Accidental non-disclosure on car insurance

[Pingu]

first some background

different insurers have different risk profiles. some just want low risk unlikely to make a claim type customers. Other are prepared to take on more risk. so the ones that want just low risk customers will load the fuck out of anyone who they feel doesnt match their model of what makes a ow risk customer. this way they deter them from insuring with them and if they do the premium is so high then they can justify it.

each insurance company will have data that gives them the ability to build various models to predict the likelyhood of a person claiming. Age, area in which you live, type of car, cost of repairing that car if it should be in an accident etc.

I am assuming none of the above is news to most people but just in case...


data can be analysed to fuck. If you have the right tools and data you can predict the influencers on pretty much anything. You will never be 100% accurate but the odds of you being correct will increase significantly.

for example. An austrailian bank (The National Austrailia Bank) used its customer and bought in data during the launch of a new credit card a few years ago (2003). They used a database engine and analytical tool that allowed them to look at each customer individually and then determine the most likely method of approach that would work for that individual. Firstly they profiled the prospect list to see who they wanted as customers, so if you were a high risk of defaulting you never made the final prospect list.

They then used various methods of contacting the prospects ranging from emails to personal visits. They used an incentive that had been tailored for that individual in order to increase the likelihood that they would take up the new credit card. This ranged form introductory interest rates though to vouchers for money off stuff that they knew the person would be likely to buy.

The number of people taking up the offer was 25 times higher than for any other campaign they had run.


Since then the technology used has moved on and the tools they used are in use at at least 5 UK insurance companies.

So they do have the means to do an individual risk profile should they want to. What they actually do is to group similar people together based upon profiles etc that are determined by some very bright people (actuaries etc) and these profiles/risk models are then loaded into the computer.

when you lob in your data its compares your data against these models and then computer determines how much the baseline premium should be loaded based upon how you fit into one of these models. Or if they will even offer you cover based upon that companys risk model


there is more and more data being collected about us and its only a matter of time before each of us has tehir own profile on some of these systems. The opportunities for cross marketing products etc makes this sort of thing very attractive to thsoe with the power and money to do the analysis


for anyone really interested the technologies we use to enable individual customer segmentation are:

Teradata (database)
Teradata relationship manager (CRM tool)
SAS (whizzy analytical tool)
KXEN (predictive analytic suite0
Ab-Initio (ETL tool for moving and doing stuff to data - used in loading data from external systems)
Goldengate (Same as above)

 

[noodles]

data can be analysed to fuck. If you have the right tools and data you can predict the influencers on pretty much anything. You will never be 100% accurate but the odds of you being correct will increase significantly.

for example. An austrailian bank (The National Austrailia Bank) used its customer and bought in data during the launch of a new credit card a few years ago (2003). They used a database engine and analytical tool that allowed them to look at each customer individually and then determine the most likely method of approach that would work for that individual. Firstly they profiled the prospect list to see who they wanted as customers, so if you were a high risk of defaulting you never made the final prospect list.

They then used various methods of contacting the prospects ranging from emails to personal visits. They used an incentive that had been tailored for that individual in order to increase the likelihood that they would take up the new credit card. This ranged form introductory interest rates though to vouchers for money off stuff that they knew the person would be likely to buy.

Sorry, I'm not sure how the The National Australia Bank example supposedly differs from how motor insurers currently operate, in terms of risk assessment?

The initial process that you mention is more a marketing function than a risk assessment. Insurers action their marketing techniques along the similar lines, although a few insurers do not seem to join-up the thinking of their marketing and underwriting divisions to the same extent.

The risk analysis element that you describe is the same as how insurers operate.

 

[Pingu]

Sorry, I'm not sure how the The National Australia Bank example supposedly differs from how motor insurers currently operate, in terms of risk assessment?

The initial process that you mention is more a marketing function than a risk assessment. Insurers action their marketing techniques along the similar lines, although a few insurers do not seem to join-up the thinking of their marketing and underwriting divisions to the same extent.

The risk analysis element that you describe is the same as how insurers operate.

what I was demonstrating is that the computing power and applications are already there to produce individual profiles and have been for a while. the austrainail ban one is different to the norm because they actually risk scored and profiled each individual customer rather than grouping certin similar types together. OK so the end result was that some would get the same offers etc but rather than just making that offer to ABC1s living in a certain area etc they applied their models to each prospect as an individual using their personal sepnding habits, and other information individual to that person to determine the offer

Lloyds TSB do something similar on their spending and income customer analysis for certain customers but they also attempt to categorise spend outside of the LTSB group (i.e if you have a nat west mortgage they will identify this from your transaction data and then see if its worth marketing a ltsb product to you to tempt you away from nat west)

to use an insurance example (though this isnt being done yet as far as I know). You apply for insurance. your normal details create a base price but then your spending habits are analysed to modify that price. Is tehir a pattern of spending at garages that may indicate that you keep your car well maintained, are their loads of purchases from NOBTOYSFORBOYZ that may indicate you are into modding cars. Is their a lot of fuel purchases from rural petrol stations - could indicate that you drive mainly out of cities or a lot of purchases from motorway service stations.

al the above is unique to you and could be used to derive a modified risk profile for you as an individual.

in LTSBs example the querying of the additional data when someone who its attached to makes an application etc takes about 0.25 seconds so its quite possible that should insurers also have access to this sort of data that it can be used to amend your risk profile and therefore your premium\terms of insurance.

 

[bigbry]

Agreed, but the OP said the mother was put on "to make it cheaper" which implies she was put on as the main driver. Otherwise I can't see any benefit to putting her on.

Putting a female on an insurance policy can make it cheaper - my 22 year old stepson put his 21 year old girlfriend on his policy when he renewed ( so they could share the driving on days out or holidays) and it reduced his premuim - I don't know by how much as I've forgotten what he told me but they quoted him a price and when he asked about putting his girl friend on they reduced the premium. It's because female drivers are judged a better risk.

 

[Fez909]

Fucking hell! You ask for some advice and everyone assumes the worst!

My friend is the main driver on the policy. His Mam is a named driver. She can, and does, drive. But not his car! This has been disclosed to the insurance company. This is not fraud.

The non-disclosure came when she got a speeding ticket in her own car and her son didn't know/forgot to tell them - I'm not sure which it was.

And yep, the policy does get cheaper in this way, as my girlfriend is currently insuring our car and asked me to pay half. I couldn't afford it so she was going to get insurance on her own, but it is more expensive!

So can you please all fuck off with the fraud allegations.

 

[jimadore]

If you dont' tell them they wont' pay ,out but if you do tell them you will pay extra? Unless it is full comp not worth the paper its printed on after all they little claw backs. keep paying the crooks? NOT ME!!!