October 19, 2026 · Tupll
How Portfolio Optimization Helps Decide Where to Expand, Relocate, or Close
Most location analysis obsesses over the next site. Portfolio optimization asks a bigger question: is the network as a whole deployed where it should be? Same capital, same brand, same operations, but are the pieces on the right squares?
For any multi-location operator, the honest answer is usually no. Networks accumulate history rather than strategy: sites chosen under old market conditions, footprints inherited through acquisitions, leases renewed by default. Optimization is the discipline of looking at the whole board at once.
Three decisions, one framework
Expand, relocate, close. Treated separately, each decision gets its own politics and its own selective evidence. Treated together, on one model, they become comparable: every existing and potential location scored the same way, on the market it serves and the performance that market supports.
That single basis is what makes the hard calls legible. A store can be profitable and still be misplaced, earning less than the same investment would in a market two hours away. A struggling location can deserve patience because its market scores well and the problem is operational. And a beloved legacy site can be revealed as occupying a market that moved away from it a decade ago.
Reading the portfolio honestly
Scoring an entire network against its markets produces a matrix worth more than any single forecast: performance versus market quality. Strong store in a strong market: protect it. Weak store in a strong market: fix operations, not real estate. Strong store in a weakening market: enjoy it, but plan; do not double down. Weak store in a weak market: the exit candidate, and the model gives you the evidence to retire it without relitigating history.
The relocation case is the most overlooked. When a market has drifted, the demand pocket often moves a few miles rather than vanishing. The same customer base, served from a better point on the map, can transform a marginal P&L without touching the brand or team.
Overlaps and gaps
A portfolio view also exposes what site-by-site analysis structurally cannot: places where your own stores compete with each other, and markets your footprint should own but does not. Cannibalization within a network is a silent tax paid monthly; unserved high-scoring markets are the reverse, revenue forfeited to whoever arrives first. Both only appear when the network is modeled as a network.
Cadence, not crisis
The failure mode of portfolio work is doing it once, during a crisis, under time pressure, with closures predetermined. The better pattern is boring: re-score the network on a regular cadence, because markets move even when stores do not. Each lease renewal then becomes a genuine decision informed by current evidence rather than a rollover.
Capital goes where the analysis says the market is. Over a decade, the operator who reallocates deliberately ends up with a network that looks designed. The one who does not ends up with a network that looks remembered.