BRANCH NETWORK OPTIMISATION · COVERAGE DECISIONS
Shrink the branch network, keep the clients
Model every closure and merger before you commit: which branches carry their catchment, which duplicate a neighbour, and where your clients would actually go next.
WHY NETWORK REVIEWS HURT
The cost of closing the wrong branch
Cost pressure says fewer branches. Clients, regulators and the board say prove it. Every closure is a bet on where thousands of clients go next.
Closing on cost alone
The cheapest branch to run can be the only access point for a whole district. Cost rankings cannot see coverage, so they quietly propose the wrong closures.
Clients do not move where you assume
Closed-branch clients rarely walk to the nearest pin on a map. Travel patterns and habit decide who stays, who switches channel and who leaves.
Two branches, one catchment
Duplicates hide inside network averages. Two branches sharing a catchment both look mediocre, while together they carry one strong territory.
Every region defends its own
Regional managers argue with anecdotes, HQ answers with cost sheets, and the review stalls where it always stalls: no shared evidence.
HOW NETWORK REVIEWS RUN TODAY
A cost ranking in a spreadsheet
Anecdotes from regional managers
A consultant study every few years
Backlash discovered after the closure
Coverage argued branch by branch
WITH NETWORK IMPACT MODELLING
Every branch scored on coverage and demand
Closures simulated before they are committed
Client travel and absorption quantified
Mergers planned where catchments overlap
A wave plan the board can defend
MODEL · TEST · COMMIT
Every closure tested before it happens
One review workflow. The whole network scored on one basis, every closure and merger simulated for client impact, and decisions rolled out in waves. The next review starts from a living model, not a blank sheet.
SCORE · WHOLE NETWORK
The whole network on one basis
Score every branch on demand, coverage and overlap
Find duplicates sharing one catchment
Rank closure and merger candidates by network impact
REVIEW RANKING · CLOSURE CANDIDATES
Arcade
Parkside
Mill quarter
SIMULATE · EACH CLOSURE
Test the closure before it is real
Simulate closing: where do the clients actually go
Quantify travel added and clients at risk
Drop closures that would strand a catchment
Simulations run on Network Impact Modelling, the same what-if logic behind the Generali network reduction.
CLOSE B·07 · ARCADE
Clients absorbed nearby
96%
Average travel added
+4 MIN
Coverage lost
NONE
SAFE TO MERGE · INTO 2 BRANCHES
ROLL OUT · IN WAVES
Commit wave by wave
Wave one bundles the safest closures and mergers
Client movement is monitored against the simulation
The next wave starts from an updated model
NETWORK · WAVE 1
Arcade
MERGED
Parkside
CLOSED
Mill quarter
KEPT
14 CLOSURES · COVERAGE HELD · WAVE 2 IN MODEL
KEEP, MERGE OR CLOSE
Why one branch closes and its twin stays
No branch is judged alone. A branch that looks weak on cost can be the sole access point for a district, while a busy city branch may add nothing a neighbour would not absorb. The verdict depends on the network around it.
The model is calibrated on your own client base: real visits and transactions joined with trusted location data and travel times from five to thirty minutes by car or transit. Keep, merge and close all read from one basis, so the review is consistent across regions.
SCORE COMPOSITION
What moves a keep-or-close verdict, by branch type
Urban branch
Regional branch
CLIENTS
Active clients in the catchment
88
84
SOLE COVERAGE
Clients with no alternative nearby
90
38
ABSORPTION
How fully neighbours can take over
62
90
TRAFFIC
Visits and transactions at the counter
80
52
OVERLAP
Catchment shared with your own branches
66
72
COST TO SERVE
Running cost against branch output
70
55
Urban: overlap and absorption decide
Regional: sole coverage decides
SAMPLE WEIGHTS · CALIBRATED ON YOUR CLIENT BASE
THE PAYOFF
What a modelled review delivers
NETWORK COST
A smaller network that still covers everyone
Closures and mergers land where neighbours can absorb the clients, so the savings arrive without stranding a single district.
CLIENT BASE
Cost cuts without the churn spike
Client movement is simulated before every closure, so retention holds and the branches that remain inherit clients they can serve.
BOARD + REGIONS
Reviews that survive the boardroom
Regions and HQ argue over one shared model instead of each other, and every wave carries the evidence boards and regulators ask for.
BEYOND THE COST SHEET
A cost spreadsheet cannot see client travel
COST RANKINGS
Good for budgets, blind to coverage
A ranking by running cost proposes closures wherever rent is high, and never notices the district losing its only branch.
BI DASHBOARDS
Good for branch KPIs, silent on what-if
Dashboards report how each branch performed. They cannot simulate a closure, a merger or where the clients would go.
CONSULTANT STUDY
Good for one wave, gone by the next
A network study answers this year’s review and expires. The network keeps changing, and the next wave starts from scratch.
GUT AND SENIORITY
Good for speed, weak for defence
Experience matters, but a closure defended only by seniority collapses at the first board question about client impact.
CLEVERMAPS
One model from scoring to wave plan
Score the network, simulate every closure and plan the waves on one shared basis, with client impact quantified before anyone commits. The core workflow of the Performance solution.
Branch network optimisation is the core workflow of the Performance solution, built on trusted location data and shared decision logic.
PROVEN AT SCALE
Proven on a national branch network
FEATURED CASE STUDY
Reducing branch network by 35% while maintaining client access and profitability
Generali Česká pojišťovna modelled its whole branch network, cut it from 800 to 520 sites and kept client access and profitability. Every closure was simulated before it happened.
INSURANCE · NATIONAL BRANCH NETWORK
ALSO IN BANKING
Data-driven business decisions in banking, with Česká spořitelna
Read the story →
30-MINUTE WORKING SESSION
Put your network on the table
Bring your branch list. In 30 minutes we will score the network live, find the overlaps and simulate the first closures, on your own data.