Modelling the “Should”: How LoculChoices Finds Hidden Retail Upside

Matt Copus
2 days ago
4 min read

Most analytics stops at describing what happened.

How many visits did we get?
Where did they come from?
Which centres grew, which fell?

Useful, but incomplete. It tells you what did happen, not what should have happened.

If you don’t know what a centre should be doing given its catchment, competition and format, you can’t answer basic questions:

Is this asset genuinely strong, or just average for its position?
Is that suburb “difficult”, or are we just under-performing it?
Is this new development idea realistic, or wishful thinking?

That’s why we built LoculChoices – our Spatial Interaction Model: a modelled view of how customers should behave, location by location.

LoculChoices in plain language

LoculChoices estimates the probability that a customer chooses a particular destination, given all the choices available to them.

For each origin (a suburb, mesh block, trade-area segment), the model looks at:

How attractive each centre is (scale, mix, anchors, category strength, amenity).
How hard it is to get there (distance, travel time, friction).
What the alternatives are (competing centres and retail strips).

Then, using machine learning, we calibrate LoculChoices against real behaviour and sales outcomes so it learns how customers actually trade off convenience and attraction in a given market.

We don’t badge it with academic labels. We simply call it what it is:

LoculChoices – our Spatial Interaction Model, created by AI and curated by people who actually understand retail.

The output is a clean probability surface: for every origin and every centre, “this is how customers should distribute their spend and visits if the market is functioning normally.”

It’s hypothetical by design—and that’s the point. It defines the opportunity frontier.

“Should” vs “Does”: where Loculyze gets interesting

On its own, LoculChoices is a very smart map.

The magic happens when we combine it with loculWays, our measurement view of how customers are actually behaving from mobile signal data and other feeds.

LoculChoices (Spatial Interaction Model) → how customers should behave.
loculWays → how they do behave.
The gap between the two → where value is hiding.

That gap tells you:

Which suburbs are under-penetrated for a centre.
Which centres are under-performing relative to their potential.
Where competitors are punching above their weight.

In other words, it turns “interesting model” into a prescriptive asset-growth engine.

Use case 1: Targeted growth for shopping centres

For landlords and centre managers, LoculChoices + loculWays lets you:

Find underperforming suburbs.Places where LoculChoices says you should have a strong share, but loculWays shows weak visitation and sales contribution.
Prioritise capital and effort.Direct leasing, marketing and capex into the suburbs and customer segments with the biggest gap between “should” and “does”.
Measure uplift properly.After interventions, track whether visitation and sales are closing the gap towards the modelled potential.

Instead of broad, expensive campaigns, you get a shortlist of where growth is mathematically most likely to pay off.

Use case 2: Value-add investors hunting for upside

If you’re a value-add investor, you don’t just want great assets—you want fixable assets.

LoculChoices and loculWays help you:

Scan markets for centres where modelled potential is high but realised performance is weak.
Separate “structurally challenged” assets from “operationally undercooked” ones.
Build an investment thesis around specific suburbs, segments and competitive battlegrounds.

You’re no longer guessing where the upside might be; you’re buying assets where the gap between “should” and “does” is already quantified.

Use case 3: Greenfield and scenario planning

For developments and network planning, historical data is often thin or irrelevant. That’s where a Spatial Interaction Model shines.

LoculChoices lets you:

Test greenfield proposals:Drop in a hypothetical centre, adjust size, mix and anchors, and estimate how customers in surrounding suburbs are likely to respond.
Explore multiple scenarios:Different formats, competitor openings/closures, or repositioning of nearby centres.
Understand strategic trade-offs:How far will customers really travel? What happens to neighbouring assets? Where does cannibalisation bite?

Because LoculChoices is model-based, you can explore a wide range of options before you invest in detailed design and mobile-data deep dives.

Light-touch national coverage

One of the biggest advantages of LoculChoices is that it doesn’t require heavy local data to get started.

Because we’re modelling structure—travel cost, competition, attraction—not just counting devices, we can:

Model any shopping centre in the country with a light data touch.
Run broad, cost-effective scans across markets for investors and retailers.
Flag high-potential / under-performing assets for a second phase of detailed mobile-signal analysis in loculWays.

Think of LoculChoices as a national screening tool: fast, scalable, and good enough to decide where it’s worth rolling in the heavier artillery

Predictive and prescriptive by design

We talk a lot about being a predictive modelling company, but prediction is only half the story.

Every Loculyze product, including LoculChoices, is built for prescriptive analytics:

We don’t just say “here’s a pattern”.
We say “here’s what this asset should be doing, here’s what it is doing, and here’s where to act.”

For landlords, that’s a ranked list of suburbs and segments to grow.For investors, it’s a shortlist of assets worth deeper due diligence.For retailers, it’s a map of where brand reach is structurally strong versus structurally under-realised.

Beyond mobile: a full growth engine

Mobile data is a powerful lens on customer behaviour, but it’s only one part of the story.

LoculChoices is the other half of the engine:

loculWays → performance (what’s happening).
LoculChoices (Spatial Interaction Model) → opportunity (what’s possible).
The gap between them → strategy (where and how to grow).

That’s the growth loop we’re building for retail property:

Model reality. Measure reality. Compare the two. Then act.

If you want to see how your centres – or your targets – stack up on the “should vs does” curve, we can start with a light-touch LoculChoices run, then decide where a deeper mobile analysis in loculWays will add the most value.