Who we are

Digital marketing company PLATZKART has been operating since 2004. We help companies with complex (and not very complex) products and services in the B2C and B2B sectors to develop and implement a marketing strategy, increasing the efficiency of investments in marketing and sales. In addition, we make marketing clear at all stages for company owners.

We have experience in niche areas:

  • real estate,
  • agricultural equipment,
  • fire safety,
  • videoconferencing,
  • relational databases,
  • medical devices,
  • transportation companies,
  • and other high or low competitive markets.

Our company works with projects using the SOSTAC methodology. The methodology allows with 80% probability to guarantee the success of promotion of goods and services on the Internet. We rely only on quantitative data, thoroughly study the market, prepare and implement promotion strategies, and use the accumulated data on the project to optimize it.

As part of the project, we use the following tools to implement the strategy:

  • naming;
  • design: corporate identity development, website design;
  • Creating a website or landing page to sell products and services; ;
  • seo-optimization to present the site in search engines and for participation in auctions of advertising systems;
  • setting up independent analytics to optimize SEO promotion sources and advertising channels, as well as to provide reports to our clients.

PLATZKART is recommended for further cooperation by 90% of our customers. We can provide their contacts on request.

To find out more about our work email us:


Introductory information on the research

In some places, information is withheld under the terms of an NDA (non-disclosure agreement), which is a non-disclosure agreement. It governs how confidential information that is specified in the agreement is handled.

Places with classified information are marked with a special message and highlighted in color.

Sources of information

The data for the study was taken from analytics services, published studies, Client's reports, search engines and competitors' websites. References to the source data and third-party studies are provided at the end of the document in the "Additional Materials" section.

On subjectivity

When analyzing the data we used expert opinions of people found in open sources and the accumulated experience of our company, which means that the results may be partly subjective. We believe that this is, firstly, inevitable, and secondly, not critical for the research value of situational marketing analysis.

Purpose of the study

To collect facts about the current situation on the Internet in the real estate market of Krasnodar region. To study the competitive environment in the Internet. On the basis of the collected facts and the study of the competitive environment to derive theses suggesting the growth of apartment sales in Microdistrict "N". The statements are put into a suitable framework (in this case, the matrix of Osterwalder and Pinier) and offer the best options for strategy and tactics.


Abstracts derived from the study

  • To improve the current situation in advertising campaigns, focus on increasing website conversion rates and improving ad response.

  • According to Yandex research, there is a trend in the market for buyers of all classes of housing to be afraid of missing out on a favorable offer. The main emphasis should be made through the message "don't miss out on a good deal". Use to "explicitly" capture the user's contacts. "Explicit capture" means getting the person's contact information.

... the text is shortened.


Revenue Breakdown

Revenue Breakdown - an analysis that will help you see the "bottlenecks" in the system, acting on which you can get the fastest results.

This analysis is based on the concept of "bottle neck" by Eliyahu Moshe Goldratt. The essence is that for any processes there is a "key constraint" - a stage whose throughput limits and normalizes the entire process.

We created a revenue breakdown model for an advertising campaign in Yandex based on the reports presented in excel tables and data from CallTouch's personal cabinet. The goal is to find the points of impact that will yield maximum results in the near future without requiring a lot of effort.

Disclamer

We realize that apartment sales from this advertising source do not usually go straight to "come > see > buy" and we are dealing with pent-up demand, but by improving your advertising campaigns now you will lay the foundation for increased sales in the 1 to 3 months following the revisions.

Table: revenue breakdown

The data in the table are provided as an example.

Revenue Expenses Shows Clicks CTR CPC Conversion lending Leads CPL Quality leads % into qualitative CPL2 Sales % into the deal CPO Av. check Rejection %BR To increase revenue Difference
120 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% – –
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 30% Reduce the failure rate by 25,49%
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 86 55% Reduce the number of failures by 74
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 800 000₽ 160 55% Increase the average receipt by 30 0000₽
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 7 292₽ 1 500 000₽ 160 55% Reduce the cost to the client by 1 458₽
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 74% 8 750₽ 1 500 000₽ 160 55% Increase the percentage in the deal by 12,31%
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 96 62% 8 750₽ 1 500 000₽ 160 55% Increase sales by 16
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 4 487₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Reduce the cost of a quality lead by 897₽
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 54% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the percentage of quality leads by 8,97%
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 156 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the number of quality leads by 26
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 011₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Reduce the cost per lead by 402₽
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 348 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the number of leads by 58
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 22₽ 1,16% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase page conversion by 0,19%
144 000 000₽ 700 000₽ 6 000 000 30 000 0,50% 0,50% 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Reduce the cost per click on 12₽
144 000 000₽ 700 000₽ 6 000 000 30 000 0,60% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase CTR by 0,10%
144 000 000₽ 700 000₽ 6 000 000 36 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the number of clicks on 6 000
144 000 000₽ 700 000₽ 700 000₽ 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the number of hits on 1 200 000
144 000 000₽ 840 000₽ 6 000 000 30 000 0,50% 22₽ 0,97% 290 2 414₽ 130 45% 5 385₽ 80 62% 8 750₽ 1 500 000₽ 160 55% Increase the average daily budget by 140 000₽


