- Problem: before transferring the project to our #ADINDEX agency, the client decided to update the site. Forecasts were made based on the statistics obtained from the old site, and ads were launched on the new one. The first results showed that we did not reach the planned indicators.
- Solution: A/B testing of the websites. Experiments were launched in all search campaigns: branded, by general queries, competitors. Traffic is distributed 50% to 50%.
- Result: for three weeks of the experiment (January 14, 2022 – February 3, 2022), the following results were obtained:
– the new site brought 7 completed forms, the conversion rate was 0.47%;
– the old site brought 28 completed forms, a conversion rate of 1.81%.
The conversion cost on the old site is 4 times lower.
Conclusions: This experiment confirms that different sites can convert differently. Therefore, it is important and necessary to experiment.
Analyzing two sites, we determined what factors could affect the conversion rate:
- More useful and understandable information on the first screen of the site, so that the user wants to stay on the site or take a targeted action.
- The benefits/advantages that the client receives are described in detail and clearly. For ease of perception, you can supplement with infographics.
- Target button on every page of the site.
- Clear and appropriate call to action “Get a presentation”, “Fill out a form” and “Submit data”, etc.
- The presence of blocks “Frequently asked questions” and “Reviews”.
- Contact details at the top of the site (phone, messenger buttons, social networks).
- A single style of the site is observed: pictures in the same style, color.
- Since the user needs time to make a decision in this case, it is advisable to make intermediate conversions, such as a feedback form, ask a question, etc.
- Describe as clearly and in numbers as possible how much money and effort should be invested and what benefits the client will receive.
- When a new site is rolled out or significant changes are made to an existing one, conduct A/B testing.