89 plots worth 0.5 billion rubles in an open field in 3 months

How to sell expensive country real estate? Where is it better to advertise on a website or landing page? The developer wanted to sell 70 plots over the summer. The first advertising text on the site brought in few sales. We came up with a cool micro-goal and improved the results 12 times. We made a landing page with a short sales funnel. And from both resources they sold 89 plots for 550 million rubles.

The developer wanted to sell more expensive land in the Moscow region and sell 70 plots of the new village over the summer

The owner of a company selling cheap land in the Moscow region decided to change his business strategy in order to earn more. The developer's marketing department studied market trends, data from Domklik, and reported that over the past year, the volume of proposals for suburban real estate has almost halved. And prices have increased, especially in: Solnechnogorsk (+17%), Voskresensk (+15%), Ramenskoye (+12%) and Dmitrov (+10%).

The company also observed a tendency to reduce offers. Just a year ago, 12 economy-class villages were for sale on the site, and in 2024 only 5 remained. The remaining 7 were successfully sold and have not yet rushed to open new projects.

The board of directors decided: we will move to a new, higher price segment. One “expensive” village is much more profitable than 3 democratic ones. The company had land for individual housing construction on Novoryazanskoye Highway. We decided to make a comfort-class village there and, without waiting for communications, start sales in May 2024.

We have been promoting developer projects on Yandex Direct for the fourth year now. And at the end of April we received the task: to quickly launch an advertisement for a new village, which has 198 plots. And during the high season (May-August 2024), no less than 70 plots need to be sold.

Note that 70 is a fairly high figure. Previously, our records were: 50-60 sales per season. In addition, prices in previous villages were much lower – about 600K per plot. In the new village, the price was set at 1.2 million rubles. for one hundred square meters, that is, 10 times more expensive. We understood that for this project we needed a completely different target audience, much more affluent. And they thought that on a general site where all the cheap offers were collected, she might be disappointed. What to do?

We began to think about where to advertise: applications from the site are of higher quality and more expensive, while applications from the landing page are cheaper and there are many times more of them

In addition to the general website, we had a small landing page with one inexpensive village. By that time, we had already sold all the plots on it, with the exception of a few of the most illiquid ones.

They began to figure out where it would be better to advertise the new village: on a common website or remake the landing page for a new project. On the one hand, landing pages are trusted less than websites, especially in expensive topics. But on the other hand, many more applications come from the landing page. Even if they convert 2 times less into a deal, they generally give more sales.

And leads from landing pages are cheaper: 1,500 rubles. versus 3000-4000 rubles. This is due to the fact that the one-page website is not overloaded with content, there is only one project. And the result is a faster funnel: you have received the minimum necessary information, leave a request or call – then you will receive a master plan with prices. And the site gives much more data and choice. The user has a reason to take a break to study and think about different proposals. Therefore, there are fewer applications from the site, but the sales conversion is higher.

Since we needed to sell 35% of the village in 3-4 months, we decided to remake the landing page for a new project and conduct all advertising on it. Well, while it was being reworked, we were forced to work with the site.

Neural networks drew fantasy-style visuals and taught them realism. And added valuable USPs to the offers

When preparing texts for advertisements, we used USPs that had proven themselves best in campaigns of previous projects. They prescribed: distance from Moscow, minimum price per hundred square meters, the most popular infrastructure elements.

Additionally, we added benefits that we call “for the soul.” Triggers that appeal to deep emotions, the desire to be involved in cultural values: “near the water”, “historical place”, “nearby the writers’ dachas”. Not far from our village there was indeed a settlement where writers and poets of the early 20th century, and then Communist Party leaders, came for the summer.

For advertising formats with visualization, we usually use renderings made by the client. But this project came together so quickly that by the time the advertising was launched, they only had time to film the object. And designers needed 2-3 weeks to develop 3D visualizations. Therefore, we decided to bring in AI to help: the Yandex neural network and CoPilot.

