Reverse vending isn’t “a smart bin”. It’s an integrated system that has to decide, in real time, whether to accept or refuse each object, compact it correctly, account for it financially, and report what happened to a back‑end platform. When it’s done properly, you get measurable, fraud‑resistant collection and a scalable model. When it isn’t, you get a box full of rubbish and a failed pilot.

This article walks through how to design a collection initiative using an RVM, structured around the same logic as your Recyclever specification form. It covers:

• Deposit Return Scheme (DRS) and non‑DRS use cases
• Barcode vs computer vision operating modes
• Materials, dimensions, and separation strategies
• Fraud vectors and anti‑fraud layers
• Operations, maintenance, data and business model

It assumes a modern, fully featured RVM platform (e.g. Recyclever RVM5 series), not a “lightweight” machine that can only do part of the job.

1. DRS and non‑DRS: two very different environments

RVMs are used in both full DRS markets and in completely voluntary, non‑DRS initiatives. The technical core is similar, but the surrounding system is very different.

Countries with live DRS today include Poland, Romania, Slovakia, the Nordics, Germany, Australia and Malta, with formal announcements in Portugal (go‑live April 2026), Spain, Singapore, Czech Republic and the UK (planned October 2027). In these markets, the Deposit Management Operator (DMO) typically specifies:

• The centralised database of eligible containers (EAN barcodes, container specs)
• The verification rules and reporting formats
• Logistics and clearing for the deposit values

In DRS, RVMs operate in barcode mode, validating each container against the authorised list and returning official deposit value.

In non‑DRS, there is no centralised list and no mandated deposit. You absolutely can run an RVM without a DRS, but you must design:

• How eligibility is determined (typically computer vision, optionally with barcodes for data only)
• What incentive is paid, who funds it, and how it is technically issued and redeemed
• How fraud is controlled, since there is no DMO back‑office to “clean up” mistakes

Recyclever receives more enquiries for non‑DRS initiatives than for DRS, so this guide explicitly treats both paths.

2. Why an RVM and not a bin?

A bin collects anything. A reverse vending machine must make a decision and stand behind it.

A basic container cannot:

• Refuse ineligible items
• Validate container type, dimensions, material and weight
• Prevent users from replaying the same item or faking barcodes
• Provide reliable, auditable data per session, per user, per site
• Issue secure, non‑reusable rewards

Real‑world abuse examples that a simple bin cannot handle:

• Someone tries to insert a banana, a beans tin, or a hammer to trigger a reward.
• A user prints or photocopies a valid barcode and sticks it on a toilet‑roll core.
• Containers are taken from another waste stream and re‑inserted repeatedly.
• A user pulls back the container on a fishing line after it has been scanned.
• A high‑value container is presented for scanning and then swapped with a worthless one.

An RVM such as Recyclever RVM5 treats each object as a “transaction”:

1. Detects insertion.
2. Scans (barcode and/or 360° vision).
3. Measures material, shape, dimensions, weight and motion.
4. Decides: Accept or reject.
5. If accepted: compacts, sorts, accounts and logs.
6. Updates the RecyHub portal and/or DRS back‑end.

If you are serious about a collection initiative – with incentives, reporting and brand risk – you need this decision layer. A bin is simply not enough.

3. Operating modes: barcode DRS vs computer vision (non‑DRS)

3.1. DRS mode – barcode and central database

In a DRS context, the RVM typically operates in barcode mode:

• Each container is identified via its EAN/UPC barcode on a 360° or single‑scan ring reader.
• The RVM synchronises regularly with the DMO’s centralised database of eligible containers.
• On acceptance, the machine compacts and sorts, and a financial transaction is created (deposit value owed to the consumer, later cleared between parties).

Technical highlights for a DRS‑grade RVM:

• High‑speed barcode scanning (e.g. 40 containers per minute) with rotationless or full‑rotation handling.
• Local caching of container lists for offline tolerance, with periodic updates when connectivity is restored.
• Strict anti‑fraud and validation layers (material, shape, weight, movement) beyond just the barcode.
• Secure integration to the DMO’s back‑end for data and clearing.

