Retailers lose millions of Rands every year because they approve a site that looks good on its own but makes little sense inside their wider retail network. The location has traffic, access, visibility, and a plausible target market, so the deal moves ahead. As the site matures, it cannibalises sales from existing stores because its trade area overlaps with those stores. This is often because it is situated in a location or shopping centre that looked attractive on paper and was marketed as such by property owners, but never had the likelihood of achieving its full retail potential.
That mistake usually starts with a key property development error. A site feasibility study is conducted to validate the location, rather than a proper retail network optimisation study, which is related but doesn’t have the same return-on-investment impact. Our experience shows that retailers need to move beyond intuition, simple population counts, or one-site feasibility studies toward delineating trade areas, integrating consumer demand and competitor mapping, to network-wide decision-making.
Retailers that treat these two exercises as interchangeable often overstate demand, understate cannibalisation, and misallocate capital when conducting only a feasibility study. International best practice shows that simple location analytics can identify the locations, estimate potential sales, and recommend store format or size, while modern retail-network analysis tests how one location affects the rest of the retail network across markets and channels, ensuring a stable and financially viable network.
Key takeaways
- A site feasibility study asks, “Can this site work?”
- Retail network optimisation asks, “Does this site improve the retail network, in this market, in this store format?”
- A store can pass feasibility KPIs and still destroy brand value through cannibalisation, duplication, or poor market fit.
- The right decision comes from analysing trade areas, competitors, access, consumer demand, and the whole store footprint together.
- Retailers using retail network optimisation make better expansion decisions, and only when using this approach can store relocation and rationalization be done effectively.
What a site feasibility study is designed to do
A site feasibility study usually evaluates a single location. It looks at access, visibility, rent, zoning, traffic generators, nearby demographics, parking, and whether the surrounding catchment areas are large enough to support a store.
That research matters. No retailer should sign a lease or approve a capital spend without confirming that a site is physically workable, commercially plausible, and consistent with the brand’s operating requirements.
But feasibility studies still only use a one-site lens. Even when it uses good data, it tends to answer a narrower question – does this specific site have enough local promise to justify further investment?
What network optimisation is designed to do
Network optimisation works at a different level. It evaluates the whole store footprint, the spaces between stores, the markets not yet served, the markets overserved, and the likely effect of adding, relocating, resizing, or closing outlets.
That is why international best practice treats retail footprints as interacting systems rather than a collection of isolated pins on a map. McKinsey notes that markets are becoming more complex with online purchasing and delivery systems – retailers now need to evaluate how online behaviour affects how brick-and-mortar locations and vice versa, and that network decisions must reflect market-level differences rather than blanket assumptions about individual sites.
GeoScope has made the same case locally in Crafting a Winning Retail Network Strategy: Key Steps for Success, where trade areas, target-market definition, competitor mapping, and greenfield-versus-brownfield thinking are treated as core inputs to retail decision-making rather than optional extras.
Why retailers keep confusing the two
The confusion persists because both exercises use similar words. Both talk about catchments, demand, competition, and customer profiles, so decision-makers assume they are just different labels for the same thing.
They are not. A feasibility study may identify a promising corner, but network optimisation asks whether that corner adds net value after accounting for overlapping trade areas, substitution, market saturation, store format, and the existing role of nearby competitors’ outlets. Clustering of retail outlets near competitors or attraction facilities can have a positive impact on revenue potential, but if done wrong, significant revenue losses can occur.
That difference becomes critical in mature or highly competitive markets. KPMG’s retail location-planning work defines location analytics as a discipline that not only identifies good locations and estimates potential sales, but also suggests the right retail store format and size based on many combined data sources. That moves the discussion well beyond a simple “good site” logic.
The questions each discipline should answer
A sound feasibility study should answer questions like these: Is the site accessible? Is it visible? Does the surrounding area contain the right target market? Can the store physically operate here and generate revenue? Will the lease economics and likely turnover justify the investment?
A sound retail network optimisation exercise asks harder questions: Will the new location grow the market or simply redistribute sales? Which existing stores gain or lose? Is this a whitespace opportunity, a defensive move, a relocation, a format correction, or is consolidation needed?
The distinction matters because a retailer can easily approve the wrong store for the right reasons.
The site itself may be viable, yet network decision may still be poor because another location would have more incremental sales, better coverage, or less internal competition.
Cannibalisation is the cost most retailers underestimate
Cannibalisation is where the confusion becomes expensive. If a new outlet captures customers who would have bought from an existing brand outlet, the business may post new-store opening losses and destroy the brand’s net value.
This is not just theory. Esri’s Business Analyst treats cannibalisation as a measurable trade-area-overlap problem, with tools that identify overlap between polygons, calculate overlap metrics, and add demographic or lifestyle variables to show what potential revenue losses are at stake inside the shared area.
That is why network optimisation must test interaction effects, not just local demand. As an example, GeoScope in its blog on Brooklyn Mall’s Struggles argues that its rising vacancy rate from 3.6% in 2019 to 18.7% in 2024 is a consequence of Castle Gate Shopping Centre cannibalising its market. This should have been visible in advance by conducting retail network optimisation and capacity-defined trade area analysis.