The first results of their work were imperfect. For the initial test, CoPilot produced a very surreal picture with purple beds and a fluffy tree; it failed in advertising. After retraining, I began to draw pictures that were still a little fabulous, but more adequate, and they found their fans – their effectiveness rose to the highest level.

The Yandex neural network coped with the first task much better. The illustration turned out to be quite realistic, attractive and received a good score from Direct. Subsequent ones became even more effective. Especially with the image of a house under construction.

Conducted cascade training of Ya.Direct algorithms on conversions and bids. The method brought a lot of traffic, but few sales

So we prepared and tested almost perfect advertisements. We launched advertising in May. Involved: Campaign Master, 2 separate campaigns in EPK on Search and in YAN + retargeting. As described earlier, at this time the client was reworking the landing page for a new village, so traffic was directed to the general website.

We launched our own method of cascade training of Ya.Direct algorithms on different types of conversions and bids. We tell you what we did step by step:

1. Targeted advertising campaigns to pay for 4 macro conversions. Calls and requests from the site that bring the hottest audience. And also messengers and quizzes – the target audience for them is rather warm.

2. Set the highest possible rates. The hottest requests are always more expensive, and for them the rate is large – 10,000 rubles. We pay less for warm ones – 5,000 rubles each. We encouraged the Republic of Kazakhstan to catch on with the coolest target audience and bring them to the site.

3. In the first week of training we received 5 conversions: 2 calls, 1 application from the website and one lead each from messengers and a quiz.

4. Started to lower rates: calls and requests for 9,000 rubles, messengers and quizzes for 4,000 rubles. So we gradually lowered the bar until we reached the numbers we needed: 6,000 rubles. for hot requests and 3000 rubles for warm ones.

5. Monitored impression statistics in Yandex. We collected search queries for which advertisements were shown and removed ineffective ones. We watched where we showed up and cleared out bad areas. We drew conclusions about what goal to set next.

6. Switched to paying for microconversions. We studied the statistics and saw that before contacting, users spent at least 10 minutes on the site. We set up time goals for the campaigns on the website – 10 minutes.

The advantage of this method is that it provides a huge amount of cheap traffic. It also helps to notify interested audiences about the emergence of a new project. Having this knowledge, the target audience then independently searches for information about it in other ways.

So, we did everything right. But in the first month there were very few requests – only 10, and qualified leads (from people who are ready to buy from us) – 3. And the vast majority of users who went to the site through advertising were not financially ready to pay 6-8 million for land in a new village. The situation was dire. There were only 3 months left to sell 70 plots…

Yandex algorithms have been weaned from bringing “economy” to the site. We came up with a micro goal that increased results by 12 times

We began to analyze the reasons for the failure. There were no errors in the campaign settings. This means the problem needs to be looked for elsewhere. We turned to the site.

We put forward 2 hypotheses:

1. Yandex's memory played a cruel joke. The advertising account stores accumulated statistics on the target audience of the resource. For several years we sold plots on this site for 600K rubles and the algorithms remembered this well. Now we needed to sell plots 10 times more expensive – 6 million rubles each, but the algorithms, by inertia, brought us an economical audience ready to buy land no more than a million. Of course, they were not interested in the new village and there were no sales there.

2. Yandex doesn't understand people. Direct's auto strategies provide a general analysis of the effectiveness of a website visit. If the user has fulfilled the goal, for example, left a request, then he is good and we will bring people like him. But on which page of the site the goal was fulfilled is not important to Yandex. But it is extremely important for us. A user who came from an advertisement for an expensive village, and then went to the page of a cheap one and filled out a request form there, is not targeted. How can we train algorithms to understand the intricacies of human behavior?

We came up with a script that creates a separate goal – time spent on the page of a new village. Moreover, the counter stops if a person moves to another project. The target value was set – 5 minutes. We abandoned the previous micro-goal and left only this micro-conversion with pay-per-clicks to get more traffic. And as quickly as possible, train RK on people for whom the prices of the new village are suitable.

