Objective
Applications for the purchase of a car on lease through the website at the lowest price.
About the project
The project has a high-converting website which is the main platform handling the inbound traffic. The site is translated into 3 languages – Latvian, English and Russian. There are over 300 different cars available in stock. Google Ads and Facebook Ads are used for advertising. Customers are located all over Latvia.
At the start of work advertising on Facebook and Instagram generated leads at the cost per lead of about 2,7€. It was these social media that we started working with. The advertising budget is around 1500€ a month.
Stages of work
- Analysis of the ongoing advertising campaigns
- Choice of the advertising campaigns type to work with
- Preparation of the hypotheses for testing
- Analysis and upscaling of the results
Analysis of the ad campaigns
The ongoing campaigns generated leads at an average cost per lead of 2,7€ for the past month. During the detailed study we took into account cost per conversion, the most converting placements, cost per mille (CPM), effective types of ad campaigns, detailed targeting options, and the use of retargeted advertising campaigns, the most converting creatives. We built a further work plan according to the analysis.
Choice of the advertising campaigns type to work with
After the analysis, it became clear that the best results could be achieved with the three types of campaigns:
- Advertising of goods from the catalog
- Retargeted advertising of goods from the catalog
- Advertising with the conversion purpose, using the designer banners
Such campaigns built a good customer funnel. Advertising of goods from the catalog as well as advertising with the designer banners ensured the inflow of targeted traffic and the lead flow. Retargeted advertising of goods from the catalog displayed the previously viewed item to those who had visited the site without submitting an application. This way our ads reached those interested in a specific product who didn’t perform the target action right away.
Preparation of the hypotheses for testing
- It was Facebook that turned out to be one of the most converting placements, so we decided to test this placement separately.
- We identified some additional interest categories, connected with the purchase of a pre-owned car, together with credits and loans. Besides, the interests of purchasing a pre-owned car were transformed into a separate audience.
- We worked separately with a wide audience limited by age and geolocation throughout Latvia.
- Apart from it, we created several look-a-like audiences: based on the people who had spent the longest time on the site (TOP 25%), and on those who had submitted the application on the site.
Here are some examples of the promotional posts which we used to test the hypotheses:
Analysis and upscaling of the results
All these hypotheses were tested with the purpose of conversion and advertising of the catalog. Categories of interests and look-a-like audiences showed the best results; Facebook as a separate placement had no significant effect on the results. Initially, the wide audience generated expensive results, but 3-5 days after the optimization the cost per lead dropped significantly. After analyzing these results, we could expect the reduction in the cost per lead by approximately 25%, compared to the initial data.
The same types of advertising campaigns are still running, the visual is updated on a monthly basis to overcome banner blindness. In order to keep the funnel working, around 75% of the budget is being spent on work with cold audiences and about 25% on the retargeted advertising campaigns. The monthly budget is around 1500€.
Results
For over 2 months there has been a steady lead flow – more than 850 applications a month.
Received 2058 applications in total
Budget spent: 3610€
Cost per lead: 1,75€
Thanks to our testing of the new hypotheses as well as the existing advertising campaigns optimization, we managed to reduce the cost per lead from 2,7€ to 1,75€ already in the first month. An application on the website is now 35% cheaper than before the start of work.
