The question

With a Deposit Return Scheme (DRS) on the horizon, should a retailer:

• Pilot a small number of Reverse Vending Machines (RVMs),
• Move straight into a phased deployment, or
• Go for a near‑simultaneous full network launch?

And how early should RVM procurement and installation begin relative to the DRS go‑live date?

The direct answer

For any sizeable estate, the only rational strategy is:

1. Pilot first, in a small but representative set of stores, to learn from real customer behaviour, real store operations and real machine performance.
2. Lock down SOPs, alerting and configurations based on that learning.
3. Negotiate RVM procurement 12–16 months before go‑live, so capacity and approvals are in place.
4. Begin the physical rollout ~6 months before DRS go‑live, installing machines and either:
   • Keeping them switched off until launch, or
   • Running them in “trial mode” for internal learning or non‑DRS incentives.

Rolling out hundreds of RVMs “cold” and trying to improve as you go, store by store, almost guarantees avoidable cost, operational friction and inconsistent customer experience.

Why piloting is non‑negotiable

A pilot is not a marketing experiment; it is industrial prototyping at scale. Its purpose is to:

• Understand consumer reactions:

  • How quickly people understand the process.
  • Where they get stuck.
  • How they react to rejections and vouchers.

• Build technical understanding of the machines:

  • Stability across real store network conditions.
  • Sensitivity to environment (lighting, temperature, dust).
  • How alerts and remote diagnostics work in practice.

• Learn what assistance the RVM requires from store teams:

  • Frequency and duration of bag changes.
  • Printer paper usage and change patterns.
  • Real cleaning needs, not just the manufacturer’s manual.

A good pilot gives you evidence, not opinions. It answers questions like:

• Do we really need bin alerts at 70%, or is 85% enough if the store responds reliably?
• Should paper‑low alerts go to the duty manager, a central helpdesk, or both?
• Is it better to keep spare paper rolls and cleaning kit inside the machine cabinet, or in the backroom?
• How much backroom space is required for sealed bags, and how often are they moved?

If you skip this stage and move directly to a network rollout, you end up “learning on live customers” in 500 stores at once. That is expensive and hard to unwind.

Recommended timeline from DRS announcement to go‑live

A pragmatic timing model for a national DRS could look like this:

As soon as DRS is announced – start trialling RVMs

• Identify a small group of pilot stores:

  • A mix of formats (convenience, supermarket, high‑traffic urban, edge‑of‑town).
  • Different demographic and sales profiles.

• Deploy RVMs in these stores while DRS is still in design or legislative stages.

• Use them to:

  • Test technology in your own environment.
  • Refine layouts, signage and flows.
  • Start building internal knowledge in operations, IT and property.

If the scheme is, say, 22–24 months away, those months are valuable learning time.

12–16 months before go‑live – negotiate RVM procurement

• Use pilot learnings to write precise technical and operational requirements:

  • Alert thresholds.
  • Data and integration expectations.
  • Physical constraints (door swing, width, depth, back‑room space).

• Run your tender or supplier negotiations early:

  • Lead times for hundreds of machines, shelters and integration work are long.
  • DMOs will usually have an approval process; you want time to align with it.

• Secure contractual agreements that:

  • Lock in pricing and service levels.
  • Confirm capacity and delivery schedule.
  • Cover pre‑DRS trial use where required.

Early negotiation avoids last‑minute bottlenecks and “panic buying” on poor commercial terms.

~6 months before DRS go‑live – begin physical rollout

• Start installing RVMs into the wider estate around six months before the scheme starts:

  • This spreads the engineering and building workload.
  • It gives your teams time to become familiar with machines before the legal go‑live.

• For each installed RVM, you have two options:

  • Keep it switched off until DRS launches, using the period for back‑end testing and staff training; or
  • Run it in “trial mode” with:
    • Internal recycling incentives.
    • Non‑statutory loyalty rewards.
    • Soft customer communication (“coming soon – national deposit refunds”).