Results

When creating the model, we set a goal to increase revenue from this sales channel by 20%. We got the following results:

The data in the table are for illustrative purposes. Not relevant to reality

To increase revenue Difference
Reduce the failure rate by 25,49%
Reduce the number of failures by 74
Increase the average receipt by 30 0000₽
Reduce the cost to the client by 1 458₽
Increase the percentage in the deal by 12,31%
Increase sales by 16
Reduce the cost of a quality lead by 897₽
Increase the percentage of quality leads by 8,97%
Increase the number of quality leads by 26
Reduce the cost per lead by 402₽
Increase the number of leads by 58
Увеличить конверсию страницы на 0,19%
Reduce the cost per click on 12₽
Increase CTR by 0,10%
Increase the number of clicks on 6 000
Increase the number of hits on 1 200 000
Increase the average daily budget by 140 000₽

The color coding in the table indicates the level of difficulty:

  • green is a relatively quick and easy way to increase revenue by 20% based on current performance
  • yellow - medium level of difficulty

So, based on my own experience, increase site conversion by 0.19% and raise ctr by 0.1% is more feasible than reducing cost per click by 20%. or reduce the share of bounce or garbage requests. Given the results of QFD analysis to do this work will not be difficult.

Let's assume a way related to increasing the budget. So by increasing the average monthly budget by 140,000₽. we can achieve our goals, provided that the market volume will allow us to buy an additional number of impressions. Also, the "game" with the budget refers to the cost per click.

... the text has been shortened.


Predictive analysis of appeals

Let's say we managed to raise the conversion rate of the site to n. What results can we expect, given that now, on average, we receive m hits per month from Yandex Direct?

Let us perform calculations in RStudio using the Poisson probability distribution method:

The results show that there's a 92% probability of getting between 340 to 410 conversions monthly from "Yandex Direct". If we take as a constant the conversion rate of the site equal to n.

... the text has been shortened.


Analysis of appeals by 6 sigmas

The main task of Six Sigma analysis is to learn how to distinguish between systemic and random, i.e. what data can be considered statistically non-random. Understanding the power and limits of the system, for example, what the average number of hits an advertising campaign in Yandex can produce, we can see anomalies and analyze them in order to eliminate or improve the performance of the system. So with the help of this analysis we can step by step improve the performance of advertising campaigns and other sources of traffic.

Sigma is the ratio of the mean spread for some number of subgroups divided by the D2 factor of the Schuchart maps for that number of subgroups.

What the analysis of appeals in the Yandex advertising campaign showed us.

Graph descriptor

The red dashed line represents the average number of requests from May 2020 to April 2021. Blue indicates the three sigma boundaries.

Going beyond these sigmas means going beyond the normal production capacity of the system. That is, if you see such an overrun, you need to understand its cause. If it is beyond the upper limit (i.e., a positive impact), then understand the cause and implement it in the system. If it is beyond the lower limits, then understand that and do not allow it from now on. Thus, step by step, gradually improve each traffic source.

... the text has been shortened.

Graph interpretation

The pandemic period heavily skews the six sigma boundaries and the arithmetic mean, as does the spike in interest. So we remove this period and analyze the data from November 2020 through April 2021.

Graph interpretation

We got more uniform data, but still there are frequent exits outside the system boundaries. We see that "Yandex.Direct" generates from 5 to 16 requests per day - this is the capacity of the advertising channel for the current period. It is not known whether it is good or bad, but this is the efficiency with which this system works.

... the text has been shortened.


Analysis of the arithmetic mean of requests

Schuchar and Deming identified two criteria that should indicate growth and control of the system:

  • arithmetic mean increases or decreases depending on the parameter.
  • sigma boundaries are getting narrower and narrower. By reducing the variability of the parameter we take control.

Knowing the power as well as the trend of the arithmetic mean we can tell if we are controlling the advertising source, i.e. working on improvement.

The graph shows the average number of requests for 22 weeks from November to April. From November to December, the average value of references gradually decreased. Then January and February with low values of the average. In March there is a return to the previous values.

... the text has been shortened.


Analyzing quality leads

Analyzing "qualitative leads" (blue line on the graph) we noticed a high correlation with "primary referrals" (black line on the graph). The results of the correlation analysis by the nonparametric Spearman method rho=0.7592431 with p-value=3.186e-12.