We spent quite a lot of money on this hypothesis, but we saw that engagement on our landing page increased. We are convinced that the flow of requests is growing. And we changed the campaign goals to macro conversions. In July, the number of leads increased by 6 times, and qualified ones by 12.

We launched a branded landing page with a minimal sales funnel. The site was also left in advertising. And both resources together increased leads by another 4.5 times

By August the landing page was ready. Based on the experience of the previous project, which was sold out from this landing page, we recommended the client to make the one-pager more personalized. And at the same time, do not overload the user with information. So that there is a feeling of slight understatement, which can be filled by requesting a call back, presentation of the master plan, excursion, or by calling yourself.

Here's what was done:

1. Individual corporate style village with its own logo, corporate colors and elements. The client did not practice this for other villages. It was important for us to create a stable association with the brand among users. This style was broadcast both in the visuals of advertisements and on the main website.

2. Minimum sales funnel. On this project For the first time, we abandoned the funnel through messengers. She gave a lot of contacts, but mostly cold ones and even fraud. We convinced the client to test only the good old forms this time: an application from the website and a call. And our idea of ​​focusing on the hottest requests worked.

Previously, we planned to use only the landing page for advertising. When it was ready, we decided not to rush into shutting down the site for fear of destroying the accumulated results. We thought that we would work for a short time with two places of sale, and then smoothly move only to a one-page site.

To upgrade the new landing page, we started testing 2 tools on it. Master of Campaigns (MK) and Unified Performance Campaign (EPC). We divided campaigns and ads into groups based on broad semantics. The EPK was switched to semi-manual mode, and the semantics in the keywords coincided with the title of the offers. In MK there was no such possibility, because everything there is controlled by automation.

Otherwise, everything was the same: settings, texts and pictures of offers that were liked by those who were interested in our village. Weekly budgets and target cost per lead are also equal – 50,000 rubles. and 1500 rub. respectively. We were wondering which tool would work better.

In a fair fight, the Campaign Master won, he brought 30 leads in a week for 1100 rubles, but EPK – only 6, at a price of 2000-5000 rubles.

It would seem that EPK is a new Yandex tool and should work better. We did not give up hope of getting one and began to use settings to reduce the cost of a lead. But as a result, we only received a decrease in the number of requests. This is how automation defeated our more logical segmentation by keys in EPK.

And when we took the reports for August on the landing page and the website, we saw that together they brought a stunning result. Compared to July, the number of leads increased by another 4.5 times and amounted to 284 pieces. Of these, there were 2.4 times more qualified, that is, 85.

Website + landing page = lead generation synergy. Over the summer they sold not 70, but 89 plots for 550 million rubles. despite the abolition of preferential mortgages and falling demand

It was necessary to decide how to develop the digital strategy further. This is not to say that these two resources did not compete with each other. In August, the number of hits from the site decreased by 35%. But it was from there that we received the most qualified leads. There were 6 times more requests from the landing page and 3 times cheaper, but the ratio for qualified leads was lower.

Nevertheless, the site and landing page worked perfectly together and there was no point in abandoning such successful synergy. So the temporary solution became permanent, and both sites remained in operation. And together with the client, we decided that when launching new objects we would use the same mix.

In May, when we were just starting, there were 198 vacant plots in the new village, and by the end of September, 45% of them were sold for the amount of 550 million rubles. That is, we exceeded the client’s plan and helped sell not 70, but 89 plots during the season. And all this against the backdrop of the cancellation of preferential mortgages and falling demand for real estate.

If you need an increase in calls, order for 0 rub. our in-depth audit of contextual advertising on makodigital.ru

How will this be useful?

• We will find technical errors in campaign settings and segment them by level of danger: minor, significant and critical.

• Let's check the statistics and find the reason for the increase in the cost of circulation and the fall in profits.

• Important! We will recommend positioning, content and development (conversion rate, usability, functionality) for the site.

• We will develop a contextual advertising strategy with new growth points.