It took several weeks to solve this problem. We wrote a script that taken into account all available parameters and sorted them out. We highlighted every feature that a user drew attention to as the main parameters for the offer development. We also had default parameters, which were worth for analyzing. For example, for mobile phones we took into account their screen size, the operating system, RAM, their price. Having collected all necessary parameters, the script searched for the corresponding models and notifications sent if the cost for some of them had been changed. If these parameters were considered as new, we sent a notification to users as to a new arrival. Only during a couple of days when the algorithm was working, we increased mailing conversion rate from 2% to 15%. Just imagine, if at least some restaurant companies did it. They boost sales in times through online. If you are an owner of small cafe - how much does restaurant website cost.
We identified each user by a unique code that we had written in cookies in order to keep accurate parameters. For mass mailing we used automatic and unique tags for each email. Thus, when the user clicked the link and went to the site, we had already known who exactly came to us. After conducting a work on scripts developing, we started sorting the product based on its history and preferences. The more user was available on the site, the more information about him we collected. And this made it possible to form unique proposals more accurately for each person. Doing this, we increased loyalty, average views and number of purchases, what is more important.
So we decided to move forward. We already had a ready to purchase audience which was tracked by our team. Having a discussion with the CEO of the company, we had a question arisen: is it possible to improve our result? Therefore, we decided to perform even deeper iteration and identified the most active users. Further, we sorted them into the following groups: -the most active, involved and slightly active. So that we had profound and detailed information about their behavior, such as which products were viewed, how long they were interested, whether they read reviews, etc. Basing on this statistics we noticed that after 5-6 weeks a user reduced its activity. During this very time he managed to choose and buy the appropriate product. In this case we decided to send a letter with a request as to providing us with a comment or give a review on goods in order not to lose connection with the client. This made possible to increase the number of useful reviews. This concept is acceptable for cryptocurrency exchange platform.
So let’s sum up the gained results:
Mass mailing conversion: 15%
Involvement: 6.3 pages per user
Activity time: 2.45 minutes
It was a quite complicated task, which gave us good results and increased number of sales.