A complete business CCTV system is designed around the evidence and response the organisation needs—not the number of cameras or the largest megapixel figure in a catalogue. Start by defining each scene, target and operational purpose. Then select the view, lens, lighting, recording, network, power, monitoring, privacy and maintenance controls needed to achieve that outcome.
The result should answer a practical question at every location: what must an authorised operator be able to see or prove here, under which conditions, and what action should follow?
1. Define the security outcomes
Each camera should have a documented purpose. Typical outcomes include:
- Detecting movement at a perimeter or restricted area;
- Observing behaviour in a yard, aisle or reception;
- Recognising a known person or vehicle;
- Obtaining identification-quality facial or number-plate evidence at a controlled point;
- Verifying an alarm before dispatch;
- Supporting safety, process or incident investigations;
- Providing live situational awareness to guards and responders.
One wide-angle camera may show that “something happened” but fail to identify who was involved. A design should specify the required level of detail for the target at the intended distance. Industry designers often use the detect, observe, recognise and identify—or DORI—concept to match image detail to purpose.
2. Complete a site survey
Walk the site during the conditions that matter: daylight, darkness, shift change, deliveries and periods of low activity. Mark critical assets, routes, entrances, exits, public areas, restricted zones and likely approach paths.
For every proposed view, record:
- Target and required outcome;
- Distance, width and height of the scene;
- Day and night lighting conditions;
- Backlight, glare, shadows, dust, rain, vibration or steam;
- Mounting position and safe maintenance access;
- Possible obstruction by vehicles, stock, trees or doors;
- Network and power availability;
- Privacy-sensitive areas that must be excluded or masked.
Link the survey to the site security risk assessment. This stops attractive but low-value views from consuming the budget while critical gaps remain.
3. Choose the right camera for each scene
| Camera type | Best suited to | Design caution |
|---|---|---|
| Fixed dome or turret | Entrances, offices, retail and general indoor or sheltered views | Confirm low-light performance, vandal resistance and reflection control |
| Fixed bullet | Perimeters, yards and visible deterrence | Protect from tampering and avoid an unnecessarily narrow or wide lens |
| Varifocal camera | Scenes where the exact field of view must be adjusted during commissioning | Recheck focus and target detail after adjustment |
| PTZ | Live tracking over large areas by trained operators | A PTZ looks in one direction at a time; use fixed overview coverage where evidence cannot be missed |
| Thermal | Detection in darkness, haze or difficult perimeter conditions | Thermal detection may need a visible camera for recognition or identification |
| Dedicated number-plate camera | Controlled vehicle lanes with defined speed, angle and lighting | A general overview camera rarely produces reliable plate evidence in every condition |
| Panoramic or multi-sensor | Wide situational awareness with fewer mounting points | Check target detail at the edge and the effect on storage and bandwidth |
Resolution matters, but it is only one variable. Lens choice, target size, sensor performance, compression, shutter settings, lighting, mounting angle and movement all affect usable detail.
4. Design the field of view
Start with the target and work backwards. If the requirement is to identify a face at a door, make the door the controlled capture point. Mount the camera near face height while protecting it from tampering, avoid a steep overhead angle and manage backlighting from glass or the exterior.
For vehicles, separate context and identification where needed. An overview camera can show vehicle colour, direction and events, while a dedicated lane view captures the plate under controlled geometry and lighting.
Avoid these common placement errors:
- Mounting every camera as high as possible;
- Using one ultra-wide view for both context and identification;
- Pointing a camera directly into sunrise, sunset or headlights without testing;
- Ignoring seasonal trees, stacked goods or parked vehicles;
- Creating blind spots at corners, doors or fence transitions;
- Expecting digital zoom after an incident to create detail that was never recorded.
5. Treat lighting as part of the CCTV system
A camera cannot recover detail that lighting and exposure have lost. Test the darkest and brightest expected conditions. Consider white light, infrared, wide dynamic range, shutter speed and the reflective properties of the scene.
Infrared can help in darkness, but insects, dust, nearby walls, glass and reflective clothing can degrade the image. Built-in infrared may be unsuitable for long or complex scenes. White light can improve colour evidence and deter activity, but it must be designed for safety, neighbours and site operations.
6. Calculate recording and retention
Retention should follow operational, legal, privacy, insurer and investigation needs. “Thirty days” is not a universal rule. Choose a justified period and ensure the system can deliver it under real recording conditions.
A basic storage estimate is:
Storage = average bitrate × recording seconds × number of cameras × retention days.
Convert bits to bytes and include headroom for overhead, scene variation, redundancy and future additions. Variable bitrate can change significantly when a busy, detailed scene replaces a quiet one. Motion recording saves storage only if detection is reliable and configured to include enough pre-event and post-event footage.
Define:
- Resolution, frame rate, codec and image-quality target;
- Continuous, scheduled, event or hybrid recording;
- Primary and protected evidence retention;
- Redundancy and drive-failure behaviour;
- Health alerts for camera, disk and recorder failure;
- Export format, watermarking or integrity controls and playback tools.
7. Design the network and cybersecurity controls
IP cameras are networked computers. Treat them accordingly. Estimate peak and average bandwidth, separate camera traffic where appropriate and protect management interfaces from unnecessary exposure.
