DUoS charges, explained.

What DUoS pays for, and who sets it

Electricity reaches a site over two networks. The transmission system carries it across the country at very high voltage and ends at a grid supply point. From there the distribution system carries it the rest of the way, stepping the voltage down as it goes, until it reaches a meter. DUoS is the charge for that second journey.

Great Britain is divided into fourteen distribution network areas, each run by a licensed distribution network operator. The operator owns the wires, the substations and the transformers in its region, and is paid an allowed revenue set by the regulator through the price control. DUoS is how it recovers that revenue from the people who use the network, whoever supplies them.

The instrument

The rates are not a commercial decision. Standard Licence Condition 14 of the electricity distribution licence obliges each operator to publish a use of system charging statement, and the charges in it are calculated under methodologies set out in the Distribution Connection and Use of System Agreement, the industry code that governs connection to and use of the distribution networks.

• Schedule 16, the CDCM. The Common Distribution Charging Methodology, which sets the charges for low voltage and high voltage designated properties. This is the methodology behind almost every business bill, and the one this guide describes.
• Schedule 17, the EDCM. The Extra-High Voltage Distribution Charging Methodology, for the Designated EHV Properties the licence assigns to it. The statement puts that threshold at 22kV, which is where it defines Extra-High Voltage as beginning. EDCM charges are calculated site by site rather than from a published tariff table.
• Schedule 29, the PCDM. The Price Control Disaggregation Model, which sets the discount percentages used for licensed distribution network operators embedded inside another network.

Forward-looking and residual

Every DUoS tariff is built from two different ideas, and telling them apart explains most of what looks strange in a schedule of charges.

The forward-looking part signals to users how their behaviour affects future network costs. It is the reason peak-time units cost more than overnight ones, and the reason capacity is charged for at all. Ofgem is explicit that the revenue recovered from forward-looking elements is typically less than the allowed revenue due to be recovered, because a lot of network cost does not change with use.

The residual part closes that gap. It is the difference between what the forward-looking charges bring in and what the operator is allowed to recover, and since the Targeted Charging Review it is collected through fixed charges rather than through anything a site can avoid by shifting demand. Domestic customers sit in a single band. Non-domestic customers are allocated to one of four residual charging bands for each charge structure, with boundaries calculated nationally under Schedule 32 of the DCUSA. A site certified as non-final demand pays no residual at all and goes onto a No Residual tariff.

Which tariff your meter is on

A schedule of charges runs to dozens of tariffs and only one of them is yours. Four things decide which, in this order, and the operator decides all four rather than your supplier.

1. 01The network area
   The first two digits of the MPAN are the distributor identifier, and they name the operator whose schedule applies. Southern is 20, north Scotland is 17, London is 12. Nothing else about the supply changes this.
2. 02The voltage of connection
   Low voltage, low voltage substation, or high voltage. The operator decides it, generally from where the metering sits and where responsibility for the equipment passes from the operator to the customer. Extra-high voltage leaves the CDCM entirely and is charged under the EDCM.
3. 03How the meter is settled
   A site whose consumption reaches the operator as site-specific half-hourly data is billed the full set of components. A site settled on aggregated data is billed a fixed charge and three unit charges only, with no capacity and no reactive power charge.
4. 04The residual charging band
   One of four for non-domestic customers, allocated by agreed capacity for site-specific billing and by consumption for aggregated billing. It is the reason the same tariff name appears four times in a schedule with four different sets of numbers.

The result of those four is a Line Loss Factor Class, also called a DUoS Tariff ID: a short code in the operator schedule that names exactly one row of rates. If you can find that code on a bill or from your supplier, you can look your own rates up in the published document and check them line by line. If you cannot, nothing else in this guide is much use, so it is the first thing worth asking for.

The components a DUoS charge is made of

A site-specific DUoS charge is not one number. Under the CDCM it is up to seven separate charges on four different bases, and they can move independently of each other.

Two of those are worth pausing on. The capacity charge is not a charge for electricity: it is rent on a reservation, paid daily against the maximum import capacity agreed with the operator, and it is owed in full by a site that ran at a fraction of it or did not run at all. The reactive power charge is not a charge for electricity either. It applies to the reactive units a site draws above a threshold, and Southern sets that threshold at 33% of active units, which is what a power factor of 0.95 works out to.

Both are among the very few lines on a business bill a site can act on directly. Agreed capacity can be reduced by agreement with the operator, and reactive power charges are described in the Southern statement itself as generally avoidable where the electrical installation is designed to hold a power factor between 0.95 and unity. Both are covered in more detail in the agreed capacity and reactive power glossary entries.

Red, amber and green, and why the windows differ

The three unit rates are the part of DUoS most people have heard of and the part most often described wrongly. The bands are not national. Each operator publishes its own windows in its own schedule, and they differ in start time, in length and in what happens at weekends.