Теоретически это означает, что увеличив кол-во входящих «первичных обращений» увеличим и кол-во «качественных обращений». Как это сделать быстро показано в части «Revenue breakdown».

… текст сокращен.


Target audience analysis

In this part we will look at the target audience and their characteristics. We will calculate the market size using the TAM SAM SOM model. We will talk about the Clayton Christensen JTBD model and why it is important. This part is developed based on our own research and Yandex research. All references to research and "raw" data are given in the "Additional Materials" part.

JTBD

Christensen formulated a theory he called "Jobs To Be Done" (JTBD). JTBD touches on more different aspects of the market than the study of individual groups, so it is necessary to study not only target audiences and socio-demographic groups, but also the types of work that customers do with a product, remembering that the same customer can do different jobs with the same product at different times and in different circumstances.

For example, let's say we have a restaurant in an office building. Every day we have office workers coming in. Some for lunch, some for breakfast, and some for breakfast, lunch and dinner. So we can have the same person "hire" the restaurant up to three times a day.

That's why it's important not to describe fictional profiles of people in the style of "this is Sveta, she's 20 to 45, she has two children aged 1 to 20 and a machine... a washing machine", but to know why your company is being hired. Because Sveta is a fictional character who lives in the head of the profile writer and the company will not benefit from this knowledge.

But knowing the JTBD we can more accurately determine the volume of the consumer market, which is impossible to do knowing only the portrait of the CA. Let's go back to our restaurant example and calculate the market size. The restaurant knows that 1000 people work in the office building every day. Then we have 3×1000=3000 potential requests. This is the JTBD volume, aka market volume. Let's say the restaurant sells 300 lunches, dinners and breakfasts per day, so the restaurant has a market share of 10%.

In this paradigm, we can also think about the real estate market. Having studied the market and real estate research from Yandex, we have identified the following important JTBD customers in the market of newly built houses.

The table has been abbreviated

JTBD JTBD in detail
Improve housing Improve housing conditions
Change rented accommodation for your own, not to spend money on rent
… …

However, it is not enough for us to simply know about JTBDs. We need to understand the volume of each JTBD in order to identify the largest segment in terms of potential profit. To analyze the market capacity, we will apply the TAM SAM SOM model and data from Yandex research (links to the research are provided in the "Additional Materials" part).


TAM SAM SOM

Let's decipher the model's acronyms:

  • TAM - Total Addressable Market
  • SAM - Served/Serviceable Available Market, that we could occupy with unlimited resources.
  • SOM - Serviceable & Obtainable Market, that we can take now with our current level of resources.

Using the data on the percentage of JTBDs from Yandex research, we will calculate the market volume for each JTBD, taking into account the approximate number of queries from the Google search engine. According to Yandex Radar data, Google occupies about 40% of the market, which gives us about 60% of the current volume of queries from Yandex.

Table: number of queries from Yandex and Google per year in SFD

The table is abbreviated

JTBD TAM % JTBD SAM
Improve the housing 19 200 000 47% 9 024 000

…

Table: number of requests from Yandex and Google per month in SFD

The table is abbreviated

JTBD TAM % JTBD SAM
Improve housing 1 600 000 47% 752 000

…

ЕApproximately 1,600,000 searches are generated monthly from both sources, with the majority of searches - 47% related to JTBD "Home Improvement".

Combine the obtained data with the data on the average CTR of ads from the personal cabinet of the CallTouch service.

Table: current market coverage

The table is abbreviated

JTBD TAM % JTBD SAM Reach SOM
To improve housing 1 600 000 47% 752 000 19 270
Total 2,76% 41 000

…

Thus, we have the idea that "N" covers approximately 3% of the target audience. Taking into account the volume of SAM we know that there is definitely a potential for growth.

It should be taken into account that these are approximate data, nevertheless they create a coordinate system in which we can orient ourselves in relation to other market participants. More precise data on the volume of segments will be given in the part of the analysis of selection factors, as the data will be based on the results of a survey of the target audience in the Krasnodar region.


Characteristics of consumers of comfort-class housing

In February 2021, the companies "Yandex" and "Brand Pulse" conducted a study in the course of which the characteristics of buyers of different classes of housing. Below is the data on buyers of housing of the "Comfort" class, as the microdistrict "N" belongs to this class.

Data on sex-demographic characteristics and social status

Comfort class consumers are younger and have a more pronounced core audience than other CAs.


Leisure activities and digital consumption

The main types of leisure activities are visiting shopping malls, watching videos and surfing on social networks.


Sensitivity to advertising sources

Consumers are sensitive to promotions, noticing video and banner ads equally.


Services used

Comfort class buyers:

  • listen to Yandex.Music and Youtube Music;
  • use Yandex.Taxi;
  • are buying on Avito and AliExpress;
  • actively consume content on the Youtube video platform.

