How to increase the number of leads by 637% for a real estate agency in Dubai without losing your mind?

It was an extraordinary year of 2022. Our fellow citizens were rapidly leaving their homeland. Many left with the desire to find, if not a new, then at least a spare house. Quite logically, in the midst of such events, a real estate agency from Dubai asked us for help.

The client has already attracted buyers through other channels and wanted to scale. We were required to increase the share of requests from Russian-speaking clients in order to occupy the capacity of the Russian-speaking department.
We decided to work with the Google Ads service. The demonstration of targeted advertising, which, as you know, is impossible in Russia. That is, we had to search for the Russian-speaking audience in other countries

The first thing we did was to launch several advertisement campaigns for the Russian—speaking and English-speaking audiences. The advertisement campaigns for the US and Europe had already been lauched in Google . The advertisement was working mainly for the English-speaking target audience. But do not forget that it was the Russian-speaking segment that the client needed :

• three advertisement campaigns for the English-speeking segment
• one advertisement campaign for the Russian-speaking segment

In order to target the Russian—speaking audience, we launched campaigns almost all over the world: 50 countries of Europe, CIS, and Asia. In the campaign settings we chose the “Russian” targeting and envisaged how interesting and ornate our path would be.

In the first month of promotion, we launched search campaigns from the client:

• by one of the properties ("the name of the property" Dubai)
• with shared keys (apartments in Dubai price)
• with keys on moving (moving to Dubai)
• English-speaking campaign with shared keys and targeting the USA

We added new search campaigns for the Russian-speaking audience by:
● brand — for the audience that is already familiar with the brands of specific residential complexes / our agency
● competitors — for the audience that is familiar with the brands of other real estate agencies
● comfort class real estate — for the audience that is looking for real estate in Dubai, but is still at the stage of choosing a specific residential complex/real estate agency (general commercial requests, for example, "buy apartments in Dubai")
● premium class real estate — for the audience that is looking for premium real estate in Dubai (villas, penthouses, etc.), but is still at the stage of choosing a specific developer/ real estate agency
● investments — for the audience that is looking for real estate for investment.
In the first month, it became clear to us that it is much easier to reach the English-speaking paying audience from the United States than to find the Russian-speaking audience with an income level of $ 5,000 per month outside of Russia. The campaign in the English language in July brought 70% of leads.
Having decided that this was good, in August we added campaigns in English by analogy with Russian-speaking ones:
● by brand
● comfort and premium class real estate
● investments
● by one of the properties
The number of leads from search campaigns increased to 102, of which 96 still came from English-language advertising campaigns. The Russian-speaking audience stubbornly refused to be converted into leads. The promised KPIs were, of course, met. But the Russian-speaking department of the client, as before, was tediously waiting for a work surge. The result was that the client was not satisfied. And then we responsibly transfered most of our resources to the comprehensive elaboration of the Russian-language advertising campaigns.

One of our first and main hypotheses was that the Russian-speaking audience, which is already familiar with the brands of properties/developers/agencies-competitors in Dubai, would suit our request. Then we started collecting maximum traffic in these directions. The agency has more than a thousand properties in Dubai, so collecting semantics and preparing campaigns turned out to be quite a labor-intensive process. There were more than 100 names of properties alone. Even now, after 4 months of work, we continue collecting semantics and adding new groups of ads. Then the time for separate search campaigns has come:

• brands of residential complexes
• developers in the UAE
• brand of the Real Estate Agency

There are more than 100 groups in them now.

General campaigns aimed at the Russian-speaking audience practically did not bring results. In July, we received only 5 leads, having spent almost 15,000 US dollars. A good enough reason to panic.
But we, without losing our spirits, studied the statistics by country. We identified the most prioritized ones for promotion in the form of individual advertising campaigns. In terms of the number and quality of contacts, Germany, Spain and the UAE turned out to be the most targeted. For them, separate offers in the ads were written. At the same time, in general campaigns, we excluded countries that brought non-targeted traffic (Bangladesh, India, Kyrgyzstan and other developing countries). The client regularly gave us feedback on the quality of requests, we, in turn, promptly corrected the advertising campaign.

The restrictions did not help to cut off the traffic from unsuitable countries by 100% with the first attempt . We harnessed our willpower and made the targeting tougher — “People who are in target locations and regularly visit them” instead of “People from target locations and those interested in them". Now ads were shown only to those who were in the countries that we were interested in.
In the general advertising campaigns for comfort-class real estate (here are the broadest requests like "real estate in UAE"), we tested 3 landings. We chose to go with the agency's main website for PC-traffic and the real estate agency's branded landing page for mobile traffic.

We optimized out advertising campaigns on a daily basis:
● we worked on increasing the relevance of ads — tested various landing pages, increased the relevance of headlines, improved the ads
● we tried to improve the quality of traffic — we added all the keys to the phrasal correspondence, which decreased the number of rejections, but reduced coverage, regularly added negative keywords
● we switched off inefficient keywords
● we added bid adjustments by device type, age, and family income

The result of October was 24 leads 354 US dollars each. Compared to five leads for 15,000 US dollars, it looked like a win. But it's not the victory yet.

As if deciding that the above experiences were not be enough, periodic bursts of spam applications interfered with our work.

Surely, the quality of traffic from online campaigns is often low. And we usually cut off non-targeted traffic at the start:
— we add excluded YouTube channels and placements (we have collected our own list of low-quality sites), settings of excluded content
— we limit the frequency of ad demonstrations per user

But any attempts to increase the coverage of GDN led to waves of spam again and again. In response to that, we created a checklist for any changes in the advertisement campaign. With it, we managed not to miss a single setting that leads to a surge in leads. Excluding countries, by analogy with general traffic, did not help even with stricter targeting methods. Therefore, together with the client we identified the most prioritized countries (Spain, France, Germany) and left targeting only for them. Branded GDN, working on the "Maximum conversions" strategy, began to show good results.

The share of high-quality applications was much lower than on the search, but their final cost was significantly cheaper in comparison the the search one:

The main system of end—to-end analytics for the project is Roistat. It is synchronized with the client's AmoCRM and advertising cabinets, which allows you to link data on the quality of leads with sources. Also, through the integration of Whatsapp with CRM, we receive information about contacts in the chats (the main communication channel in this market).
All fine and dandy, but we didn't, don't and won't have access to AmoCRM.
This means that we cannot send the data about the applications to Google Analytics and transfer them to the cabinets for optimization. Unfortunately, Roistat also does not know how to transfer the aggregated data received further into analytics systems. And so, we get a bundle that works only one way — the data from advertising cabinets and CRM are aggregated in Roistat, but there is no data on applications from Roistat in the advertising cabinets.
The situation becomes even more complicated due to the number of analytics systems — about 20 systems, one for each landing page. The problem was partially solved by setting up goals — to send the form and to click on the Whatsapp application (there was no possibility to set up a goal to send a message from Whatsapp). As a result, our strategies in Google Ads were trained only on a part of the data. In order for the data in Roistat and Google Ads to match, we chose a roundabout way. We set up the transfer of offline-conversions with a certain status from AmoCRM via the Albato service.
In this case, the conversion data is updated in online mode. This means that advertising campaigns can be optimized immediately for high-quality leads. This task is still being implemented by us.

In July, the number of high-quality applications was 0.7% in GDN and 33% in the Search. In October, the share increased to 39% in GDN and 65% in the Search.
During 4 months of promotion, there were 6 sales from advertising campaigns totaling about 2.4 million US dollars, which was 14 times higher than the advertising costs.