Two rules hold across the operators covered here. All times are UK clock time, which means the windows move with the clocks rather than staying fixed in UTC. And bank holidays count as weekdays, which catches out anyone who assumes a public holiday is charged like a Sunday.

Three hours a weekday sit in the red band in both areas, which is 66 hours out of the 720 in a thirty day month, or nine per cent of the month. The pricing does not follow that proportion at all. On the Southern LV Site Specific Band 2 tariff the red rate is 6.243p per kWh and the green rate is 0.028p, so a unit moved out of the red window and into the green one costs a two hundred and twenty third of what it did.

That ratio is the single most useful thing to know about distribution charges. It means the share of a DUoS bill that came from red-band consumption is almost never the share of kWh that landed in the red band, and it means load shifting is worth far more against DUoS than the headline unit rate on a supply contract suggests.

A month of DUoS at published rates

Here is a complete DUoS charge for one month, at the rates published for the 2026/27 charging year. The site is an assumption and is labelled as one. Every rate, every band window and the reactive threshold are the published figures.

Assume a half-hourly low voltage site in the Southern area, on the LV Site Specific Band 2 tariff, LLFC H87. Assume it holds 500 kVA of agreed capacity, that it draws a steady 400 kW around the clock through September 2026, and that it runs at a power factor of 0.90. September 2026 has 22 weekdays, 8 weekend days and no bank holiday, and a steady 400 kW over its 720 hours is 288,000 kWh.

Red, 66 hours at 400 kW

26,400 kWh

Amber, 360 hours at 400 kW

144,000 kWh

Green, 294 hours at 400 kW

117,600 kWh

Red at 6.243p per kWh

£1,648.15

Amber at 0.509p per kWh

£732.96

Green at 0.028p per kWh

£32.93

Fixed at 0.00p per MPAN per day, 30 days

£0.00

Capacity at 10.63p per kVA per day, 500 kVA, 30 days

£1,594.50

Reactive at 0.251p per kVArh on 44,438.4 chargeable kVArh

£111.54

DUoS for the month

£4,120.08

Where the reactive figure came from

A power factor of 0.90 means reactive import runs at 0.4843 kVArh for every kWh, so 288,000 kWh comes with 139,478.4 kVArh. The Southern allowance is 0.33 kVArh per kWh, which is 95,040 kVArh over the month, and only the excess is chargeable. 139,478.4 less 95,040 is 44,438.4 kVArh, which at 0.251p is £111.54. Correcting the power factor to 0.95 would remove that line entirely.

That single subtraction works here only because the load and the power factor are assumed constant. Paragraph 2.53 of the statement does the calculation in every half hour and sums the results, so on a real site the good half hours cannot pay for the bad ones. Rebuilding a month from monthly totals nets them off and returns a smaller chargeable figure than a correct invoice, which reads as an overcharge when it is a rebuilding error.

What each component actually contributed

The interesting part is not the total. It is how little the proportions look like the consumption.

Nine per cent of the units carried forty per cent of the bill. Forty-one per cent of the units carried less than one per cent of it. And more than a third of the charge came from two components that are not measured in kWh at all, so no amount of energy efficiency touches them.

The same site, moved to the north of Scotland

Now take the identical site, the identical month, the identical load and the identical tariff name, and put it in the north Scotland area instead. The band windows are different, so the consumption falls differently, and the rates are different too.

Red, 66 hours at 400 kW

26,400 kWh

Amber, 306 hours at 400 kW

122,400 kWh

Green, 348 hours at 400 kW

139,200 kWh

Red at 8.049p per kWh

£2,124.94

Amber at 2.321p per kWh

£2,840.90

Green at 0.299p per kWh

£416.21

Fixed at 0.00p per MPAN per day, 30 days

£0.00

Capacity at 13.83p per kVA per day, 500 kVA, 30 days

£2,074.50

Reactive at 0.442p per kVArh on 44,438.4 chargeable kVArh

£196.42

DUoS for the month

£7,652.97

Nothing about the business changed. The same 288,000 kWh cost £4,120.08 in distribution charges in one area and £7,652.97 in the other, a difference of £3,532.89 in a single month, because the two operators run different networks with different costs and recover them from different numbers of customers. This is what people mean when they say network charges are regional, and it is why a DUoS rate quoted without a network area attached is not a rate at all.

Where it appears on your invoice

The operator publishes seven component names. Suppliers do not use them. The same component appears on invoices under several different labels, and the same label can mean different components at different suppliers, which is why a bill that shows a distribution total and nothing else is effectively unauditable.

Unit Rate 1, 2 and 3 are the trap in that table. On a supply contract those three labels usually mean day, night and weekend energy rates. On the distribution section of the same invoice they usually mean red, amber and green. Reading one as the other produces a comparison that looks arithmetically fine and is answering a different question.

