Overview
Property is the industry where the internet changed buyer behavior most completely and agency operations least, and Hoshna Realty wanted the whole thing built properly.
Buyers now do almost all of their searching online, filtering, comparing, shortlisting, before they ever call anyone. But behind the scenes, a large share of the industry still runs listings on spreadsheets, WhatsApp and memory. Hoshna Realty, run by Abdur Rahman and based in Türkiye, wanted a portal buyers actually search on, an app for the way people browse now, and an admin system the business runs on daily. The brief sounded like one project. It was three, and they all had to agree with each other.
Challenge
Sale and rent share the word property and almost nothing else, sale listings care about price, financing and long decision cycles, while rentals care about availability dates, tenancy terms and speed, modeling them as one thing with a toggle looks efficient in week one and becomes a permanent tax on every feature after that. A listing edited in the admin panel has to appear correctly on the web portal and in the app, with the same photos, the same price, the same availability, any drift between them is visible to customers and embarrassing to the business. Property listings are also photo-heavy by nature, ten to twenty images per listing multiplied across the catalogue on mobile connections, so image handling isn't a detail, it's the performance story. Inventory changes constantly too, properties sell, rent, come off market and change price, so a platform where updating a listing requires a developer will show wrong prices within a month. And everything in the platform exists to produce one outcome, a serious buyer contacting the business about a specific property, if that path breaks, nothing else matters.
The Decision That Shaped This Build
Sale and rent were split into separate models from day one, rather than one listing type with a status flag, and that decision was the point, not overhead. It meant a rental could never surface a mortgage calculator and a sold property could never keep accepting enquiries, an entire class of bugs removed before it could exist. It also meant the admin panel, not the website, was built as the system of record, with the web portal and the mobile app both consuming it, so the app could be added later without renegotiating where the data actually lives. Building the portal first and treating admin as an afterthought is the common path, and the one that leaves two products quietly disagreeing with each other in ways nobody can fully fix. Deciding where the data lives, and how many products it actually needs to describe, before deciding what any of it looks like, is judgement work most clients don't know to ask for.
Approach
Sale and rent were modeled as related but distinct from the start, shared where they genuinely overlap, media, location, property attributes, and separate where they diverge, pricing structures, availability, and the enquiry context an agent needs when the call comes. Search was designed around how buyers actually filter, not every field a database can hold, but the handful people genuinely use, where, what type, how much, sale or rent. The media pipeline was treated as core engineering, correct sizing per surface, appropriate formats, lazy loading in galleries, since on a property portal this decides both the browsing experience and the hosting bill. The platform was also built bilingual, Turkish and English, serving local buyers and overseas buyers looking at Turkish property from one listing database rather than two separate sites. And the admin was built for the client's team, held to the standard of whether the business can add a property, change a price, or mark something sold without calling us.
Result
Hoshna Realty is live in production across all three surfaces, web, app and admin, at hoshnarealty.com. The client's team runs listings, pricing and enquiries in-house, with no developer required for day-to-day operations, and all three surfaces stay in agreement because they consume the same underlying data. Specific figures for live listings, monthly visitors and enquiries generated are available on request.
