Your next branch, where your current branches point.
We learn the shared profile of your branches, rank the areas that fit it best and test the result in every report. We run the analysis; you receive the report.
New branch decisions are often made with a property tour, a hunch and a spreadsheet.
A branch in the wrong place carries rent, fit-out and staff costs for years. The best source on where your business works is the branches you already run.
Location decisions should be testable. We built Zonindex for that.
Learns from your branches, searches the rest of the city.
Every analysis goes through the same four steps. Your branches decide which features matter and how much.
We map your branches
Your branch list, with revenue if available, is placed on the 0.7 km² hexagonal cells that cover Istanbul.
We learn the shared profile
We describe every cell from six angles: business mix, population and social structure, transit access, people movement, physical texture from satellite imagery, and the cell itself. The model learns the pattern that sets your branch locations apart from the rest of the city.
We measure the market
For each candidate we draw a catchment on the real road network: 10 minutes by car, or 5, 10 and 20 minutes on foot. We count residents, daytime density and competitors inside it.
We build the shortlist
We apply your must-have conditions: distance to competitors, overlap with existing branches, minimum population. Candidates are sorted into Prime, Strong and Moderate tiers.
Every cell is measured from six angles.
Each map ranks Istanbul's 1,666 urban cells from one angle; a darker shade means a higher value. The model learns from your branches which angles matter for you.
From 1,667 cells to 116 candidates.
The shortlist narrows in four steps. This example was prepared for demonstration: well-known neighbourhood centres were treated as branches and are marked in black.
If the model can't find your branches, we tell you.
In every analysis we hide your branches from the model one at a time and ask it to find them again in the city. The result appears in the first pages of your report.
Could we have found the three newest clubs in advance?
We hid the chain’s three most recent clubs from the model and reran the analysis as if they did not exist. All three came out in the top tier.
Prime tier, first 117 candidatesThree newest clubsEach square is a candidate cell: the 1,222 cells before conditions, ranked by score left to right, top to bottom. The 116 candidate areas on the opening map are the shortlist left after conditions.
Data from the chain’s public club list. The chain’s name, number of clubs and locations are not shared.
A report you can take to your board as it is.
For every candidate we write the reasoning: where it resembles your branches, where it falls behind, how many people live in its catchment and how many competitors are there.
Candidate 01
Candidate file
One page per candidate: a comparison with your branches and the three branches most similar to it.
New branch candidates
Executive summary
Four key findings, a map and the top ten candidates.
10 minutes by car
Catchment
How many people, businesses and competitors the candidate is close to.
Values on the report pages are illustrative.
We build the score around your business.
Off-the-shelf location scores apply the same yardstick to every business. Zonindex builds a separate model for each company and shows whether that model works.
10 minutes by car from Beşiktaş
The bold line is the area reached in 10 minutes on the real road network; 222,000 people live inside it. The dashed circle has the same area, but 31% of it lies in the sea and 141,000 people live inside it.
- 10 minutes by road
- Circle of the same area
- Cell shade: residents
Competitors in one neighbourhood
Businesses with “döner” in their name in a 4 × 3 km area of Şişli–Mecidiyeköy. A single source finds 10; when we combine sources, 49. Across the city, 502 records grow to 1,707, 3.4 times as many.
- Single source
- Other sources
- Building footprints
You can start with three branches.
If there are not enough branches to build a model, we rank candidates directly by similarity to your branches and present the result as a relative ranking. To strengthen the analysis we can also add branches of other chains with the same positioning as examples.
Black: your three branches. Sea green: branches of chains with the same positioning. Grey: the rest of the city.
We measure the fit
We test whether the added chain really does well in the same kinds of places as you. If it does not fit, we leave it out and explain why in the report.
We count them as competitors too
Branches of the chains we use as examples also count as competitors in the candidates’ catchments.
It gets stronger as you grow
As your branch count grows, the analysis moves to a model that learns from your branches. It can be rerun after every new opening.
A branch list, a report.
Your branch list
You send your branch list with addresses or coordinates. Sharing revenue is optional. We place the addresses on the map and check them.
Report delivery
We deliver the report, the Excel file and the suggested points to you.
Expansion analysis
For a first project, for chains of any size.
- Tiered candidate shortlist
- Validation test
- Report, Excel and suggested points
Performance analysis
For chains that can share branch revenue.
- Everything in the expansion analysis
- A model built on your best branches
- A test of whether the profile explains revenue
- Two scenarios: catchment and conditions
Expansion advisory
For chains that open several branches a year.
- Quarterly update
- Reviews of new openings after the fact
- Fast reviews of franchise and rental site applications
We are running pilot projects with our first partners. We discuss the terms in a call.
Do you only work in Istanbul?
For now, yes. We built our data layer for Istanbul at building and street level. Write to us about another city; we prioritise new cities by demand.
Is revenue data required?
No. Branch locations are enough. If you share revenue, the model learns from your best branches and we measure separately what that adds to the result.
How many branches do we need?
You can start with three. As your branch count grows, the model learns on its own the pattern that sets your branches apart from the rest of the city.
Do we still need site visits?
Yes. Zonindex gives you a ranked shortlist of areas to visit, with the reasoning. Rent, property and frontage decisions are made on site.
Set aside 30 minutes for your next branch.
Fill in the form or write directly to info@zonindex.com; we will get back to you within two business days.
Write directlyinfo@zonindex.com