Showing posts with label statistics. Show all posts
Showing posts with label statistics. Show all posts

Tuesday, December 17, 2013

Allstate Drivewise. A huge failure in potential.

I've been meaning to get this one up for some time. For a short time I had signed up for Allstate's Drivewise program. Driving very few miles as compared to most other drivers and the fact that I don't drive like an idiot, I figured it was safe (nothing to lose). The problem is the data set that they've created their perfect driver rating system around are likely based on their payouts for accidents by type of braking, time of day, mileage, and excessive speed (80+ mph). Okay sounds good so far. I kind of figured this going in.

Overwhelmingly statistics apparently aren't on my side when they're being applied by Allstate.

When you dig into the Drivewise data after receiving your first set of "grades" you'll see 4 nice looking graphics in the interface. One for mileage, one for braking, one for time of day, one for speed in excess of 80mph.

Mileage

They go on to tell you that the mileage is a calculated projection of how many miles they think you'll drive based on your daily driving. If you're sticking to what they expect this shouldn't be a problem. I don't have a problem with mileage from the device because it coincides with the mileage from my odometer (which they already had on file). Spoiler, if you tell them you only drive 7,000 miles a year, and you really drive 50,000 the program will not give you a discount and the agent will have access to your actual mileage and will likely raise your premiums accordingly.

Braking

The braking section of the graphics show two things. "Hard Braking" and "Extreme Braking" are the two categories. According to Drivewise, hard braking is when you decelerate by 8mph in less than 1 second. If you're following a bus that makes frequent stops and you can not change lanes, depending on the bus driver's performance and lack of indication you will have a hard braking event (or two or four), someone walks out in front of you, a dog in the road, you get the idea. Extreme braking is when you decelerate by 10mph in less than 1 second. So if you come up to a short traffic light that has a 3 second yellow (these do exist) from 50mph and begin decelerating, you will likely have an extreme braking event. When you have 4 hard braking events and 1 extreme braking event over the course of 3 weeks this erodes any discount you would expect to receive from the program. I do mean ANY and ALL discounts.

Time of day

The time of day expectations for the program are really optimistic for Allstate. They have 4 categories. They've said that the "Lowest Risk" for accidents is on weekends between 5am and 11pm. The same time that most teens are out driving to work or shopping, etc. The "Low Risk" time is from 4am-12pm on weekdays (When teens are driving to school.). "Moderate Risk" is from 12pm to 11pm on Weekdays (When teens are on their way home from school or on their way to work at night). If you're out past 11pm you are driving at the "High Risk" time which is from 11pm-4am on weekdays and from 11pm-5am on weekends (Drunk people dodging, but luckily most teens are at home curiously enough).

Speed >=80

This one is pretty straight forward however they give a whole range of grading here where you basically stay below 80 or you don't. I don't understand this one at all because if you go above 80 you should be in the very high-risk category for drivers. Go get a racing license and take it out on the track. Now if you're in a state like Florida or Montana where you may encounter a 75mph speed limit, then it's understandable that this may need to be changed, but they're not trying to hide anything from you here.

My Suggestions

If Allstate really wanted their Drivewise program to be highly successful for them and to actually reward people who are definitely driving safely they would look at a different set of parameters.

Speed

Since the device already knows how fast the driver is going, it should be able to tell whether they are one of those people who can't maintain a constant speed. If the driver accelerates extremely rapidly (0-60 in 10 seconds) then they should have a record of drag racing on file. This could be road rage (extremely risky) or someone not paying attention to their lane ending... this might also obviously be drag racing, but the risks are the same. If the driver is running 65 and catches someone doing 45 and does not overtake or switch lanes, then they are not paying attention. Also if they decelerate by this much and maintain the speed it means that they either entered a construction zone, or they slowed down to the flow of traffic. If the device sees people accelerating and decelerating regularly it should know that the person is in a stop and start traffic jam. It already knows the time of day, so if the person is in rush hour stop and start traffic then it should know and place them in a higher risk category (for a low impact collision).

Location vs Speed

The device already has the ability to track vehicle location because it's transmitting on a cellular signal. If Google and most GPS systems can tell how long it will take to get to a destination, the device should be able to do this as well. This means if the driver is speeding and the device knows it, then they are risky and should not receive the discount. Something like 65 in-town in a school zone and they should put the driver into the high-risk category as well (for a high impact collision).

Crazy Driving

Add a couple of accelerometers to the device and now you can actually find the people who are weaving at risky times (drunks) and the people who are weaving on their daily commute (food-eating, texting, doing their make-up, you know... people who are exhibiting risky behavior). Also you can find the people who are insanely driving and weaving in and out of traffic with fast bursts of acceleration. Like the yellow semi truck that didn't like my Chicago Black Hawks tag.

Time of day

The Allstate Drivewise program needs to get a realistic idea of when people drive and when people don't drive in order to be successful. If I'm driving safely at a time of day when there is nobody on the road but me, then I shouldn't be in the high-risk category. If I'm driving when there are fewer people who are on the road like an afternoon after rush hour, I shouldn't be in a moderate risk category. If I'm driving when EVERYONE is off of work at the same time as they are on a Weekend, then I should be in a very high risk category (more people on the road=greater chance of an accident). If I'm driving when people are trying to get to work on time or they're trying to rush home after a bad day at work those are risky times as well.