Points of growth

... the text has been shortened.


Analyzing the factors of apartment selection

Design and objectives of the survey

The purpose of the survey is to identify the most and least important criteria for choosing an apartment using QFD analysis. The survey data is also used to build a value-price map.

The online survey was conducted through the "Yandex View" service for users with interest in "real estate" in the Southern Federal District.


Sample formation

As we already know, an average of 1.6 million queries are generated monthly in the Yandex and Google search engines - this is our general population. From it we form a sample for the online survey according to the formula:

\[SS=\frac{Z^2×p×(1-p)}{c^2}\]

where

  • \(Z\)- confidence probability, indicates how likely a random response is to fall within the confidence interval (e.g., 1.96 for a 95% confidence interval)
  • \(p\) - Percentage of respondents of interest or responses (0.3)
  • \(c\) - confidence interval. It can be understood as a margin of error, sets the spread of the part of the distribution curve on either side of the selected point where the answers may fall (0.05=±5% ).

We obtained a sample of 458 people, and we need to obtain approximately 137 fully completed questionnaires. In the service we specified a sample of 500 questionnaires. As a result, we received about 600 questionnaires, of which 175 were completely filled out.


Survey design

The questionnaire included the following questions: the first distribution question: "Are you going to or have you already bought an apartment in a new building in Krasnodar?". The purpose of the question was to distribute respondents into three groups:

  • Group 1 - are planning or have already bought an apartment;
  • Group 2 - plans to buy on the secondary market;
  • Group 3 - does not plan to buy;
  • ... items deleted;
  • the last question, please indicate gender and age.

This block includes respondents of all groups. It is important to understand what target audience the service has surveyed and whether the data compares with the results of the Yandex research, which we discussed in the part "Characteristics of consumers of comfort-class housing".


Survey results

Age of respondents


The main part of the survey audience was women from 25-44 years old and men of the same age group. and men of the same age group. The data correlates perfectly with Yandex research data.

...text is shortened


Distribution by group


We can extrapolate this distribution to the general population according to statistical principles. In the role of the general population we have the number of monthly requests from search engines. Thus, we obtain the following distribution:

Gene set Intent Distribution Qty.
1 600 000 Interested in buying 17% 270 000
Secondary market 12% 190 000
The others 71% 1 140 0000

It is important to note that the table specifies the number of requests, not the number of people. That is, one person can make several requests during a month. Nevertheless, this number can serve as a reference point when analyzing advertising campaigns.

The " others" segment includes those who bought an apartment, and everyone who is somehow connected with real estate from realtors to market analysts. This segment is always larger in all niches.


Reasons for buying an apartment


The survey results highlighted the top three reasons (JTBD) for buying an apartment:

  • relocation to Krasnodar
  • …

We can now adjust the segment volume by JTBD according to our survey results.

The table is abbreviated.

JTBD TAM %JTBD SAM
Relocation to Krasnodar 270 000 29% 78 000
… … 25% 68 000

...the text is abbreviated


Quality Analysis Function (QFD)

QFD, or building a quality function deployment matrix, helps you calculate the factors by which a customer chooses your product.

...the text is abbreviated

The QFD matrix needs to be constructed if you don't know:

  • what features of the new product are important to the buyer;
  • what positioning to choose for an existing product or service;
  • what positioning to choose for your company;
  • What to focus on in the content strategy (what to talk about and from what angle);
  • What are the most important touch points for your customers.

There are two approaches to constructing a QFD analysis. The first is a shortened version of "small QFD" - where we analyze through beta-weighted survey results and identify the most important criteria. This is the approach used in this study.

...the text is abbreviated

Survey results

Consumers of comfort-class housing identified the following main selection criteria:

  • proximity of social and transportation infrastructure;
  • reputation of the developer;
  • favorable conditions when buying.

You should pay special attention to these factors when building communication with customers. Use them ...the text is abbreviated.


Analysis of the competitive environment

Selection of competitors

The methodology of competitor selection is based on the hypothesis that the more impressions a competitor's domain has for the selected keywords (i.e. the greater the share of coverage), the more traffic the domain receives, which means it is our competitor on the web.

To select competitors on the Internet, we uploaded semantics from Yandex.Metrics, requested a file with existing semantics and additionally parse the keys in both search engines. As a result, we selected for analysis in the search engine about 1400 keywords and 500 in the paid search engine. This is enough to assess the competitive situation on the web.

As a result, the semantic kernel revealed unique domains (not repetitive) competitors:

  • Yandex.Direct - \(312\)
  • Google Ads – \(71\)
  • Yandex search results - \(610\)
  • Google search results – \(606\)

The presence of this number of unique domains in each channel shows the level of competition in these channels.