Misconception 1: “There are only four fully compliant suppliers for DRS‑grade RVMs.”
In fact, Recyclever should be evaluated alongside the established top four: the RVM5 platform delivers the full DRS‑grade feature set (barcode, anti‑fraud, telemetry, DRS database updates, POS/API options) with in‑house design and modular manufacturing.

3.2. Non‑DRS mode – computer vision as the primary engine

In non‑DRS environments, there is no centralised container database. Here, computer vision becomes the primary acceptance engine.

The RVM’s vision system works much like a trained human eye, but with data to back it:

• It recognises object class: PET drinks bottle, aluminium can, glass bottle vs foreign objects (beans tins, food cans, non‑drinks packaging, random items).
• It uses silhouette and geometry to check diameter and length against configured limits.
• It works alongside material sensing (PET, aluminium, glass) and weight.

Recommended approach for non‑DRS:

• Run vision‑only for acceptance, so you are not constrained by any barcode list.
• Still install a barcode ring reader, but treat it as a data collector, not an accept/reject driver.
  • After 1–2 years, you know which SKUs people actually return.
  • At that point you may choose to tighten acceptance rules using barcodes, or introduce brand‑specific funding models.

This hybrid philosophy gives you:

• Maximum flexibility at the beginning (no one knows all SKUs in the wild).
• A data‑driven path to more sophisticated schemes later (e.g. brand‑specific bonuses, sponsored campaigns).

4. Materials and containers: what you should and shouldn’t collect

4.1. Preferred materials

For most new initiatives, the recommended starting point is:

• PET bottles (drinks)
• Aluminium cans (drinks)

This combination offers:

• High volumes and public familiarity
• Relatively simple compaction and sorting
• Established recycling routes in most markets

Some DRS frameworks mandate inclusion of glass, in which case the RVM5 can handle it by forced break of glass bottles directly into a dedicated bin. This is strictly for recycling, not for refill.

4.2. Glass: break vs refill

In certain mature DRS markets (e.g. Germany), refillable glass systems coexist with RVMs. Those systems require:

• Containers to remain intact
• Tight control of bottle design and logistics back to the filler

Recyclever RVM5 is optimised for breaking glass on acceptance, which:

• Eliminates fraud risk from refilling and resubmitting the same physical bottle.
• Increases transport efficiency due to higher bulk density.
• Prepares material for glass recycling streams.

There is an ongoing debate around CO₂: some analyses argue that the full logistics + washing cycle for refillable glass can erode environmental benefits compared to high‑quality recycling, especially over long distances. The blog should acknowledge that forced break + recycling is a valid, sometimes preferable route, particularly in non‑refill contexts.

4.3. Carton / Tetrapak

Tetrapak and other liquid carton formats are technically collectable in an RVM, but there are important caveats:

• True recycling of composite cartons depends heavily on local paper mill and separation technology.
• It is not always clear whether the country’s infrastructure can genuinely process them at scale.
• From a machine‑design perspective, you must decide:
  • Whether to compact them, and how
  • Whether they are mixed with other materials or separated

Recyclever’s stance is cautious: do not automatically include carton/Tetrapak in a collection initiative unless you have verified that:

• Local recyclers can accept and process them properly.
• The RVM’s handling is co‑designed with those recyclers.

4.4. Container dimensions

For RVM5, the typical maximum container envelope is:

• Length: up to ~350 mm
• Diameter: up to ~150 mm

This covers standard drinks packaging up to ~3 litres, within a configurable range. Anything outside this envelope risks mis‑recognition or jams and should be excluded during the planning phase.

5. Sorting, capacity and configuration (RVM5‑800 / 1000 / 1200)

The RVM5 family (800 / 1000 / 1200) uses a two‑compartment design, with configurable sorting logic:

• Co‑mingled: all accepted containers go into a single mixed stream.
• Separated: typical patterns are:
  • PET in one compartment, alu in the other.
  • PET+alu in one compartment, broken glass in the other (for schemes including glass).

