Which markets produced buyers who actually completed?

Selling apartments in one city to buyers in Singapore, Dubai, Hong Kong and London means running parallel campaigns in markets with different costs, different languages and completely different conversion behaviour.

Reported together, they produce a blended cost per enquiry that describes none of them. One market may generate enquiries cheaply and buyers rarely; another may cost five times as much per enquiry and carry the whole project.

Matching completed purchases back to the enquiry, with the buyer's territory on the row, ranks markets on completions. That is the number that decides where next quarter's exhibition budget and agent commissions go.

Overseas agent networks typically take a substantial commission on each unit sold, and they deliver buyers the developer could not reach directly. It is a known, large cost per sale.

Direct marketing into the same territories is cheaper per enquiry and far less certain per sale. Only a report that ranks both on completed revenue can settle which is better in which market, and the answer is usually different market by market.

Exhibitions remain central to selling property across borders, and they are expensive in a way that makes finance directors uncomfortable every year. A stand, a shipped model of the development, flights and hotels for four staff, translation, a local agent on the door, advertising in the host market to fill the room for two days. What comes out of it is a stack of registration forms and a folder of business cards, and what the firm records afterwards is how many people came and how many reservations were taken on the day.

Reservations taken at a show are close to the worst available measure of one, because the environment is built to manufacture urgency and a deposit signed beside a scale model on a Sunday afternoon is not a considered decision. What actually matters is the visitor who took a brochure, went home, thought about it for eighteen months, watched the construction updates arrive by email, discussed it with a spouse and an accountant, and completed on a unit long after the exhibition team had written that particular event off as disappointing. That person exists on the registration list and nowhere else in the business.

Typed up with names and contact details, that list is a source file like any campaign. Completions are matched back to it on the buyer's own phone number or email, however long afterwards, so each show is ranked on units actually completed rather than on footfall or on the deposits taken before anybody went home and reconsidered. Several years of exhibition lists uploaded together will usually settle an argument the firm has been having internally for as long as it has been exhibiting.

A buyer in one country enquiring about property in another arrives with messy identifiers. The mobile number is written with a country code sometimes and without it at other times, with a leading zero that is correct locally and wrong internationally. Messaging applications supply a different number from the one on the reservation form. An email address given at an exhibition is personal, while the one on the purchase contract belongs to a lawyer or a family member who handles the paperwork.

This is the ordinary failure mode of every attempt to reconcile these files in a spreadsheet, and it is the reason firms conclude the exercise cannot be done at all. A lookup on raw text finds a fraction of the real matches and quietly reports everything else as unattributed, so the answer is not simply incomplete. It is biased toward whichever market happens to write its telephone numbers most consistently, which means the territories that look weakest may be the ones the join failed on rather than the ones that sold least.

Normalising phone numbers to a canonical international form before any comparison happens is the first step of the match rather than an afterthought bolted on when results look poor, and it recovers the large majority of what a naive join loses. Where the remaining link is weaker than the threshold for automatic counting, it is flagged for somebody to accept or reject rather than counted quietly and reported as fact. At this transaction volume that is a short review by one person, and it is a far better answer than an assumption nobody can audit afterwards.

Cross-border purchases are family decisions and corporate ones. A parent buys an apartment for a child who will study in the city and whose name goes on the contract. A couple enquires jointly and completes in one name for tax reasons. A company buys the unit and the individual who attended every viewing appears nowhere on the deed. A relative already living in the destination country handles the entire transaction under a power of attorney because the buyer cannot travel.

Each of those situations breaks a name-based reconciliation completely, and they are not rare cases sitting at the margin of the book. In several markets they are the majority of transactions rather than the exception, which means a firm attempting this manually in a spreadsheet will conclude that most of its completions came from nowhere at all, and that its exhibition programme and its agent network between them produced almost nothing. That conclusion is wrong, and it is wrong in the direction that gets budgets cut.

Matching on contact details rather than on names handles the common version of this, because the mobile number given at the enquiry is usually still the number the developer rings to arrange the handover inspection two years later. People change vehicles, names on deeds and even countries of residence; they change their mobile number far less often. Where it genuinely is a different person on both sides, the honest answer is that the connection is not present in the data, and that revenue sits in the unattributed bucket rather than being assigned to whichever campaign happened to be running that month.