Competitors were further selected for analysis based on the following criteria:

  • the largest coverage of the semantic kernel. I.e. the frequency of occurrence of the domain by the analyzed keywords was taken into account;
  • are present in most channels of attracting traffic - i.e. they are present in the organic rendition of Yandex and Google, and in the paid rendition of these systems;
  • are direct competitors - that is, they provide the same services and products as you and in the same region;
  • domain is not an aggregator site or marketplace.

As a result, the following direct competitors are selected for analysis:

  • ...the text is abbreviated


Attraction

In this section we will answer the question from what sources competitor sites attract traffic to the site. Which sources have strong competition, and which sources are weakly used.

Traffic sources used by competitors

Data on traffic sources were collected based on the analysis of competitors' sites using SimilarWeb service. The histogram shows aggregated parametric data on traffic to desktop devices, i.e. it is an average indicator for the analyzed competitors.

The data will help you understand your traffic patterns, and help you figure out which sources are expected to compete strongly for traffic, and which sources are given little attention.

The histogram shows that companies actively use both organic traffic and paid traffic sources. SimilarWeb service records transitions from paid and free Yandex search engine output in the "Referrals" source, so the traffic from this source is correctly distributed among the "Organic" and "Paid" sources.

...the text is abbreviated


Visitor traffic on competitors' sites

The service does not accurately provide data on the number of visits to competitors' sites, in addition, there are only data on visits from desktop devices. However, by measuring the indicators of competitors' visits in the metrics of the service and comparing them with the indicators of site "N" we can understand our place in the competitive environment.

The data are as follows:

The table is abbreviated

№ Competitors Visits
1 … 49 333
2 … 44 000
3 … 31 667

It should be noted that the table presents data on visits only from desktop devices and does not take into account data on site visits from mobile devices. However, the presented data is quite enough to identify strong competitors on the Internet. It is with these sites that the fight for traffic will take place by semantics and other targeting methods.

In terms of traffic distribution by device type, 66% of traffic comes from mobile devices.

It is important to take this into account, as the first acquaintance with the company's website is usually from a mobile device, especially at the level of unformed demand. First of all, the site should be optimized for mobile devices.

Next, we'll take a closer look at the structure of paid and free traffic sources in the SERPs.


Search Engine Traffic (SERP)

Having collected the primary semantics, we used software tools for data collection (Python) and analysis (RStudio) to see which of the competitors occur most frequently in the paid and free rendition and who has a greater coverage on the semantic core.

The paid rendition involved ...text abridged

Table index "Shows" - shows the total number of shows by semantic core, i.e. if "shows" is equal to 492 it means that the domain was shown 492 times out of 520 keywords. Indicator "% coverage" - shows the share of coverage from semantics in relative terms.


Paid rendition of Yandex

A total of 312 domains were found in the paid issue for Krasnodar. This is quite a lot, which indicates strong competition in this source.

The "Average position" indicator in Yandex Direct serves as an indirect indicator of who pays the most for a click in the auction. The closer it is to zero, the higher bids a competitor makes.

Table: competitors in Yandex paid rendition

Table abridged

№ Domain Shows Av. position % coverage
1 mrqz.me 492 10 95%
2 Avito.ru 447 1 86%
3 подбор-квартир-в-краснодаре.рф 388 9 75%
  • ...text abridged

In first place is the site mrqz.me - a service for creating quizzes. Below are the companies and residential complexes that use this service (arranged in descending order of frequency of use):

Also among the domains participating in the auction a significant share of aggregator sites was noticed. These sites, judging by the average position, pay the most for a click, such as "avito" and "kayn", and have the largest coverage by keywords.

Among direct competitors it is worth mentioning the following sites …

It is interesting to note the presence of the resource zen.yandex.ru in the output. Articles with the following headings are ranked by keywords:

  • Why are housing prices rising so rapidly...
  • How we lived in Southern Russia for 12 years and why we ended up...
  • Actual layouts and prices for apartments...
  • ...the list is shortened

The articles mostly belong to bloggers. The Zen page of the company "N" is advertising material and it is the only developer that is seen in paid distribution through this platform. There is also the company "N" providing analytics and news of the real estate market.

...text shortened


Auction in Yandex Direct

We did an additional mini study to see if the rate varies by domain.

Loading in the interface of "Direct" groups of ads, where in each group one ad targeted by the key "buy one-bedroom apartment" we received the following data on rates for the volume of traffic equal to 100%.

The table is abbreviated

Residential Complex URL Forecast 100% Charged off
633 414
517 344
602 397

You can see that all landing pages have different predicted prices and debit price. The landing page has the lowest deductible price.

Further analysis of landing pages in the "Engagement" part will show that this landing page has the last place in the ranking in terms of outline disclosure (the first place is occupied by "Residential Complex Fresh"). However, Yandex gives them a lower bid. It is possible that the advertising campaign inside the interface is competently configured, or use bid-manager, which allows you to keep a low rate.


