What Your Store's CCTV Hides: Retail Analytics in Singapore
The gap every retailer knows about but can't explain
500 people walked into your store last Saturday. 43 bought something. You know the 43. They're in your POS. You have no idea about the other 457.
Where did they go? Which aisle did they browse? How long did they stand at the fixture before walking away? Did a staff member approach them? Did they queue and leave?
This is the gap between what your POS tells you and what actually happened on your floor. For most Singapore retailers, it's invisible. It doesn't have to be.
What computer vision actually does with your CCTV
Modern computer vision models can analyse video feeds in real time and extract behavioural data without storing any footage or identifying individuals. What you get:
Footfall and conversion rate Count everyone who enters, not just everyone who buys. If 500 walked in and 43 converted, your conversion rate is 8.6%. Fashion retail conversion is commonly cited in the 15 to 25 percent range. That gap is measurable, trackable, and closeable.
Aisle heatmaps Which sections of your floor do customers actually visit? Retailers often find that a large share of their floor space, sometimes a third or more, gets almost no traffic. Knowing this before your next planogram review saves you from restocking products nobody walks past.
Dwell time How long does someone stand at a fixture before deciding? Long dwell with no purchase usually means pricing, packaging, or product confusion. Short dwell usually means the layout isn't pulling people in. Both are solvable, once you can see them.
Staff engagement When does your team approach customers versus stay at the counter? Retailers using engagement data consistently find that the stores with the highest approach rates have the highest conversion rates. This isn't a coincidence.
Queue and stockout alerts Real-time notifications when queues build beyond a threshold or when a shelf has been empty for more than 20 minutes. Your phone buzzes before the sale is lost.
The Singapore-specific context
Singapore's retail environment has specific pressures that make this data particularly valuable:
- High rental costs mean every square metre needs to earn its keep. Dead zones are expensive.
- Labour costs and quotas mean you cannot simply hire more staff to improve service levels. You have to get more from the team you have.
- PDPA requirements mean any system that captures customer data needs to be compliant. Fully anonymised computer vision (no faces stored, no individuals tracked) is PDPA-aligned by design.
No new cameras needed
The most common objection: "We'd need to upgrade our CCTV system."
In almost every retail deployment we've done, the existing CCTV infrastructure is sufficient. Modern computer vision models work with standard resolution IP cameras, the same ones most Singapore malls and standalone stores already have.
What matters is camera placement, not camera quality. We assess your existing setup during the pilot and flag any positions that need adjustment.
What the data actually changes
Knowing your footfall is interesting. Acting on it is where the value is.
The retailers who get the most out of analytics are the ones who review the data in their weekly operations meeting and make one concrete decision from it. Not 10 decisions: one.
- This aisle is a dead zone. Move the seasonal display there this week.
- Dwell time at the accessories fixture went up 40% since the restock. Let's make sure staff are briefed on it.
- Store B's conversion rate is 11 points below Store A. They have the same footfall. Pull both heatmaps side by side.
That's the cadence. Weekly data, one decision, repeat.
How we deploy it
The setup process for a typical Singapore store:
- CCTV audit (day 1): We review your existing camera positions and confirm coverage. Most stores need minor adjustments to 1 to 2 cameras, not a full replacement.
- Integration and calibration (week 1 to 2): We connect the feed, calibrate the counting zones and heatmap regions for your specific floor layout.
- Dashboard goes live (week 2 to 4): Heatmaps, conversion, dwell time, and staff engagement appear on a web dashboard accessible on any device. Daily summary emails start automatically.
- First review (week 4): We sit with you and walk through the first month's data. You'll almost always find at least one thing that surprises you.
Pilot programme
We run a two-week pilot for one store at no commitment. You see the data from your actual floor before deciding whether to roll it out.
If the pilot shows nothing actionable (which has not happened yet), you owe us nothing.