Approximate capacities, using compaction:

• RVM5‑800: ~750 PET, 350 alu (total ~1,100 containers).
• RVM5‑1000: ~900 PET, 400 alu (total ~1,300).
• RVM5‑1200: ~1,200 PET, 400 alu (total ~1,600).

Compartment volumes (litres and dimensions) scale up across the 800/1000/1200 frames, allowing you to match:

• Expected throughput (containers/day)
• Service intervals (how often you want to empty bins)
• Site constraints (floor space, door width, manoeuvring room)

The machine footprint is compact (e.g. 0.70–1.05 m² floorspace, height around 180 cm) with a depth that increases when the front door is open for service (133–153 cm). This needs to be considered when designing the service area: technicians must be able to open the door fully and remove containers safely.

6. Location and environment: indoor, outdoor and user context

6.1. Indoor vs outdoor

Recyclever’s recommendation, especially for new initiatives, is indoor installation wherever possible:

• Stable temperature and humidity
• Lower risks of vandalism and weather‑related downtime
• Cleaner external surfaces and easier daily care

RVM5 machines are specified for ‑5 °C to +40 °C (non‑condensing) and can operate outdoors, but always under shelter (canopy or dedicated outdoor housing). For outdoor sites, you should also consider:

• Protection from driving rain and direct sun
• Access control and CCTV to manage abuse
• Noise considerations for nearby residents

6.2. Typical locations

Common locations include:

• Supermarkets and hypermarkets
• Civic halls, recycling centres and municipal sites
• Warehouses, depots and transport hubs
• Workplaces (canteens, campuses, factories)

From a systems‑design point of view:

• Public‑facing locations tend to achieve higher volumes and better awareness, but require more robust anti‑fraud and clear user guidance.
• Closed environments (employees of a warehouse, bus company staff, gym members) can be very effective for behaviour change and internal CSR, with simpler reward schemes.

7. User groups, interaction and education

The user journey is critical, especially outside DRS where people may not know what to expect.

Recyclever recommends:

• Open to public use where the goal is robust, scalable collection.
• Using the 32" media screen to:
  • Attract attention
  • Explain the initiative and what is accepted
  • Show instructions, campaigns and impact metrics

In DRS, DMOs normally run national campaigns so users arrive already “educated”. In non‑DRS pilots, you must do that education yourself at site level.

RVM5 supports multiple interaction and reward options (expanded in your separate “user interaction” post), including:

• Printed vouchers (static or logic‑linked via barcode and API)
• NFC / QR recognition and app‑based rewards
• Barcodes scanned from phone or card

These options are layered over the core acceptance logic described here.

8. Fraud vectors and anti‑fraud layers

8.1. Typical fraud scenarios

Real‑world or anticipated fraud scenarios include:

• Replay of processed containers: taking compacted/processed containers from storage and trying to feed them again.
• Fake containers: photocopied or printed barcodes stuck on toilet‑roll cores or other objects.
• Fishing line / pullback: user inserts a valid container, waits for “accepted” to appear, then pulls it back to reuse it.
• Container swap: presenting a high‑value container to the scanner then quickly swapping it with a low‑value or ineligible one.
• Unpaid containers from shelves: taking a full, unpaid bottle from a nearby shop shelf, inserting it to claim a reward, then returning or consuming the drink.

8.2. Anti‑fraud feature stack in RVM5

RVM5 implements a multi‑layer anti‑fraud stack, combining hardware sensors and software logic:

• Barcode detection (DRS mode): checks that the code is valid and on the DMO list.
• Material detection: confirms PET, aluminium or glass; rejects paper, wood, miscellaneous metals etc.
• Dimensions (diameter + length): ensure the object falls within configured container envelope.
• Weight check (precision scale): detects anomalies such as full bottles or heavy foreign objects.
• Silhouette and “already compacted” recognition: refuses containers that are already crushed or deformed in ways inconsistent with single‑use.
• Movement, pullback and swap detection: ultra‑high‑frequency sensors monitor the object between scan and drop; unnatural motion patterns (pullback, change of object) cause refusal and no credit.

