Real estate marketing trends 2026: where budgets are shifting

AI agents move from pilots into daily operations

The first shift of 2026: AI agents in real estate marketing have moved from pilots into daily operations. Every day they manage bids, move budget between campaigns, clean up keyword lists and review call recordings, while a human sets the rules and approves decisions with major consequences. The AI experiments that developers ran in 2024 and 2025 have, in mature teams, become a standard process with a rulebook, defined limits of authority and a log of every action. The difference from the previous wave of hype is in the connection: through protocols such as MCP, an agent reads the ad account, Yandex Metrica and the CRM directly, with no manual exports or copying of spreadsheets. Advice from a chat window has become an action in a live ad account. A pilot answered the question of whether the technology could work at all. Daily operations answer a different question: how much money it returns every week and who is accountable for its decisions.

The practical effect lies in the frequency of decisions. A human reviews campaigns once a week; an agent goes through every ad set each morning and moves the money before a burnt-out campaign eats through the budget. In a street retail case, daily reallocation within the existing budget increased the number of leads twenty-five times with no additional spend. What to do within a quarter: pick one process, most often daily campaign unit economics or call reviews, connect the agent to the ad account and CRM in suggestion mode, and after a month give it the right to make small adjustments within a defined corridor. Call reviews deliver a second quick result: the agent transcribes conversations, finds recurring objections and shows marketing what happens to a lead after it is handed over to sales. One working loop convinces a team faster than any strategy on slides.

Lead costs are rising: what to do about optimisation

The cost of a lead in real estate keeps rising: in major cities an enquiry gets roughly ten to twenty percent more expensive each year depending on segment and region, and there is no reason for it to get cheaper, because the auction is overheated and the solvent audience is finite. The working answer to this growth is a change in the optimisation goal. The teams that win drive campaigns towards a qualified lead and movement through CRM stages, not towards any enquiry at the lowest price. Benchmarks for cost per enquiry in major cities keep the same structure: comfort class roughly from 1,500 to 4,000 ₽, business class from 4,000 to 10,000 ₽, premium from 15,000 to 20,000 ₽ and above. What changes is the bar itself: every year these ranges shift upwards, and holding the economics together with cheap enquiries no longer works.

The mechanics are well known, but only a few teams apply them systematically. Offline conversions from the CRM are passed into the ad accounts: a qualified call, a meeting, a reservation. Automated bidding strategies learn from these events and gradually stop chasing cheap, irrelevant traffic. In an international case across four markets, it was precisely the focus on qualified enquiries instead of cheap ones that produced a conversion to sale of about ten percent and a fivefold return on traffic spend. What to do within a quarter: agree a single definition of a qualified lead with sales, set up the transfer of CRM events into the advertising platforms and move at least half of the budget to optimisation for deep goals. We broke down an honest calculation of cost per lead by segment in a separate article with benchmarks: it shows that the cost of an enquiry in comfort class and in premium differs by an order of magnitude, and comparing them against an overall average is pointless.

Property portals get pricier and demand negotiation

Property portals (classifieds) remain a mandatory channel for new developments, but the terms are getting tougher: the major platforms index their tariffs by roughly twenty to thirty percent a year, and their share of a project's budget often reaches a third or more. Passively paying the tariff is no longer an acceptable strategy. The channel requires negotiation and an honest assessment of its contribution to deals, otherwise the indexation eats up the entire effect of optimising every other channel. The reason for the growth is clear: the buyer audience sits on two or three large platforms, developers compete for the same listings and positions, and the platforms monetise this position through tariffs, paid boosts and lead generation formats.

Judging property portals by last click is unfair in both directions: they genuinely create some deals and claim credit for others, because the buyer had already chosen the residential complex and simply clicked on the listing. The working tool here is post-view evaluation: matching impressions and listing views against deals in the CRM, asking 'where did you hear about the project' during qualification, and running control periods with reduced placement. What to do within a quarter: collect statistics on deals with a portal touchpoint, bring those numbers to negotiations with the platforms and lock in package terms, and in parallel launch post-view measurement so that by the next round of negotiation you have your own estimate of the contribution, not the platform's estimate. Even a discount of ten to fifteen percent on a channel that takes a third of the budget frees up noticeable money for tests elsewhere.

GEO: buyers ask AI models about new developments

GEO is optimisation for generative search engines: working so that AI models cite your project when a buyer asks them which residential complex to choose. The share of this traffic is still measured in single-digit percentages, but it multiplies year on year, and the first developers are already restructuring their websites for citation. The buyer puts the whole request to the AI model at once: district, budget, a school nearby, completion dates, and receives a short list of projects. Getting onto that list is becoming a new field of competition, and for now it is wide open. This is not only about ChatGPT: AI answers are built into Yandex and Google search itself, so even a person who never went anywhere in particular increasingly sees a generated summary before the links to websites.

The good news: basic GEO hygiene coincides with honest SEO. AI models cite pages where the answer to the question is given directly and in a structured way: prices, timelines, instalment terms, transport access, without marketing fog. What to do within a quarter: check how the project appears in the answers of the main AI models for a dozen typical buyer questions, rewrite the key website pages in a direct-answer format, add FAQ blocks and markup, and create an llms.txt file. Then repeat the brand mention measurement once a month. This is cheap monitoring of a channel that in two or three years may become a noticeable source of enquiries, and it is wiser to take a place in the answers before all the competitors arrive.

