$3k profit on Car Insurance offer with Facebook ads
Today we present a case study from our internal media buying team on running Car Insurance offer with Facebook ads.
We run different whitehat campaigns on Pay Per Call and CPL basis, so we finally got time to test one of the most popular verticals – auto insurance. This offer is 100% whitehat, so you don’t experience account bans in any traffic sources. You need to use a prelander page, as Facebook doesn’t allow you to redirect links in ads.
We started promoting several car insurance offers in October 2020. During that time we tested several approaches and found which ones worked better and then scaled them in November. We splitted traffic between several car insurance offers, so our payout per lead was in the range of $6.5 to $8.5 per lead. Average EPC of the offers was $3.8 (we excluded days when Facebook bots actively visited our lander). Payout varies depending on weekdays and time (night hours and weekends are usually paid lower than average). With each offer we got payout and caps bump once they had checked our traffic quality.
Campaign and ad sets settings:
– Campaign type: conversions
– Optimization with a pixel installed on page after lead submit
– Targeting M/F 25-55 y.o.
– Placements: All Feeds
Our tactics for launching new creatives was the following: launch with a $50 budget per day, check 2 day performance, if the ROI is negative, then pause and delete the ad. Make the same ads and try again 3-4 times until positive ROI. Sometimes Facebook algorithm may target the wrong audience, after some attempts you will stick to the most converting one.
Important part is that you should always have conversion pixel installed to allow Facebook algorithm to learn and optimize.
Our system allows you to automatically place your pixel on the page after lead submit to track your leads.
Ad creatives:
We can’t share exact creatives that we used but can share similar approaches that we found on spy tools:
Ideas for the text:
- Get auto insurance for as little as $XX/month! Find your BEST rate in less than 2 minutes.
- Many people in {State} are overpaying on their car insurance by over $XXX per year! Don’t be one of them
- Found out today that I was overpaying for car insurance. Thanks to website I went from $XXX to $XX per month same coverage
- No tickets in the last year? Enter your age & ZIP for a new rate like I did.
- This is what 2020’s cheapest auto insurance looks like, rates as low as $XX / month. Check the lowest rate in less than 60 sec!
We found that images with people have higher CTR than images with cars. There was no big difference in CTR between images with male and female.
Video ads had higher CTR but were harder to create as you need to pay freelancers to make them and wait several days.
If you have spy tools, you will easily find similar creatives for the inspiration.
Want to get some examples of videos from spytools? Ask our affiliate manager!
Prelander
We used a simple prelander with a one step form where the customer should fill his zipcode and then we redirected him to the offer page. Example of such a prelander is below (note: not our real prelander).
Our prelander showed a conversion rate of 43%, while offer pages conversion rate was in the range of 45-53%.
Some offers allow to pass customer data in tracking link, so you can collect more info on your prelander like this:
We didn’t test this approach as we were running several offers, most of them didn’t have option to pass customer data.
Results
Period: November, 2020
Payout per lead: $6.5-$8.5
Revenue: $18 494
Expenses: $15 384
Profit: $3 435
ROI: 20%
Traffic source: Facebook ads
We used only one ad account and didn’t have any issues with bans and suspensions. It was our second month running this vertical and we got 20% ROI. We consider it as a positive result, as it’s whitehat offer in a huge market and we are going to scale this campaign in several months. Our profit also was affected by Black Friday (27 Nov) and Cyber Monday (30 Nov), in that period CPM was really high which resulted in negative ROI.
