Automatic research.
Visible evidence.

Estate Signal monitors Egyptian property sources and turns measured search demand into an explorable market board. It does not predict sales or guarantee advertising returns.

Automatic discovery

Public Nawy sitemaps and Real Estate Egypt directory pagination expand discovery beyond the official SODIC, Emaar Misr and Mountain View catalogues. The worker follows supported project and developer pages automatically, retaining source links and observed names. No manual project or keyword entry is required. Source-reported counts are not validated national totals. Pending pages, blocked sources and identity gaps remain visible; no source establishes that every Egyptian project has been found.

Directory-reported identities can enter Google research when a stable source explicitly associates a project, developer and Egyptian location. They remain directory-reported, even after Google returns statistics. Official corroboration is shown separately. Conflicting or incomplete identities stay visible without a recommendation. Explicit phases are counted separately; unclassified records are not silently counted as independent projects. Matching uses developer, location and conservative name normalization, preserving ambiguous records.

Discovery is not a launch date

The first successful import of a source is a baseline. Later additions are marked newly discovered, with an observation timestamp. Neither proves when a development launched. Current Invest-Gate news-page headlines are collected independently; launch-related titles are marked reported and unverified. Existing projects disappearing from a catalogue are retained with a stale-source warning after three days for official catalogues or eight days for directories, not declared sold out or withdrawn.

Automatically generated keywords

Maintained profiles pair catalogue identities with natural English names and Arabic aliases, including spelling variants such as iCity. Ambiguous names keep a developer qualifier; phases keep their own identity rather than inheriting a parent development’s demand. Price, sale, booking and payment phrases are generated locally. Unrecognized projects use qualified names and source-provided Arabic aliases without invented translations. Profiles are research hypotheses, not proof of demand. A profile update automatically queues existing projects for new research without requiring manual entry or a catalogue recrawl.

Google demand evidence

Approved Google Ads access and account authorization are required. The worker requests historical keyword estimates for Egypt on Google Search, across languages, using an EGP account. Returned keyword groups may combine close variants; each canonical group counts once per project. Unreturned keywords stay visible as unavailable. Different projects can overlap, so their combined volumes are not a deduplicated market total. Searches are not unique buyers.

Intent and opportunity

Local rules classify purchase phrases (for sale, buy, حجز, للبيع), consideration phrases (prices, payment, اسعار, تقسيط), general research and irrelevant employment queries. Buying-related searches sum the returned purchase and consideration keyword groups. This is phrase-based evidence, not observed purchasing behavior. Creating a purchase phrase without returned volume adds no buying evidence.

Ad test priority is a heuristic out of 100: up to 30 points for logarithmic total demand (capped at 100,000 searches), 35 for logarithmic buying-related demand (capped at 1,000), 15 for buying-keyword bid affordability, 10 for momentum and 10 for evidence. Affordability uses the volume-weighted high-end top-of-page bid, declining logarithmically from EGP 1 to EGP 1,000, scaled by the share of buying volume with bid estimates. Missing bids contribute no cost points. Evidence adds five points for two positive-volume buying groups and five for comparable momentum. Momentum contributes between zero and ten points (five plus percentage growth divided by ten); unavailable momentum contributes zero.

Missing volume means no score; name-only demand is capped at 39. Priority tests score at least 60, focused tests at least 40; both require positive buying-related volume and fresh evidence. The shortlist excludes source-stale projects. Competition is shown as advertiser activity, not a measure of buyer quality. The score omits conversion rates, inventory, commission and lead quality. A bid range is not actual CPC or cost per lead. These rules prioritize experiments, not profitability.

Timing, history and uncertainty

Momentum compares the latest three months with the preceding three only when every included measured group has six consecutive, complete months. Unreturned groups do not erase measured history; missing points within measured groups remain gaps. The latest search month is displayed separately from the request date.

With twelve complete months, a June–August share of at least 50% flags summer-heavy demand and deducts 15 points outside May–September. One year is not proof of recurring seasonality. A latest month above twice the previous two-month average deducts eight points; below half that average deducts ten. Snapshots older than 45 days or search history more than three months behind today deduct 20 and require a refresh before recommendation. These are transparent caution rules, not statistical forecasts. Missing evidence is never treated as zero cost or guaranteed opportunity.

Refresh schedule and costs

The separate worker checks schedules each minute, reads developer catalogues daily and news every six hours. Google snapshots refresh after 30 days. Directory indexes refresh daily and detail pages weekly. Rate limits pause research for the indicated cooldown; project-specific failures do not stop other projects. Account-access failures require attention. Valid empty Google results are cached too. The board refreshes from stored data each minute; it is not a real-time search ticker.

The app uses public sources and a Google Ads API integration, with no paid data or AI APIs. Your server and advertising spend are separate. Robots policies, bounded downloads, public-network validation and request spacing constrain source collection. The coverage page shows the worker heartbeat, failures and next checks.