Paid Google rendition

The section is abbreviated


Yandex organic traffic

The table is abbreviated

№ Domain Shows % coverage
1 avito.ru 1577 115%
2 krasnodar.cian.ru 1430 105%
3 realty.yandex.ru 1367 100%

As we can see, aggregator sites have the largest coverage on the analyzed kernel. That is, when a person enters a query related to the purchase of an apartment, then with a high degree of probability on the first page he will see sites from this list. As you can see, there is not a single competitor site in TOP-20.

...text abridged


Google organic traffic

The section is abbreviated


Sites of direct competitors in search engine rankings

As for the analyzed competitors' domains, they are practically not present in the search results, which is evident from the table on the share of semantics coverage. However, site N stands out among all competitors, especially in Google, which indicates good optimization of the site for this search engine.

The table is abbreviated

№ Domain Coverage in Yandex SERP Coverage in Google SERP
1 … 5,3% 26,7%
2 … 1,2% 6,9%
3 … 0,8% 0,6%

...text shortened


"Vertical SEO"

Vertical SEO strategy means placing directly on aggregator sites, marketplaces and other similar sites. Considering the situation in SERPs, this is an optimal strategy, as developers' sites can receive search traffic mainly on branded queries. By placing on as many of these sites as possible there is a chance to get additional traffic.

First of all, you should consider the following resources on which you can place ads:

  • avito.ru
  • krasnodar.cian.ru
  • realty.yandex.ru
  • ... the list is shortened


Social media

The section is abridged


Media advertising

The section is abridged


Brand Lift

This metric shows how well known a brand is among Yandex users in Krasnodar and what the trend of the studied brands is. The data is taken from the "Yandex.Wordstat" service.

The red line in the frequency zone from 4,000 to 6,000 denotes the brand of the company "N". The trend line clearly tends to grow, the brand is becoming more recognizable on the Internet in Krasnodar.

Three companies are located in the frequency zone from 2,000 to 4,000:

  • ... abbreviated

Point of growth

...text abridged

The situation in all Russia is somewhat different, in the frequency zone from 10,000 to 15,000 to 15,000 is company "N".

In the zone from 5k to 10k, companies "N" (red line), "N" and "N" are located.

...text abridged

Table: correlation analysis of brand mention frequencies by company

Abbreviated...

If we aggregate the frequencies of all brands, we get a metric that indirectly shows what's happening with demand in the real estate market.

In this niche there is a good trend of demand growth, which means that in the coming year we expect an increase in users interested in buying apartments on the primary market. We will further investigate this hypothesis through trend analysis in RStudio.


Involvement

In this section, we will show how companies conduct initial communication with customers, what engages them and how they keep their attention. We will analyze the market for the degree of its maturity through a value-price map.

Positioning

Our Moscow colleague I. Balakhnin, a marketing expert, has made a wonderful point about positioning in the development sector. We will set out his viewpoint, with which we fully agree, below.

One of the key challenges of the development market is the developer's transition to the logic of client-centeredness. Increasingly, positioning is built not around abstract words and slogans, but around solving specific client tasks.

…

The current TTX of the objects are taken for granted by buyers; what could delight and be a kind of "Moment of Truth" 5 years ago, today has completely stopped working.

Moreover, the decision-making model has also become more complicated: in the conditions of growing supply and high level of price transparency, clients increasingly pay attention not only and not so much to the positioning of residential buildings. In this regard, we identify 5 so-called contours around which developers should build their positioning:


Contours of real estate objects

The inner contour of the "Apartment". Clients are increasingly interested in layouts. It is important for buyers to understand how they will be able to live in this apartment, to what extent it covers or does not cover their immediate needs.

"Entryway". The client is concerned about how the public areas are decorated, whether there are storage rooms, whether the contingent of residents in the entrance corresponds to the social status of the buyer, and whether they simply look like him or not.

...text abridged


Competitor analysis by contours

Having adopted these outlines with minor adjustments (we added a neighborhood outline and merged the entrance with the house) we analyzed the content of landing pages, drew conclusions on the quality of each outline and overall positioning.

We will evaluate each circuit on a staggered ten-point scale. The grading system is as follows:

  • content presence - give it 3 points;
  • completeness - a grouped score consisting of several additional parameters. We add to the previous criterion from 0-3;
  • pitch - a grouped score of 0-3 points;
  • WOW effect - add 1 point to the accumulated score.

We ended up with several tables:

  • The first one is the total page quality summary for all developers;
  • the second is the overall totals in detail by element;
  • third - results for each developer and place in the rating.

In the table below you can see the share of use of each contour on the site for all participants of the study and the quality rate of contours (Quality Rate). Quality Rate - shows what share of sites has a high (more than 7) score for the specified element.

