Predictive business modeling

This is the flagship work of the desk. It is for owners, finance leads, and operators in Batumi and the rest of Adjara who must decide before a season turns — hotels that live on occupancy, traders who buy containers, shops whose till changes with the boulevard, houses that borrow against a summer they have not yet seen.

The result you leave with

You receive a documented forecast model for a named horizon (usually ninety days, sometimes a full season), three scenarios that a manager can argue with, and a board pack: charts, a variance table, and a two-page letter in plain English. The pack is printed and also sent as a PDF. The working tables remain yours.

It is not a live feed, not a login, and not a promise that the rain will miss July.

Who sits with you

Nino Beridze leads the build. Tamar Kakabadze draws the charts and sets the type for the pack. Giorgi Tskitishvili joins when the books involve stock, port timing, or a second site. You meet the same people at the scoping visit and at the reading.

What is in the room, and what is not

Included: the scoping visit (up to ninety minutes at Melikishvili Street or at your premises in Adjara); cleaning and naming of the series you supply; the model itself; three scenarios agreed in week two; two working reviews; the printed pack (twelve copies) and the PDF; a two-hour reading of the pack with your named directors.

Excluded: bookkeeping, tax filings, software licences, hiring decisions, and any chart of a series you did not give us. We will not invent occupancy you did not record. We will not put a competitor’s unpublished rates into the model.

How the weeks run

Week one is the books: occupancy sheets, till tapes, warehouse counts, loan schedules — whatever you actually keep. We say, in writing, which gaps make a forecast dishonest. If the gaps are too wide, we stop and bill only the scoping visit.

Week two names the drivers: season, cancellations, border traffic, a kitchen that cannot flex, a container that lands late. You choose the three scenarios. We do not choose a “best estimate” and hide the others.

Weeks three to six are the build and the drawings. You see a draft pack once. You mark it. The last week is typesetting, the reading, and the moment the copies leave the desk.

Duration, place, preparation

Four weeks is honest when the series are already monthly and a single site. Eight weeks is honest when there are several outlets, a mix of lari and foreign receipts, or a season that has to be split by weekday. Work sits in Batumi; we travel within Adjara without a separate travel theatre if the site is a working day from the office. Beyond Adjara we say so in the estimate.

Bring to the first visit: two years of the series if you have them, last season’s plan even if it failed, and one person who knows why a number was written down. Do not tidy the spreadsheet until we have seen the mess. The mess is usually the model.

Constraints we will not talk around

A model that is used to raise money from strangers will be labeled as such, and we may decline. A model that must “show growth” is not a model we will sign. If your series begin last April and you want next April, we will say the winter is a guess.

Price and the next step

You do not buy this off a card. After the scoping visit we send a one-page estimate: fee, what would raise it (another site, another currency, a pack in Georgian and English), and the deposit (one third, due before week two). Typical finished work sits between 4,800 and 18,000 GEL. The first forty-five minutes of a fit conversation, by telephone or in the rooms, is not billed.

File a brief on the contact page. Name the decision and the month it must be taken. We reply within two working days.