Oikion

Oikion replaced the patchwork of spreadsheets and WhatsApp groups that small Greek agencies, five to fifteen agents, had stitched together just to get through the week. What shipped is three modules on one screen — an MLS listing inventory, a CRM for contacts and deal pipeline, and Oikosync, a real-time team feed — over four systems the agency never has to think about: an import engine, the Polis inter-agency network, a self-hosted signing flow, and end-to-end encryption. All of it designed in Greek first, because Greek runs about a third longer and every compact element has to survive the harder language.

The one place I will put a number is onboarding. Replacing the incumbent’s eighteen-field signup wall with a three-field account step and a progressive flow models to roughly a 28% lift in completion. That figure is modeled against form-usability benchmarks, not measured on Oikion: the direction is well established, the magnitude is my estimate.

Greek agencies were not short of software. The two serious incumbents have each run for about eighteen years, sit inside 850 and 724 agencies, and syndicate to two dozen-plus property portals. I was never going to out-feature them in a year, so the bet was the model.

The tell is a single word: mandate. The incumbents treat it as a listing the agency holds, modelling the business supply-first. Greek agents mean the opposite: a brief a buyer hands you, a demand signal. So Oikion makes the Request a first-class, demand-side entity from the data model up. Same word, opposite centre of gravity.

Eighteen fields became three

Every legacy tool I audited opened with the same wall: one page, every field a Greek agency could ever need, demanded before the product had earned a single keystroke of trust. Oikion’s three-field account step moves the rest into a seven-step flow, and the progress bar deliberately stays hidden until step three — because choosing a language and saying hello are explicitly not work.

The full teardown, side by side with the competitor wall →

Sharing got two dimensions instead of one

Polis, the inter-agency network, turned out to be less a matching feature than a trust problem. Fiercely private, competing offices will only share if they stay in control, so one control decides whether a record is shared at all and a second decides how much of it others see: bare specifications, plus the office name, or plus the listing agent. It resolves when a record is read, so a change of heart takes effect the instant an agent makes it. Matching was the reward. Control was the product.

How Polis resolves privacy at read time →

All of it behind one PIN

Agencies hold financial positions, tax numbers, and private negotiations, and I wanted a real guarantee instead of a comforting phrase: that Oikion itself cannot read the most sensitive collaborative data. True end-to-end encryption is famously cruel to users, so the whole system hides behind a single four-to-eight digit PIN, set once. And the one truth that cannot be designed away, that losing both the PIN and every recovery code means the data is genuinely gone, the interface says plainly, at the moment it matters.

The key hierarchy, and the edge case that could lock users out forever →

None of the four systems underneath ever announces itself. That was the point: an agent sees three modules and a feed that stays current on its own, never the machinery keeping it that way.

Five working agents trialled it. Small and informal, with no formal scoring: the read was consistent that Oikion was quick to learn and felt more polished than the incumbent they came from. I am explicitly not presenting that as a metric.

Set against those incumbents the honest picture is mixed, and saying so is the whole point. They are ahead where eighteen years buys reach and features; Oikion is ahead where a coherent model and design discipline buy fit.