Why not take off the real estate portal. Part 1

I often began to ask why I did not work with the real estate portal (Best-House.org) and I decided not to answer a piece in different comments, but to write an article with my own thoughts and experiences.

So, as promised, I share my experience. I had no problems creating a real estate portal, for obvious reasons. Difficulties arose in the next stage.

By the way, the article, as I planned it all, saw and implemented, here: Startup on my own

The first problem is attracting to an empty portal those who will publish their property offers. How I solved it. I created a system for parsing email on the website. I started the parser on all real estate portals, on bulletin boards on the subject of “Real Estate”, in short, wherever it was possible to get the user's address. It was at that moment that I realized how vulnerable 99% of resources in the internet were to leak users' personal data. Only a couple of major real estate players paid due attention to the issue of the safety and security of personal data of their users. All the rest did not care.

Having stuffed the base for several million addresses of any garbage, I wrote a system for checking the quality of these addresses. What did she do. The system deleted duplicate addresses, deleted addresses with the wrong format, deleted addresses with any abracadabra and muddy domains. After that, the base has shrunk four times. Further, the system began a long but detailed test for the existence of these addresses. Why all this, I will explain further. How did it happen technically.

From the address database I received an address that was not yet verified, then made a request for the domain of this address, if the domain answers me without delay, it means it exists, if not, then the address is immediately assigned the status of bad. If the domain responds, then I make a request for the existence of a specific email, without sending a real letter. If I get an error, then the address does not exist; if I get an answer, the address goes to the white list.

The problem is that you can not do a check in several threads, because mailers such as Google buy that the bot is accessing, and this is not a real request from a live user. And the interval between calls should also be within reasonable limits. Therefore, the test was very long.

As a result of such dances with a tambourine, I received a list of one hundred thousand living addresses. I understood that if I started doing mailing from my server, then the ip of my server would be quickly banned. Then you can close the project immediately. Further, my address verification program allowed to split all addresses into files with the maximum number of addresses: 2 thousand records.

I needed a stable mailing of letters, reliable and free. After going through all the circles of hell to find a suitable action algorithm, I still found the right path. I began to use MailChimp, a statewide project, which allowed me to send 2 thousand letters for free no earlier than 14 days after the last mailing. I want to say that the service is not stupid and easily calculates all sorts of tricks and tricks. But I also did not do a finger. From under different ip addresses I created about a dozen accounts, with different addresses for a profile, with different text for distribution.

The service allows you to upload subscriber addresses to it in any way. I used plain text files. Immediately upon loading, the service checks the addresses for validity. There were many addresses where a phone number was used before the dog's badge. So, Mailchimp doesn’t like that very much, and he stops the import process and shouts that you are not a good person and you try to fake soap with him. This is the first nuance. To avoid this, I had to mix such soaps with standard soaps. 95% standard and 5% "digital" addresses.

Since milchimp is a very trusted service and it has a system for unsubscribing messages and a powerful system for analyzing and behavioral factors, the effectiveness of the distribution is very good. It should also be remembered that if users press a button on a button (link), it is Spam, or even worse, complain to the hoster about such a letter, then if you exceed 3% of the abuse message (complaint), your account will be blocked. Sending letters to non-existent addresses also acts as an abuse. Now it became clear to you why it is so necessary to approach the quality of addresses!

What is the final? Of the 100 thousand messages sent over the course of several months, I received 1500 real registered users, who were, for the most part, realtors or Real Estate Agencies.

I rejoiced at the work done as a child, as at this stage my work and strategy paid off. Real difficulties just waited for me.

Yet the story again comes out immensely, and in order not to write the sheets of the text and not over-exaggerate the readers, I decided to break it into pieces. I think this will be enough for the first part, in the second part I will write about how I began to develop relations with registered users-sellers of real estate, about what wishes on the project they wrote to me, what I improved, and how things went further.

To be continued…

Reprint from your own blog. Original here