Minimum controls should include:
- Unique credentials and removal of default passwords;
- Role-based accounts instead of shared administrator logins;
- Network segmentation and firewall rules;
- Secure remote access without exposing devices directly to the internet;
- Supported firmware and a documented update process;
- Time synchronisation across cameras, recorders and access systems;
- Configuration backups and controlled change records;
- Logs and alerts for access, failures and suspicious activity;
- A decommissioning process that removes data and credentials securely.
Agree who owns cyber maintenance after handover. A secure installation can become vulnerable if nobody monitors updates or user access.
8. Build power and communications resilience
Document what the CCTV system must do during load loss, generator transition, network failure and control-room outage. Use suitable backup power for cameras, network switches, recorders, displays and communications—not only the recorder.
Calculate backup duration against actual load and battery ageing. Protect outdoor links from lightning and environmental exposure where relevant. If remote connectivity fails, define whether local recording continues and how missed events are reconciled when the link returns.
9. Connect cameras to an operating response
Cameras do not respond to incidents. People and procedures do. Define which views are watched live, which events generate an alert, who verifies them and who can dispatch a guard or response team.
For analytics, measure more than detection accuracy. Record false alarms, missed events, operator workload, acknowledgement, verification and response. Analytics should be tuned to the scene and rechecked when lighting, vegetation, layout or activity changes.
Integration with alarms and access control can improve verification. For example, a forced-door event can automatically display the relevant camera and recent access record. This is one reason an integrated security model may reduce response hand-offs.
10. Address POPIA and privacy by design
Where CCTV captures identifiable people or entities, the footage can be personal information. POPIA requires appropriate and reasonable safeguards and, among other things, transparency about the collection of personal information in applicable circumstances.
Before commissioning, document:
- The lawful and specific purpose for each camera group;
- Which areas must not be viewed or recorded;
- Notices and privacy information appropriate to the site;
- Who may watch live footage, search recordings and export evidence;
- Retention, deletion and legal-hold rules;
- How access and exports are logged;
- How requests, complaints and security compromises are handled;
- Written operator terms where a service provider processes footage.
Avoid cameras in places where people reasonably expect a high degree of privacy unless a qualified adviser confirms a lawful and necessary basis. Use privacy masking and narrow fields of view where they can meet the security purpose with less intrusion.
11. Specify acceptance testing
Do not accept the system because every camera appears on a monitor. Test whether each outcome is achieved.
| Test area | Evidence to capture |
|---|---|
| Image purpose | Day and night samples showing the target at the designed distance |
| Recording | Playback and export from representative cameras and dates |
| Retention | Measured storage use and projected days under normal activity |
| Failure | Camera, disk, network and power alarms reach the responsible person |
| Response | End-to-end alarm or analytic scenario with timestamps |
| Access | Roles prevent unauthorised live view, search, export and configuration |
| Documentation | Accurate camera schedule, map, addressing, credentials process and configuration backup |
Keep accepted sample images so future maintenance can detect deterioration in focus, angle, obstruction or lighting.
12. Plan maintenance for usable evidence
Set a preventive maintenance schedule based on environment and risk. Clean lenses and housings, check focus and field of view, test recording and export, review storage health, confirm time accuracy, apply approved updates, inspect backup power and verify user access.
After site changes, revalidate affected cameras. A new sign, shelf, boom, floodlight or parking layout can quietly defeat an otherwise healthy system.
Business CCTV procurement checklist
- A documented purpose and target for every camera;
- Site plan with camera locations and fields of view;
- Day and night performance assumptions;
- Camera, lens, housing and mounting specification;
- Bandwidth and storage calculation with headroom;
- Network, remote-access and cybersecurity design;
- Power backup and failure behaviour;
- Monitoring, response and escalation procedures;
- POPIA roles, notices, access and retention controls;
- Acceptance tests, training, warranty and maintenance SLA;
- Asset register, configuration backup and exit documentation.
Frequently asked questions
How many CCTV cameras does my business need?
The number follows the required outcomes, site geometry and target detail. A site survey should identify critical views first. More cameras do not guarantee better evidence.
Is a higher megapixel camera always better?
No. Usable evidence depends on target size, lens, light, sensor, compression, shutter, placement and recording quality. Higher resolution can also increase bandwidth and storage.
How long should CCTV footage be kept in South Africa?
There is no single retention period suitable for every business. Choose and document a period based on purpose, operational discovery time, legal and contractual needs, privacy and storage. Delete footage when there is no valid reason to retain it, subject to lawful holds.
Can a CCTV camera read number plates at night?
It can when the camera, lens, angle, shutter, lighting and vehicle lane are designed for that purpose. A general overview camera may not capture reliable plates under headlights or movement.
Should CCTV be monitored live?
Live monitoring is valuable where an operator can detect or verify an event and initiate a timely response. For low-risk scenes, event-driven review or recorded evidence may be sufficient. Define the operating outcome before paying for monitoring.
Design for the moment the footage is needed
A business CCTV system succeeds when it consistently produces the required view, recording and response under real site conditions. Design from risk, test the result and maintain the full chain from camera to authorised action.
Planning a new system or upgrade? Request a CCTV site survey from Azomakhanye Security Services for a solution tailored to your risk profile and budget.