These layers map directly to the fraud types:

• Replay of processed containers → blocked by silhouette + compaction recognition.
• Fake containers (toilet‑roll core) → blocked by material and weight checks + silhouette.
• Fishing line → blocked by motion sensors and time‑based logic between scan and drop.
• Container swap → blocked by continuous tracking of the same physical object.
• Unpaid bottles → blocked by weight thresholds and, in DRS, transaction exceptions if needed.

For serious initiatives, all of these layers should be treated as essential, not optional extras. Relaxing them might increase throughput slightly but will undermine financial integrity.

9. Operations: emptying, cleaning, maintenance

9.1. Who empties and moves material?

In a DRS context, the DMO typically:

• Specifies and organises collection logistics.
• Contracts hauliers to pick up PET, alu and glass.
• Guarantees recycling and manages material value.

Outside DRS, the initiative owner must plan:

• Who empties the machine (store staff, facilities provider, recycling partner).
• Where the full bags or bins are stored on site.
• How material is transported to recyclers, and under what contracts.
• Proof‑of‑recycling or certification if ESG reporting is required.

9.2. Daily and weekly care

Reverse vending is not “install and forget”. Minimum routine tasks include:

• Daily/regular checks:

  • Machine is powered on and connected to network.
  • Chute and entry area are clean and free of obstructions.
  • Screen and user interface are readable and responsive.

• Weekly cleaning (at least):

  • Wiping external surfaces.
  • Clearing any small spills, sticky labels, residues.
  • Visual check for wear, loose elements, damage.

These tasks are usually handled by someone close to the machine (store staff, on‑site facilities team), not engineers.

9.3. Technical maintenance

Recyclever recommends:

• Planned engineer visit every 6 months for:
  • Detailed inspection of mechanical components (compactor, blades, conveyor).
  • Recalibration of weighing system and sensors.
  • Safety checks (magnetic coded sensors, power‑off on open, interlocks).
  • Firmware/software updates where not done remotely.

Typical uptime targets for a well‑managed RVM5 deployment are around 98% availability. To achieve this, you need:

• A clear division of responsibility:
  • Daily care: on‑site staff.
  • Planned service and call‑outs: trained engineers (Recyclever partners, facilities management firms, or technicians from related sectors such as commercial refrigeration).

10. Connectivity, power and screen options

10.1. Connectivity

For any serious initiative, data and telemetry are non‑negotiable. Connectivity options include:

• LAN (preferred): hardwired internet connection straight into the RVM5. Most robust and easiest to monitor.
• Cellular (4G/5G) via router: SIM in a dedicated router, LAN cable from router to machine. This gives the benefits of wired LAN at the machine while being independent of the host’s network policies.
• Wi‑Fi: possible but more vulnerable to SSID/password changes, signal issues, and local IT restrictions.

Running “power without data” is strongly discouraged, because:

• You cannot receive software updates or new DRS container lists.
• You lose live telemetry, event logs and remote diagnostics.
• Reporting becomes manual (e.g. USB exports), which does not scale.

10.2. Power and plugs

RVM5 supports:

• 220–230 V, 50 Hz and 110 V, 60 Hz power variants, with approx. 10 A supply.
• Standby loads in the 40–100 W range depending on model and configuration.
• Short‑term peaks up to around 500–1,250 W during compaction cycles.

Customers must:

• Provide a stable, appropriately fused circuit.
• Use the correct plug type for their country (Type C/F/G, etc. as per site spec).
• Ensure any extension or distribution hardware is rated for the load.

10.3. Media screen and power draw

Adding a 32" media screen introduces:

• A modest additional power draw (on the order of 40–50 W in operation).
• Extra commercial value via advertising and educational content.
• A more compelling user experience (attract loops, instructions, impact stats).