In both cases, by the time DRS goes live:

• The kit is in place.
• Staff are trained.
• SOPs and alerts are already tuned from the pilot experience.

What a good pilot should include

A robust pilot is not just “we put three machines somewhere and watched”. It should be structured around:

Representative store selection

Include:

• At least one high‑volume flagship.
• One or more typical supermarkets.
• A handful of convenience formats.
• Any special cases (travel, city centre, retail park).

Clear KPIs

Examples:

• Containers returned per day per store and per machine.
• Average and peak return session length.
• Frequency and duration of “full bin” and “no paper” states.
• Time taken for staff to respond to alerts.
• Incident rates (misuse, attempts at fraud, vandalism).
• Customer complaints and NPS‑type feedback around the return experience.

SOP and alert tuning

Use pilot data to make decisions such as:

• Bin full alert at 80% vs 70% vs 90% – what gives best balance of safety margin and staff workload?

• Paper‑low alert:

  • Should we change rolls at first alert, or run them to zero and change once?
  • Where does the alert go: local, central, both?

• Cleaning routines:

  • Is daily superficial clean enough, or do stores prefer to link it to bag‑change events?
  • Which cleaning materials work best with the hardware and sensors?

Backroom and logistics planning

Learn:

• How many sealed bags does a store accumulate before collection?
• How much space is needed, and where is it safest to store them?
• How does this interact with existing cage movements and waste streams?

All of this then feeds directly into your network‑wide SOPs and training materials.

Why “full network launch then improve as we go” is a bad idea

Rolling out, for example, 500 RVMs without a pilot and then optimising “on the fly” sounds agile but is operationally risky because:

• Every change (alert thresholds, SOPs, data flows) must then be communicated to hundreds of stores and teams.

• Inconsistent behaviour across the estate leads to:

  • Uneven customer experience.
  • Confusion for central support and helpdesks.
  • Difficulties in reporting and reconciliation.

• Basic decisions (for example where to keep spare paper rolls, how to set up backroom bag areas) become improvised locally instead of being defined centrally based on evidence.

The cost is not just in money; it is in lost trust—from customers, from staff, and potentially from the DMO if performance varies widely.

Pilot → hardened model → phased rollout

The optimal pattern, therefore, is:

1. Pilot (small scale)

   • 5–20 stores, depending on estate size.
   • Run long enough to see normal and peak patterns (weeks, not days).

2. Harden the model

   • Freeze key parameters and SOPs based on pilot outcomes.
   • Finalise integration patterns with POS, loyalty and, where applicable, DMO systems.
   • Document everything: staff guides, troubleshooting flows, cleaning and safety procedures.

3. Phased rollout

   • Roll out by region, banner or format in waves, not all at once.
   • Use early waves to catch any residual edge cases before you reach full coverage.
   • Maintain a feedback loop so that small improvements can still be made—but with version control and governance, not ad‑hoc tweaks.

4. Final network launch at DRS go‑live

   • All or nearly all machines in place.
   • Staff trained and rehearsed.
   • Communications to customers coordinated with the scheme start date.

Summary

For a serious multi‑site retailer, the optimal RVM rollout strategy around a DRS is:

• Pilot early as soon as DRS is announced, to learn from real customers, real stores and real machines.
• Negotiate RVM procurement 12–16 months before go‑live, using pilot insight to set clear requirements.
• Start installations about six months before go‑live, bringing machines on‑line in a controlled, phased manner and, if desired, using trial mode.
• Avoid “500‑store experiments” by freezing your SOPs and alerting logic based on pilot evidence before you attempt full coverage.

This approach reduces risk, stabilises operations and gives you the best chance of entering DRS go‑live with an estate that is compliant, cost‑effective and ready to scale.

The unique features of Recyclever RVMs: https://www.recyclever.com/why-recyclever-reverse-vending-machines