A feasible site can still be the wrong market move
Retailers often assume that if a site sits in a busy node, the decision must be sound. That logic fails because activity alone does not tell you whether the node adds incremental demand to your network.
One of the clearest lessons from the GeoScope research is that measured demand matters more than surface optics. In Where South African’s Shop – Mapping Malls and Their Trade Areas for Smarter Retail Decisions, we argue that mall locations become truly useful only when they are linked to trade-area demographics, mall classifications, key anchor tenants, and quantified consumer demand.
The opposite also appears in our blog on The Cost of Poor Planning: The Rise and Fall of Shopping Malls in South Africa, where we argue that Tshwane’s China Shopping Mall was accessible but constrained by weak demand and powerful nearby competition, which is exactly the kind of failure that a narrow feasibility lens can cause.
Good optimisation starts with better data, not more opinion
Retail network decisions improve when retailers combine internal performance data with external market intelligence. That means store sales, format, and profitability data on one side, and demographics, mobility, trade areas, outlet censuses, mall data, category spend, and competitor locations on the other.
In our recent retail blog series, this point is made repeatedly. We argue that expansion in South Africa requires more than simple population statistics – it needs granular consumer insight, outlet mapping, performance benchmarking, and continuous adaptation as markets shift. One of the only sources of consumer data is from the Marketing All Products Survey (MAPS) of the Marketing Research Foundation (MRF), which GeoScope is a partner of.
McKinsey reaches a similar conclusion from a global perspective. Its geospatial retail work argues that retailers should assemble wide-ranging internal and external data, model actual and potential sales by local market, and simulate the impact of adding or removing stores rather than relying on broad network-wide assumptions.
The right process is sequential, not simplistic
Retailers should not abandon site feasibility studies. They should place them in the right sequence.
First, run network optimisation to identify true retail whitespace, overtrading nodes, consolidation risk, and the best role for each market. Then screen candidate sites inside those priority zones. Finally, run site feasibility studies on the most highly ranked potential sites to determine which one best delivers the brand’s network strategy.
That order changes the quality of decision-making. It prevents teams from falling for a single site offered to them by property agents of mall developments before they have tested whether the market actually needs another store, a different format, a relocation, or no action at all.
What this means for South African retailers
This distinction matters even more in South Africa because the market is uneven, diverse, and often poorly understood at the local level. Formal malls, township retail, transport corridors, informal trade, and shifting income patterns create demand landscapes that punish generic assumptions.
GeoScope’s research consistently argues for measured trade areas, local consumer intelligence from MAPS, and detailed mapping of outlets and facilities across both formal and informal environments. That is visible in posts on retail censuses, mall trade areas, fast-food competition, grocery positioning, and broader retail expansion strategy.
Retailers that keep using a feasibility mindset for network decisions will keep asking the wrong questions. They will ask whether a site is attractive when they should ask whether the network becomes stronger if that site opens there, in that format, at that time.
The practical test every retailer should use
Before approving a location, executives should require the project team to answer five questions:
- What unmet demand does this site capture?
- Which stores does it affect?
- How much sales transfer should we expect?
- What alternative location or format creates better net value?
- What happens if we do nothing?
If the team cannot answer those questions with evidence, it is not doing network optimisation. It is doing a dressed-up feasibility study.
That distinction may sound technical, but it goes straight to retail development discipline. Retail growth does not come from opening stores that look good in isolation. It comes from building a network that covers the right markets, reaches the right customers, and avoids cannibalising itself.
A site feasibility study is a necessary gate. Network optimisation is a strategic choice.
Retailers need both, but they must use them in the right order and for the right purpose. The winners will be the brands that stop confusing “this site can work” with “this is the best network decision.”
FAQs
1) Can a retailer skip the feasibility study if it has already done network optimisation?
No. Network optimisation tells you where the next high-ranking opportunity exists that has the largest market share in the shortest distance in the portfolio, but it does not replace on-the-ground evaluation of a specific site. You still need to test access, visibility, lease economics, store fit, site constraints, and local trading conditions before you commit capital, because optimisation narrows the search and feasibility validates the final choice.
2) What is the biggest warning sign that a business is doing site feasibility instead of retail network optimisation?
The clearest warning sign is that the whole decision revolves around one proposed site. If the executive site motivation documents show traffic counts, surrounding demographics, rental terms, and attractive visuals but say little about overlap with existing outlets, transfer of sales, whitespace alternatives, or the effect on nearby stores and channels, then the business is still thinking like a feasibility team rather than a network team.
3) Is network optimisation only relevant for large national chains?
Not at all. The discipline becomes more valuable as soon as a business has multiple outlets, different formats, or expansion choices. A regional chain with ten stores can still overserve one corridor, ignore another market, or open a store that weakens an existing branch, and those mistakes can hurt a smaller business even more because each store carries more weight in the overall portfolio.
4) What data should a retailer prioritise first?
Start with the data that links demand to performance. That usually means store sales, format, and profitability data; customer or loyalty data where available; competitor and same-brand outlet locations; trade areas and travel times; local demographics and income; and major attractors such as malls, transport nodes, and commercial areas. After that, add richer layers such as category spend, footfall, mobile movement, and informal retail mapping, where they materially improve the decision.
