Scottish Power billing

50Twuncle

I am paying SP £33.07 per month dual fuel - and managed to build up a £76 credit over the last 12 months
So - I reduced the payment to £30 - which raised an automated email from SP - informing me that if I did this - there was a likelihood that at the end of the next 12 months - I would be £85 in debit !!
OK - How ?

backfoot

Ring them up and ask. They will do a call back if busy.:)

[Deleted User]

50Twuncle wrote: »

I am paying SP £33.07 per month dual fuel - and managed to build up a £76 credit over the last 12 months
So - I reduced the payment to £30 - which raised an automated email from SP - informing me that if I did this - there was a likelihood that at the end of the next 12 months - I would be £85 in debit !!
OK - How ?

We are not yet out of the cold weather, and we have had the warmest winter since records began so you cannot judge this winter on the future ones. I ve not forgot the winter about 4 years ago where it never let up from ther end of November ( thick snow ) to mid April with cold Easterly/North Easterly air flow for four months. Suppliers will not be using such a short period as one years ultra warm winter weather as a norm.

Shrimply

Why? Because a group of primary school kids could probably come up with a better predictive model than the energy companies seem to use.

I've been with Eon since October and by paying the set direct debit (the first payment of which wasn't until the end of November) I've never yet been out of credit. Yet they still recommend I need to increase my direct debit by over 20% to break even for my yearly review. Totally bonkers.

Cardew

Shrimply wrote: »

Why? Because a group of primary school kids could probably come up with a better predictive model than the energy companies seem to use.

I've been with Eon since October and by paying the set direct debit (the first payment of which wasn't until the end of November) I've never yet been out of credit. Yet they still recommend I need to increase my direct debit by over 20% to break even for my yearly review. Totally bonkers.

As stated in post#3 we have had a mild winter and thus not representative.


The algorithm for all company's 'predictive' model is apparently approved and audited by Ofgem and has input from theMet Office.


I suspect another reason is if the DD is set too low and has to be raised dramatically before the review date, there will be howls of protest like those posted on MSE!

Shrimply

Cardew wrote: »

As stated in post#3 we have had a mild winter and thus not representative.

But they aren't predicting how much I'd pay in a really cold year, they are predicting how much I need to pay for this year, when the winter has almost passed.

I believe a lot of people are probably being paid a lot of money to come up with complex "algorithms" whereas you would get better predictions using a simple fast model for each individual user in a given year.

[Deleted User]

What if the OP was on a fixed tariff which is coming to an end soon?

In that case, SP will assume switching to a standard tariff which would be more expensive and hence allow for a shortfall in 12 months time.

Lack of sufficient information from the OP to allow for any accurate replies.

molerat

£33.07 dual fuel !! That is around the average for electricity in a dual fuel house.

Cardew

Shrimply wrote: »

But they aren't predicting how much I'd pay in a really cold year, they are predicting how much I need to pay for this year, when the winter has almost passed.

I believe a lot of people are probably being paid a lot of money to come up with complex "algorithms" whereas you would get better predictions using a simple fast model for each individual user in a given year.

As you state -

I have been with Eon since October and by paying the set direct debit (the first payment of which wasn't until the end of November.

That is hardly much data to feed in to an automated process. So unless we have 20 million bespoke predictions a year;) You will get the nonsense output like yours - untouched by human hand!

Also people change companies more and more these days; so less data to feed in to the computer.

Former_E.ON_Company_Representative:_Malc

Shrimply wrote: »

Why? Because a group of primary school kids could probably come up with a better predictive model than the energy companies seem to use.

I've been with Eon since October and by paying the set direct debit (the first payment of which wasn't until the end of November) I've never yet been out of credit. Yet they still recommend I need to increase my direct debit by over 20% to break even for my yearly review. Totally bonkers.

Shrimply wrote: »

But they aren't predicting how much I'd pay in a really cold year, they are predicting how much I need to pay for this year, when the winter has almost passed.

I believe a lot of people are probably being paid a lot of money to come up with complex "algorithms" whereas you would get better predictions using a simple fast model for each individual user in a given year.

Hi Shrimply

Thought it might help if I let you have a bit of information about our Monthly Direct Debits as well as details of an online tool to give you more control over the arrangement.

Payments are based on how much energy we think you'll need up to the annual review and current prices. Arrangements are spread over a full 12 months and aim to achieve as close as possible to a zero balance by the annual review. The usage is worked out based on past history. As you've not been a customer long, we'll use information given us by the energy distributors when you switched to us. The longer you're with us the more we'll rely on your readings.

Predicting usage can never be an exact science as what happened before may not be repeated going forward. Even consistent patterns can be thrown out by sudden spikes/troughs. This is why we review arrangements quarterly and have introduced our online tool the 'Direct Debit Manager.' This lets you amend payments to better reflect changing circumstances. The aim is to stop too much credit/debit building up. If it does, we've built in a safeguard so, at the annual review, any credit over a fiver is automatically refunded and any debt is included in the ongoing payments and spread up to the next review.

As you've only been with us since last October, I suspect part of the past usage is based on last year's estimated consumption as given us by the distributors. If this was higher than now, it's probably led to the higher suggested payment as it's anticipating you'll use more than payments are currently covering even though there's credit on the account. Lower usage during the spring/summer will hopefully put things back on track to something like what you anticipated when you switched. Once we've actual readings for a full 12 months, the calculation should even out provided, that is, there are no significant changes.

If there are changes of circumstance, you can use the 'Direct Debit Manager' I mentioned above to take more control. Provided the account is billed up to the latest meter readings, payments can be changed by up to 20 per cent up or down. By up to 5 per cent up or down without readings. Before confirming the new amount, we'll let you know by how much we anticipate the account will be in credit/debit by the time of the annual review if payments are changed but usage doesn't alter as expected.

I'm guessing a lot Shrimply and sorry if my assumptions are off beam but hope this gives a bit of an insight into how we set up and look after these arrangements.

Malc

Shrimply

But you are getting a bespoke prediction from, who knows how many, variables that they feed into their very complex algorithm.

And data quality is much more important that data quantity. You wouldn't need that much to fairly accurately model gas usage on temperature, which really is probably all that really matters. Especially as smart meters are being installed you'd very quickly get a decent amount of data, to create a model that allowed more accurate predictions at the individual level. That is my humble opinion in any case.

Of course a model will never be 100% accurate, but even with 5 data points and simple linear regression of my monthly usage on average monthly temperature, which would run in the time it takes to load the webpage. I'd be able to make more accurate predictions than E.ON about my gas usage for the rest of the year, it certainly wouldn't predict an increase And the prediction would only become more accurate for each reading I submitted.