Table: total results by developers

Contours Share of utilization Quality Rate
Apartments 90% 27%
House 54% 7%
Yard 69% 3%
Neighborhood 92% 17%
Developer 88% 46%

The "apartment" contour is represented in one way or another in all of them, but not all elements of this contour are handled by developers (see the following table).

The "House" and "Yard" outlines are least often used by companies. As a rule, it is a reference with an inexplicable description and a photo.

But we speak well of ourselves, as evidenced by the final scores on the "builder" contour.

Table: totals by element

Table Description:

  • column "Study elements" - elements that make up the contours.
  • "Average Score" is the average score for the element across all builders.
  • Values in the columns "Excellent", "Average", "Acceptable" - show the number of companies whose elements correspond to the specified quality level.
  • "Share of utilization" - the share of usage of the element on the landing page among developers.
  • "Quality Rate" - shows what proportion of sites has a high (greater than 7) score for the specified element.

The table is abbreviated

Contours Elements of the study Average score Great Average Acceptable Share of utilization Quality Rate
Apartments Content on 1-2-3 apartments 5,5 2 9 1 100% 17%
Advantages of 1-2-3 apartments 5,0 1 9 2 92% 8%
Finished apartments 4,5 1 3 8 67% 8%
Apartment selection 6,7 9 2 1 100% 75%

On the contour of "apartments". Most companies pay a lot of attention to the functionality of apartment selection, but often forget to specify the advantages of apartments. From the content on apartments use schemes, very few photos, design 3D models, and only on one site you can walk around the apartment and look in the windows.

Remember that this is one of the most important contours, how the company discloses it will determine the decision whether the developer will be listed as an option, holding all other things equal.

...the text is abbreviated

Table: totals for each developer

The table is abbreviated

Companies Residential Complex Place in Rating Average score Sum of points Completeness Quality Rate Place Sum of points Place Comprehensiveness Place QR
… … 1 6,8 116 94% 56% 1 2 1
… … 2 5,2 93 100% 22% 3 1 2
… … 4 5,4 81 83% 22% 4 4 2

Growth points or what we should do with this data

Company websites do not have a clear structure according to the contours and in most cases the criteria of "completeness" and "presentation" are not realized qualitatively. For example, we start talking about apartments somewhere in the middle of the landing page, about the courtyard somewhere at the very bottom or no information at all, about the residential complex and neighborhood either at the top of the page or in the middle.

Hence, there is no clear positioning of comfort-class residential complexes. The market tends to acquire signs of commoditized markets, i.e. clients choose a developer mainly based on the price per square meter of living space.

Based on the above, we recommend that ...text abridged

We disclose each contour in detail with special attention to the criteria "data completeness" and "feedforward".


Value-price map of the market

A value-price map analyzes price at the level of competitors. The purpose of the map is to know the level of market maturity, which will help answer the question: "do customers understand what they are paying for?" and "what should we do (what strategy should we follow) in this or that case?".

To build the analysis we took the price per square meter in a one-room apartment and data on beta-weights of selection criteria. Then we analyzed the residential complexes according to the materials available in the network (website, social networks, POS) putting the coefficient of "manifestation of the criterion" on a scale from 1 to 5 against each criterion for each residential complex, where 1 - the criterion does not occur in any way, and 5 - is the main dominant criterion. By multiplying the coefficients and beta weights of the criteria we formed the "value" metric.

The obtained data were placed on the Cartesian coordinate system, where on the abscissa axis we indicate the price and on the ordinate axis the value.

The graph shows that market participants tend to line up. This means that the market is approaching a mature stage, when the participants know exactly the quality-price ratio (an example of such a market is the TV market, where there are clear standards and the market participants line up almost in a perfect straight line).

However, in our case there is room for presenting the value by better disclosing the criteria to the competitors. This can be clearly seen in the straight line, when some competitors ask for a high price for an apartment, but present it poorly. Such market players are under the red line.

As for the "N" neighborhood ... text abridged


Trend analysis

This study is conducted in order to see where the interest of users on the Internet is moving. It is growing, stable or falling. Knowing the level of interest we can adjust strategies and plan the amount of resources needed to attract new customers.

Let's investigate:

  • Trends in Yandex and Google search engines
  • create predictive models of demand for the next year


Yandex trend research

Trend analysis was performed by adding the frequencies of queries by months of the main semantic core. The region where the frequencies were collected was the Southern Federal District. Statistics were collected by parsing Yandex.Wordstat service through the Key Collector program. Having collected the data and summarized all frequencies for all keywords in the core, further processing the data in RStudio, we got a general picture of the trend direction.

Trend presentation

As can be seen from the chart, the trend has a clear upward direction, which indicates an increasing interest in this topic. This data correlates well with the data from the Brand Lift analysis.