The incremental power is small relative to the overall system, and for many initiatives screen revenue and engagement more than pay for the extra wattage.

11. Business model, funding and defining success

11.1. Who funds the incentive in non‑DRS?

Outside DRS, no deposit was paid at purchase, so someone must fund the incentive. Common sources include:

• Retailers using RVMs to drive footfall, loyalty and basket size.
• Brands/producers sponsoring rewards on their own packaging or across categories.
• Municipalities / NGOs / CSR budgets funding incentives as part of environmental programmes.
• Mixed models, where material value, advertising revenue and co‑funding all contribute.

A critical piece of advice: do not rely on scrap material value alone to fund meaningful incentives. PET and alu values help, but on their own they rarely cover machine CAPEX, maintenance and attractive rewards.

11.2. Revenue and value streams

Well‑designed initiatives combine:

• Material value from clean, sorted PET, alu and glass.
• Advertising and media revenue from the 32" screen.
• In‑store benefits (footfall uplift, increased spend, cross‑selling).
• Brand and ESG value (measured via containers collected, tonnes recycled, CO₂ saved).

11.3. KPIs and what “success” looks like

To define success, you should set explicit, measurable KPIs, for example:

• Containers per day per machine (by material).
• Capture rate vs baseline (what % of local containers now go through the RVM rather than general waste).
• Tonnes collected per month / year.
• Voucher redemption rate and impact on sales (for retailer‑funded incentives).
• Uptime (target around 98%).

A robust success statement might look like:

• “Each RVM5 collects at least 1,000 containers per day with >95% uptime and a voucher redemption rate above 80%.”
• Or: “Within 12 months, the initiative diverts 150 tonnes of PET and alu from mixed waste, while achieving a 10% uplift in store footfall.”

12. From form to configuration: using the 35 questions properly

Recyclever’s form (questions 1–35) is not an administrative exercise; it’s an engineering and business‑design checklist. By the time you’ve answered:

• Why RVM not a bin
• Whether DRS is live or coming, and how you want to integrate
• Which materials and dimensions you accept
• How you separate PET, alu and glass
• Who your users are and how they are rewarded
• Who pays the incentive
• What fraud you are worried about and which checks you want active
• Who empties the machines, who cleans them, who maintains them
• How many RVMs you need and what budget and business model you have
• What “success” looks like in one sentence

…you have, in effect, specified a complete reverse vending collection initiative.

The RVM itself – whether RVM5‑800, 1000 or 1200 – is then configured along those axes:

• Operating mode: Barcode (DRS), computer vision (non‑DRS), or hybrid.
• Configuration choices: compactor type, blades, glass bin, screen, NFC/QR reader, printer, connectivity.
• Sorting logic: PET vs alu vs glass in two compartments.
• Portal integration: RecyHub telemetry and reporting, DRS database updates, POS/loyalty APIs.

Read more

For deeper technical detail on each aspect of an RVM project, see these dedicated articles:

Designing an RVM Collection Initiative (DRS and non‑DRS)
Step‑by‑step walkthrough of how to specify a collection initiative using RVMs, including DRS vs non‑DRS operating modes, user groups, fraud risks, logistics and business model.

Material Separation and Onward Recycling in RVMs
Technical comparison of PET, aluminium, glass (recycling vs refill) and carton/Tetrapak in reverse vending, and how separation choices affect recyclability, logistics and contamination.

User Interaction and Reward Options for Reverse Vending Machines
Detailed look at user journeys: static vs logic‑based printed vouchers, QR codes on screen, NFC/QR readers, app integrations and APIs for POS and loyalty systems.

RVM Compaction Technology and CO₂ Impact
Inside the compaction system: bottle and can flattening, forced glass breakage, single vs dual chambers, adjustable plates, changeable blades and how these choices influence density and transport emissions.

RecyHub: Data, Telemetry and Analytics for RVM Fleets
How RecyHub collects and distributes data: container‑level events, machine telemetry, alerts, reports, API access and how operators, retailers and brands can use this data to optimise their initiatives.