revenue, and resulted in a 42% lower CPA than the Facebook Custom Audience.
This two-part test shows that Facebook Custom Audiences, powered by WhatCounts data, beat out Facebook generated Custom Audiences, increasing ROI, conversions, and revenue by targeting the right customers with relevant information, creating better shots-on-goal and driving down advertising costs for retailers.
To learn more about our Facebook Integration, contact us or click here to get a demo.