Time series decomposition

Time series in general can be represented as a sum of three components: trend, seasonal changes and noise (fluctuations associated with random factors). A good understanding of the structure of a time series is an important prerequisite for building reliable forecasting models.

We decompose the time series into a trend using local polynomial regression (STL) while converting the frequencies to a logarithmic scale for better visual analysis of variance.

The plot shows the components of the time series extracted using the STL method:

  • Log - analyzed time series;
  • trend;
  • remainder — leftovers.

If we sum the components, we get the original series (shown in the upper graph). The gray vertical bars to the left of the graphs represent the specific contribution of each component to the total variance in the data.

The graph shows that trend makes the largest contribution to the total variance, as the range in variance is almost identical to the time series, while noise (remainder) makes the smallest contribution, as shown by the range in logarithmic scale (covers decimal fractions between 0 and 0.1).


Seasonality

Let's plot the seasonality over two years by aggregated keyword frequencies.

You can see seasonal fluctuations in the graph. For example, an increase in interest in the summer months and a gradual decline towards the end of the year. There is a noticeable decrease in demand from March to April. At the same time, seasonality has an accumulative effect, i.e. the number of requests increases from year to year. It is also worth noting a sharp increase in requests in 2020.


Google trend research

The section is abbreviated


Predictive models

To build forecasts for the time series under study in Yandex, we use either the ARIMA (autoregressive integrated moving average) method, or the triple exponential smoothing method, or the Holt-Winters algorithm. The Holt-Winters method works well when there is a clear periodicity/seasonality in the data structure. For example, in B2B niches, customer activity declines on weekends and rises on weekdays. We observe a typical wave-like cycle in business.

It is possible to predict the number of leads, sessions, sales or calculate other metrics for a future period quite accurately using these predictive models. In this study, we predict the frequency of requests and the direction of the regression line.


Forecast for Yandex

Based on the trend data obtained in Yandex, we conclude that there is a weak seasonality in the time series and the ever-increasing level of requests in the real estate topic plays a major role in the formation of data. In this case we can use both ARIMA or Holt-Winters methods and choose the best model according to the results of the mean absolute percentage error (MAPE) test.

Forecasting model using the Holt-Winters method

On the graph, a red dashed line of predicted mean values runs in the green area, which indicates the 80 percent confidence interval (i.e., values with an eighty percent probability of being in this colored area). However, we are not interested in the exact data, but in the trend (red line). As can be seen from the graph, the trend for the next year shows the growth of queries in the search engine Yandex.


ARIMA forecasting model

We observe the same picture in the ARIMA model. Interest is expected to grow in the coming year in Yandex.

Let's check the models for mean absolute specific error (mean absolute percentage error, MAPE) in RStudio.

Both models have low specific error, but the best model in terms of MARA is the Holt-Winters.

In terms of absolute and relative values, in 12 months from June 2021, demand for SFD is expected to grow by about 45% according to the Holt-Winters model, amounting to about 1 million requests.


Google Forecast

The section is abbreviated


Points of growth

We found out that the number of queries in Yandex will grow, while interest in Google is at a fairly high level. Yandex gives more options for solving issues related to real estate promotion, due to its own services and various aggregators. Therefore, further growth of interest in the Yandex search engine is expected.

...text abridged


Further action

Strategy

Necessary:

  • Optimize the existing customer engagement system according to the Revenue breakdown report.
  • Monitor the power of advertising systems to generate quality leads according to six sigmas analysis.
  • Separate the systemic from the accidental. When we go beyond the boundaries of the system, analyze why it happened and, depending on the situation, implement or make it not to happen.
  • ...the list is shortened

We need to consider the maturity of the market in terms of the focus of competition.

According to this model, all market players go through several stages of focusing product benefits: focus on functionality. For example, our kettle can heat water strictly to a set temperature.

Over time, consumers realize what level of product functionality suits them in a given niche, and the focus shifts from functionality to safety. I.e. everyone starts to be interested in how safe, not harmful, useful and so on the product is.

Then consumers begin to realize what level of security they need, and what level is already unnecessary, and then after security the focus shifts to service.

However, over time, everyone learns to provide the service and then the market moves into the final stage of focusing on price. Such markets are called "commoditized" markets. In these markets, competition cannot be won except by price.

Given the maturity of the real estate market, in addition to contours and criteria, the focus needs to shift ...text abridged


Tactics

  • It is necessary to conduct media planning of advertising sources based on the data of forecasting models.

...text abridged


Crosscutting analytics

It is important to receive data about transaction statuses from the CRM system into one of the analytics systems in order to optimize advertising campaigns at the revenue level, not by the number of hits. The analytics scheme for Yandex Metrics can look as follows:

Text and graphs are abbreviated...

Contact us for a situational analysis via email: