For energy suppliers and balance responsible parties with smart meter customers

    Forecasting for smart meter portfolios AI aggregation and load forecasts that are 30% more accurate than standard load profiles

    As smart meters roll out, more and more customers are settled on their measured consumption instead of a standard load profile – and the volume risk moves to you. alitiq uses AI to continuously group your smart meters by similar behaviour, forecasts each group with weather and calendar data and delivers the sum as the forecast for procurement, scheduling and imbalance management.

    than standard load profiles
    30% more accurate
    AI aggregation keeps adapting the groups
    Continuously
    resolution, from intraday to year-ahead
    15 minutes

    The volume risk moves

    With smart meters, the profile turns into a volume risk

    As long as a customer is settled on a standard load profile, the profile counts – not their actual consumption. Smart meters change that: once a customer is settled on measured interval data, their real consumption enters your balance, and every deviation from what you procured becomes an imbalance you pay for.

    Profile-based

    Settlement on a standard load profile

    1. 1Consumption is represented by a standard load profile.
    2. 2You procure and schedule according to the profile.
    3. 3Differences to the actual consumption are reconciled later, usually at averaged prices.

    The quarter-hour risk is largely absorbed by the profile, not by your balance.

    With smart meters

    Settlement on measured data

    1. 1The smart meter records the real consumption every quarter hour.
    2. 2Once the customer is settled on interval data, these values go straight into your balance.
    3. 3Every deviation between procurement and consumption is charged at the imbalance price – often exactly when it is high.

    The volume risk is yours. If you keep procuring by profile, you pay the difference.

    Why standard load profiles miss smart meter customers

    Smart meters often go first to the customers whose consumption least resembles an average profile: high consumption, rooftop solar, a heat pump or an electric vehicle.

    • PV self-consumption

      Grid consumption drops at midday – but not on cloudy days. The profile does not know the weather.

    • Heat pumps

      Consumption follows the outdoor temperature and is well above the profile on cold days.

    • E-mobility

      Wallboxes create evening peaks that no household profile represents.

    • Home batteries

      Batteries shift grid consumption and feed-in across the day.

    • Dynamic tariffs

      Customers shift their consumption to cheap hours – depending on the exchange price.

    What alitiq takes care of

    From the single smart meter to the portfolio forecast

    alitiq processes the measurements of your smart meters fully automatically – from data import and AI aggregation to the finished forecast for your portfolio.

    Smart metersForecast per clusterPortfolio15-minute values from the smart meterPV prosumersHeat pumpsHomes & businessE-mobilityPortfolio forecastalitiqSLP≈ 30% less deviation than the load profile
    Schematic illustration. The AI finds the clusters in the data; the names are for illustration.

    Example

    Standard load profile vs. alitiq on two typical days

    A portfolio of around 5,000 smart meter customers: on a sunny spring day, PV self-consumption pushes the midday load below the profile; on a cold winter day, heat pumps lift it above. The alitiq forecast follows the actual behaviour. Hover over the curves to see the values.

    Show deviation:
    • Measured load (smart meters)
    • Standard load profile (SLP)
    • alitiq forecast
    • SLP deviation
    01234MW00:0003:0006:0009:0012:0015:0018:0021:0000:00
    Illustrative example with synthetic data, derived from typical behaviour of smart meter customers. A benchmark with your own measurements shows your result.
    Show values as a table (hourly)
    TimeActualSLPalitiq
    01:001.181.231.25
    02:001.071.091.13
    03:000.990.960.95
    04:000.950.950.84
    05:001.291.141.08
    06:001.711.581.74
    07:002.242.142.35
    08:002.292.322.21
    09:001.892.081.84
    10:001.661.971.64
    11:001.552.131.86
    12:001.582.312.11
    13:001.572.352.16
    14:001.532.211.94
    15:001.342.041.77
    16:001.622.031.94
    17:002.012.302.42
    18:002.712.782.92
    19:003.293.203.38
    20:003.343.253.33
    21:003.182.873.25
    22:002.752.302.48
    23:002.041.811.86
    00:001.521.521.70

    Deviation with SLP on this day

    6.5 MWh

    Deviation with alitiq

    4.5 MWh

    Less deviation with alitiq

    −32%

    Savings calculator

    What does the volume risk cost you – with standard load profiles and with alitiq?

    Estimate what imbalance costs and short-term re-procurement cost your smart meter portfolio per year if you procure by standard load profile – and how much of it an alitiq forecast saves.

    • Always compared: standard load profile vs. alitiq
    • Transparent assumptions you can adjust
    • Benchmark with your own measurements on request
    5,000meters
    500100,000

    Early smart meter customers often have high consumption, rooftop solar, a heat pump or an EV.

    Adjust assumptions

    Deviation = mean absolute quarter-hour deviation as a percentage of the annual volume. Cost per MWh = average premium of imbalance settlement or short-term re-procurement over the procurement price. The customer structure sets a typical load profile error, which you can override here.

    Estimated savings per year

    €113,400

    ≈ €22.68 per meter and year · €3.78 per MWh

    Cost of the volume risk per year

    Procurement by load profile€378,000

    Deviation: 5,400 MWh per year

    Procurement with alitiq forecast€264,600

    Deviation: 3,780 MWh per year

    Balanced annual volume: 30 GWh
    Request a benchmark with your data

    Rough estimate based on the stated assumptions, not an offer. The actual result depends on the portfolio, market prices and data quality.

    What you gain with alitiq

    A forecasting platform that grows with your smart meter rollout.

    30% more accurate than load profiles

    Forecasts from your real measurements instead of average profiles reduce deviations and imbalance costs.

    Aggregation that keeps learning

    New meters and changing customer behaviour flow into the clusters automatically. No customer groups to maintain by hand.

    All time horizons

    Intraday, day-ahead, week-ahead and year-ahead – for schedules, spot and forward procurement from one source.

    Scales with the rollout

    Data processing is built for millions of metering points – from your first smart meters to the complete rollout.

    Easy integration

    REST API, Python SDK or file exchange: the forecasts flow into your meter data or portfolio management system.

    Co-creation

    Help shape the platform – from requirements for smart meter data to the integration into your processes.

    Related: Load forecast · Reduce balancing costs · Technical documentation

    Frequently asked questions on smart meter forecasting

    What changes when customers get smart meters?

    Customers on a standard load profile are settled on the profile, and differences are reconciled later. Once smart meter customers are settled on their measured interval data, their actual consumption enters your balance. Every deviation between procurement and consumption is then charged at the imbalance price – the volume risk lies with the supplier.

    When are smart meter customers settled on measured data?

    That depends on the market rules and often on the local grid operator. In Germany, for example, the grid operator decides from when it settles smart meter customers on quarter-hour values and takes them out of the profile-based reconciliation. Customers in different grid areas can therefore switch at different times. Together with you, we clarify which metering points are already settled on measured data and forecast exactly that part of your portfolio.

    What does AI-based aggregation mean?

    Instead of grouping customers rigidly by profile type, the AI looks for groups of meters with similar consumption behaviour, for example by PV self-consumption, heat pump or charging behaviour. Each group gets its own forecasting model. The groups are recalculated continuously so that new meters and changing behaviour are taken into account.

    What does 30% more accurate refer to?

    To the mean absolute deviation of the forecast from the measured consumption in quarter-hour resolution, compared with the standard load profile for the same portfolio. How large the advantage is for you, we show in a benchmark with your own measurements.

    Which data does alitiq need?

    For each metering point, an ID, an approximate location for the weather data and the historical quarter-hour values are enough. Measurements sent continuously keep aggregation and forecast up to date. We do not need personal customer data such as names.

    Is it worth it for small smart meter portfolios too?

    Small portfolios in particular fluctuate strongly and deviate clearly from the profile, because individual effects hardly cancel out. Every additional meter from the rollout enlarges the data basis for aggregation. The savings calculator on this page estimates what a better forecast is worth for your portfolio.

    What about customers without smart meters?

    alitiq also forecasts profile-based and interval-metered customers as well as gas and heat demand – so you get a forecast for your entire portfolio from one platform.

    How accurate would alitiq be for your smart meter portfolio?

    In a forecast benchmark, we compare the alitiq forecast with the standard load profile – based on your own measurements. You see how much volume risk you save.