Reverse-osmosis efficiency is easiest to understand by measuring two streams: product water and concentrate, often called reject water. The recovery ratio tells you how much feed water becomes product water. It does not tell you how well the membrane rejects a contaminant.
The recovery calculation
At steady operation:
Recovery (%) = product-water flow ÷ feed-water flow × 100
Because feed water becomes either product or reject water, you can also use:
Recovery (%) = product flow ÷ (product flow + reject flow) × 100
Example: a unit produces 1 litre of product water while sending 2 litres to drain.
- Feed = 1 + 2 = 3 L
- Recovery = 1 ÷ 3 × 100 = about one-third
- Reject-to-product ratio = 2:1
| Product | Reject | Feed | Recovery | Reject per litre of product |
|---|---|---|---|---|
| 1 L | 1 L | 2 L | one-half | 1 L |
| 1 L | 2 L | 3 L | one-third | 2 L |
| 1 L | 5 L | 6 L | about one-sixth | 5 L |
The ratios are not interchangeable. A “2:1 pure-to-drain ratio” means something very different from “2:1 drain-to-pure”. Ask suppliers to label both streams.
Recovery is not rejection
Membrane rejection describes the reduction of a dissolved substance:
Rejection (%) = (feed concentration − product concentration) ÷ feed concentration × 100
If feed-water total dissolved solids are 500 mg/L and product water is 25 mg/L, the calculated TDS rejection is 95 per cent. That number does not prove the same rejection for every individual contaminant, and a handheld conductivity meter is not a substitute for analyte-specific laboratory testing.
The two measures can move independently. A system can have strong rejection but poor water recovery, or improved recovery while membrane scaling risk increases. Sound design considers both.
A useful current benchmark
The US EPA’s WaterSense specification for point-of-use RO systems requires certified models to send no more than 2.3 gallons of reject water to drain for each gallon of product water. That is about 30% minimum recovery. EPA says a typical point-of-use system may send five or more gallons to drain per gallon produced, while inefficient units may reach ten.
WaterSense is a US voluntary label, not an Australian legal requirement and not proof of a particular contaminant claim here. It is still a transparent efficiency benchmark because it defines the ratio and requires independent certification. EPA separately requires baseline performance such as total-dissolved-solids reduction and at least a one-year membrane life within that program; additional contaminant claims must be certified rather than assumed.
Why a brochure ratio may differ at home
RO output and recovery depend on operating conditions. Important variables include:
- feed pressure and temperature;
- dissolved-solids concentration and composition;
- membrane model and age;
- restrictor or pump design;
- storage-tank back pressure;
- prefilter condition;
- flushing and automatic shut-off behaviour.
Cold winter water usually permeates more slowly than warmer test water. A pressurised storage tank also changes the system’s behaviour as it fills. Therefore, compare products under stated test conditions and measure a complete operating cycle rather than timing the first few seconds from the drinking tap.
How to measure without opening the unit
Follow the manufacturer’s instructions and avoid altering pressurised plumbing. Where the product and drain streams can be safely captured during steady production:
- Empty or draw down the storage tank as directed so the unit is producing.
- Let flows stabilise.
- Collect product and reject water for the same measured period.
- Record both volumes and calculate recovery.
- Repeat to check consistency.
Do not use this procedure if it requires disconnecting pipes, bypassing controls or contaminating the drinking-water path. A technician can measure flow safely.
Efficiency should not compromise reliability
Pushing recovery upward concentrates salts on the feed side. In some waters, that increases scaling or fouling risk. Pretreatment, membrane selection, pressure and recovery must therefore be considered together. The “best” ratio is not automatically the highest number; it is an efficient ratio the complete system can sustain while meeting its water-quality and service-life objectives.
Daily demand also matters. If a household drinks and cooks with 12 litres of RO water per day:
- at 5:1 reject-to-product, about 60 L/day goes to drain;
- at 2:1, about 24 L/day goes to drain;
- the difference is about 13.1 kilolitres per year.
That estimate assumes the quoted ratios persist and excludes extra flushing cycles, so measured use may differ. It does, however, show why a ratio deserves more attention than a vague “water saving” label.
Questions for an RO comparison
- Is the ratio reject-to-product or product-to-reject?
- Under what pressure, temperature and TDS was it tested?
- Is the figure independently certified for the exact model?
- What contaminant-reduction claims are separately supported?
- How are filters, membrane condition and tank hygiene maintained?
- Will local feed-water chemistry need pretreatment?
Water Analytics Australia can help compare point-of-use RO and broader filtration options against source water, household demand and ongoing maintenance. Recovery is one measurable part of a good system—not a substitute for the full specification. For the broader decision, read whether RO is worth it for Australian tap water.