How to check it

DUoS is one of the few charges on a business bill that can be reproduced exactly rather than sampled, because the operator publishes every input except your consumption. Working in this order matters, because a failure early on makes everything after it meaningless.

• The tariff. Take the LLFC or DUoS Tariff ID for the meter and find that row in the operator schedule for the charging year the bill covers. A wrong tariff produces a perfectly arithmetic bill for somebody else.
• The charging year. Rates change on 1 April. A bill spanning that date needs two sets of rates, one for each side of it, and a single blended rate across the boundary is wrong by construction.
• The band allocation. Check the kWh billed in each band against half-hourly data allocated to the published windows for that operator, in UK clock time, with bank holidays treated as weekdays.
• The rates. Compare each component against the published figure to the last decimal place the schedule prints. Distribution rates run to three decimal places of a penny and get truncated to two in transcription more often than they get typed wrongly.
• Capacity. Multiply the agreed capacity by the days in the period. Then ask a separate question: is the agreed capacity anywhere near the real peak? A capacity set for equipment that left the site years ago is the most common recurring overcharge in this whole area, and it is fixable.
• Reactive power. Rebuild the allowance before comparing the amount, and rebuild it half hour by half hour rather than from the month, which is how the statement computes it. A reactive line billed on total kVArh rather than on kVArh above the threshold is a large error that looks like a small one.

One thing not to check, because it is not there. Line loss factors gross metered volume up for several charges on a bill, and distribution is not one of them. The Southern statement is explicit that the loss adjustment is used in settlement and does not affect distribution charges. A DUoS line reconstructed from grossed-up volume will be a few per cent too high, and the error will look like a supplier overcharge.

What is changing

Two things are moving in distribution charging, and both of them are visible in the 2026/27 rates rather than being a forecast.

Surplus residuals broke the model, and are being patched

When the residual comes out negative, the CDCM discounts fixed charges and unit rates to bring recovery down to target. A large enough surplus residual exhausts that discounting and the model stops producing a complete set of tariffs. Something has to give, because an operator that cannot publish a tariff for a group of sites cannot bill them at all.

For 2026/27 this was live. Ofgem directed three network operators to depart from the notice period and the charging methodologies. In the Southern and north Scotland areas the intervention was a scaler applied to the gross asset values in the Distribution Reinforcement Model, 99% in Southern and 82% in north Scotland. Without it, Ofgem recorded that high voltage site specific customers in bands 3 and 4 in Southern, roughly 550 meter points, and bands 2, 3 and 4 in north Scotland, roughly 200 meter points, would not have been billable at all.

You can see the result in the published tables. The Southern HV Site Specific Band 4 tariff for 2026/27 carries a red unit charge of 0.000p, an amber charge of 0.000p, a green charge of 0.000p and a fixed charge of 0.00p. Everything that site pays in distribution charges comes from its capacity charge of 10.06p per kVA per day and its reactive power charge of 0.098p per kVArh.

The permanent fix is DCUSA modification DCP450, approved by Ofgem on 20 October 2025, which adds a step to the CDCM letting operators reduce the gross asset value input uniformly so that tariffs can be produced without a derogation. It takes effect on 1 April 2027 and applies to the CDCM only. Nothing equivalent has yet been raised for the EDCM, which is why extra-high voltage sites are still being handled case by case.

Notice periods are being cut, one year at a time

The fifteen months of notice in clause 19.1A is the reason DUoS is checkable so far ahead. It is also the first thing to go when the model misbehaves. Ofgem can direct an operator to depart from it under clause 19.1B, and the notice period then drops to forty days. It did so for SSEN on 19 December 2025 in respect of the 2027/28 charges for both its areas, with publication expected by 31 January 2026 rather than 31 December 2025.

If you budget from published DUoS rates, this is the practical consequence: the fifteen month figure is a default rather than a guarantee, and in an affected area the final tariffs for a charging year may not exist until the January before it starts.

Deep dives and edge cases

The main path above covers the bill most half-hourly sites receive. These come up often enough to answer and not often enough to interrupt it.

1. 15 months aheadThe charges are published

   DCUSA clause 19.1A requires written notice of the charges that will apply from a given 1 April. Embedded network operators follow one month later, at 14 months.

   →
2. 1 AprilThe charging year begins

   Every rate in the schedule takes effect at once and runs to 31 March. A bill spanning the date needs both years of rates.

   →
3. Each periodThe operator bills the supplier

   DUoS is charged to the supplier, not to you. What reaches your invoice depends on whether your contract passes it through or fixes it.

   →
4. Up to 14 monthsAggregated volumes are reconciled

   Southern states that invoices on the aggregated approach are reconciled over roughly 14 months as more accurate consumption arrives, dropping to 4 months once half-hourly settlement is in place.

Questions