Refresh, scroll, repeat
Your team is looking for a certain kind of property: small commercial units in a few postcodes, or houses suitable for conversion under a price ceiling. They check the big portals with saved searches, a dozen local agent websites that list some stock only on their own site, and auction sites. The same property appears three times at slightly different prices.
The good ones go fast. By the time someone has worked through their list, a promising listing has been under offer for a day.
Why saved searches are not enough
Portal alerts only cover that portal. Smaller agents often post first, or only, on their own websites. Auction houses publish catalogues in their own formats. Each source describes properties differently, so filtering on things like tenure, condition or planning status is patchy.
Portal terms are also strict. The large UK property portals generally prohibit scraping, so a naive approach of collecting everything from everywhere is neither wise nor reliable.
| Source | Sensible route |
|---|---|
| Large property portals | Their alerts or an official data partnership, not scraping |
| Agent websites | Feeds from their software, or public pages where terms allow |
| Auction houses | Published catalogues, often downloadable |
| Land Registry and planning data | Official open data and APIs |
What manual tracking costs
Missed opportunities, especially on stock that appears only on small agent sites. Staff hours on repetitive checking. Duplicate listings wasting viewing time. And no history, so you cannot see how long properties sit or how often prices drop in an area.
Inconsistent notes are another cost. When three people track listings in their own spreadsheets, nobody knows whether a property was already viewed, rejected or offered on, and an agent gets two calls about the same unit from the same firm.
Speed is the heart of it. In a competitive market the buyer who hears first gets the viewing first, and a process that depends on someone remembering to refresh a dozen websites will always be a step behind.
How we build a listings aggregator
- We list your sources and check each one's terms and technical options. Many agents use common property software that can publish feeds, and some will share a feed on request.
- For permitted public pages, we build polite collectors that check new listings on a schedule.
- We add open data where useful: price paid history from HM Land Registry, EPC ratings and local planning applications.
- Listings are normalised to one structure: address, price, type, size, tenure, agent, link, first seen date.
- Duplicates are matched by address, postcode and features, so one property appears once with every source and price listed.
- Your criteria filter the results, and an AI step can read descriptions for things that are not in structured fields, such as mentions of needing modernisation.
- Matches trigger alerts by email, WhatsApp or Slack, and everything goes into a searchable database with history.
We will not build collectors for portals whose terms forbid it. Where a portal matters to you, their own alerts or a commercial data agreement is the right route, and we can connect those into the same system.
One list of what fits
Each morning, or as they appear, you see new listings that match, once each, with every source and price. The team spends time on viewings and offers rather than refreshing websites. Over time the history shows how quickly stock moves and where prices are softening.
Each listing can also carry a status your team sets, such as viewed, rejected or offer made, so everyone works from the same record and nobody chases a property someone else has already ruled out.
The history also supports better offers. Knowing how long similar properties sat before selling, and how often asking prices were cut in that postcode, gives your team a stronger basis for negotiating than the agent's description alone.
Does this sound like your week?
- You check many agent websites as well as the big portals
- The same property turns up several times
- Good listings are gone before you see them
- Your criteria cannot be expressed in portal filters