Short video and Telegram warm up the long cycle

Short video and Telegram have established themselves as the main channels for warming up the long cycle: a buyer chooses a new-build over three to six months, and throughout that time the project has to stay in front of them. Clips from the construction site, floor plan walkthroughs and breakdowns of mortgage programmes build recognition long before the enquiry, while the project's Telegram channel keeps a warm audience one message away from the sales team. The formats have already settled: vertical videos on VK Clips, Shorts and Reels for reach, a channel for retention. Performance brings the enquiry; these channels make sure that by the time of the enquiry the buyer already knows the project and trusts it.

The economics of these channels are calculated by their influence on the whole loop, not by last-click enquiries: growth in branded searches, a lower cost per enquiry in performance, the share of subscribers who reach a meeting. A subscriber to a project channel costs roughly 100 to 400 ₽ depending on segment and region, and this is one of the cheapest forms of a long touchpoint. In the case of a waterfront residential complex, consistently presenting the project's main advantage visually produced a sevenfold return on traffic spend. What to do within a quarter: set up regular filming without studio polish, two or three pieces of content a week, link the video and the channel to retargeting, and add a touchpoint tag for these channels to end-to-end analytics so that their contribution to deals is visible rather than assumed.

Marketing shifts towards the buyer's financing scenarios

In 2026 new-builds compete with the secondary market and with the level of the mortgage rate, so marketing is shifting from aesthetics to working through the buyer's financing scenarios: instalments, trade-in, subsidised programmes, combinations of down payment. The purchase decision comes down to the monthly payment, and the project that is first to calculate that payment clearly for a specific family's situation gets the meeting. A beautiful render no longer does this job. A buyer with free cash compares three scenarios: buy a new-build, buy a cheaper ready secondary flat, or leave the money on deposit and wait. Marketing that only answers the question 'why is our building prettier than the one next door' does not even take part in that comparison.

The shift changes creatives, landing pages and call centre work all at once. Instead of facades, the ads run on specific scenarios: instalment payments until completion, trade-in of the old flat at a clear valuation, a comparison of the payment with rent for the same flat. In the Northern Cyprus market, focusing on investment and payment scenarios produced an eightfold return on marketing spend and deals above a million euros. What to do within a quarter: build a matrix of the project's financial instruments, create separate landing pages with a payment calculator for the two or three most common scenarios, teach managers to calculate the payment on the very first call, and add financing scenarios to the qualification scripts so that a lead with a trade-in is not lost as an ordinary enquiry. The range here is honest: the share of deals with instalments or a trade-in varies across projects from a tenth to a half, and only your own CRM will show the exact figure for your segment.

Stack consolidation: a single data loop

The seventh shift: developers are tired of the zoo of contractors and are pulling marketing into a single data loop, where advertising, analytics, CRM and AI agents work together. When one agency runs search, a second runs targeted ads, a third runs analytics and the developer handles the property portals itself, nobody has the full picture, and every contractor reports on the metrics that make it look good. Consolidation gives the budget owner back a single version of the truth: what a deal costs through each channel and where to move the next rouble. This is first of all about unified data: several contractors may remain, but they report into one system and by the same calculation rules, from spend to signed contract.

A single loop pays for itself on decisions that are impossible to make without it. In the case of a premium residential complex in St Petersburg, a channel looked unprofitable in its first month, and on the contractor's standalone report it would have been shut down. The full picture down to signed contracts showed a sixteenfold return over a long window. At Matveo we start precisely here: a two-week marketing audit in which we review the channels, the analytics and the handover to sales, and show which of the seven shifts will deliver results for your project fastest and in what order to implement them. The cost of the audit is credited against the pilot, so testing hypotheses does not mean paying twice.

Frequently asked questions

What works in real estate advertising in 2026?Optimising campaigns for deep CRM goals, daily budget management through AI agents, financing scenarios in offers, and warming up with short video and Telegram. Property portals continue to bring deals, but require negotiation and an assessment of their contribution. The general principle: a combination of channels with a single data loop wins, not one magic channel. There is no universal recipe for budget shares; the proportions depend on the segment, the region and the stage of the project.

Lead costs are rising, what should be done first?Change the optimisation goal. Pass events from the CRM into the ad accounts: a qualified call, a meeting, a reservation, and drive campaigns towards them rather than towards any enquiry. The cost per enquiry may even rise, but the cost per deal will fall, because the budget will stop going to cheap, irrelevant traffic.

Is it worth dropping property portals because of rising tariffs?In most projects, no: for new developments they remain a channel with a large share of ready demand. The sensible strategy is to assess their real contribution by deals, including post-view, and negotiate package terms with those numbers. A full switch-off is tested in a controlled way, over a short period, measuring the effect on branded searches and deals.

Does a developer need to work on GEO right now?Yes, but with limited effort. The share of buyers who ask AI models is still small, but the cost of entry is low: structured pages with direct answers, FAQ markup and a monthly brand mention measurement. Whoever does this hygiene first will hold a place in the answers by the time the channel goes mainstream. The same changes also improve ordinary search results, so the work pays off either way.

Where to start rebuilding marketing for the 2026 trends?With diagnostics, not with buying tools. First you need to understand where the loop loses the most: in the optimisation goal, in overpaying the property portals, or in the gap between marketing and sales. A two-week audit gives you this map, then one shift is implemented per quarter, starting with the one worth the most money. An attempt to rebuild everything at once usually ends with nothing being finished.