I think if Allstate had actually taken the time to utilize the system instead of cutting corners, they could actually reward the people who are indeed safe drivers and profit from the people who aren't.

Suggested Upgrades

Add a couple of wireless cameras to the device. Let's put one in the front and in the back. Let's actually get some documentation on why someone is stopping abruptly. Don't outsource the research to a country overseas. Now you can have people in the US work from home and analyze the footage. It would help keep people off of the streets and off of the roads and it would also help with those fender benders that don't get reported. Not to mention auto theft, erratic driving and whether someone's towing a trailer at high speed. Yeah I'm talking about the people in the fast lane running 80mph towing the trailer that's rated for 45mph max.

Make the device aware on its surroundings. Add a hygrometer. Let's see if people are driving in the rain or driving when it's dry. Let's take some barometric pressure readings on the Drivewise device. Zero visibility thunderstorm, do you slow down? They should know. Let's add a thermometer. Driving on ice? The device should know. If you drive excellent on ice, then you should be rewarded. If you're more like a skating star doing twirls, whirls, and 720 degree spins, you should be penalized.

Make it driver aware, add something to the keychain so when a certain driver is in a closer proximity it knows who is driving the car. Sure you could swap keys, but this would definitely help if you had teens driving the car. That way they could tell who was a safe driver and who wasn't. Want the discount back, don't let junior drive your car.

All-in-all I'm 100% positive about making the roads a safer place.

What is the Allstate Drivewise really about?

The Allstate Drivewise device is not out to make the roads safer. In actuality if we look at the device from a completely different approach, it's a now a gimmick that invades the privacy of the driver. Allstate isn't interested in whether someone is a safe driver. They're interested in finding ways to make you pay high premiums. The higher risk they can make you in their book, the better off they are (monetarily). My agent seemed disappointed that I was healthy when applying for Life Insurance... gee, I wonder why that is? The same logic applies to car insurance. If you're a truly bad driver, the system will punish you, but if you're a good driver, then it's up to the insurance company to make up for the loss.

Realistically if we look at the stats from the US National Transportation Safety Board, most people aren't at risk of getting into a major accident on the road statistically. Only the select few. If we can keep those people off of the street, then sign me up. Until then, I'm keeping the Allstate Drivewise out of my vehicle because it makes me think about something when I'm behind the wheel that isn't related to my driving performance at all, and that's whether or not I'm going to be financially penalized about something out of my control. And when they do penalize me for something which is not a risk at all, I appear to them to be an "unsafe" driver, which helps them to justify charging more.

Get Wise Allstate.


A note on the edits
Originally I had mentioned that it might have been up to the device programmers, but that's not really fair. Once a product like this passes enough scrutiny panels in the production phase, good intentions of the designers are left in a pile for the sake of a little bit of savings. As long as the device gives a plausible illusion of savings, then the company will proceed.

Thursday, December 1, 2011

Holiday Shopping Help - How to get the good deals.

Here's a cheat sheet for doing the math. In our household we do a lot of shopping (looking for the best deals). I have the ability to memorize prices items, quantities, and container packaging. Because of this I've noticed a bit of trickery. Around holiday time people are desperate for deals and sales and will unknowingly purchase things at a much, much higher price.  Some companies increase the prices a month or two in advance (in regard to Christmas, this can be as early as late September) to offer the item for the original price they are claiming is a new sale price. Here are a few of the phrases that companies uses that aren't always intuitive, but seem like a good deal.

Buy 1 Get 1 Free = 50% off of two.
You'll want to watch this, usually they've increased the price by double which makes it so you only get 25% off in the end. Not bad? Well, considering the items they do this to have a huge mark-up it's rare you'll get your money's worth.

Buy 1 Get 1 for $1 = 50% off of both plus $1.
Same as above, they've already increased the price astronomically and people are typically preoccupied with the plus $1 to think about what the original price was.

Buy 3 Get 2 free = 40% off of each. You're still paying 60% of the price for each. If it seems reasonable, then go for it.

Buy 1 Get 1 for 50% off = 25% off both.
Not a real deal if the place you're shopping is 25% higher than the competitor.

Be sure if you're going for a sale not to get lured into purchasing other items you might normally buy if they are going for a higher price. High sales also mean they have to make-up the cost, so that savings is spread around to other items in the store. Statistically, if you buy soda, and then Tuna Fish, Pizza or Toilet Paper, one of the items might be on sale, whereas the others will carry an inflated price.

It's a good idea to always look around for the best price before going out to shop.

Another trick I've noticed is varying quantity amounts. Some products (eg. Bullion Cubes) might be sold in 5 or 6 different quantity sizes at various stores. When you visit one store for price comparison, the price may seem lower, but you're getting a considerably less amount. Each package might seem like a scaled down version of the other and since the side-by-side counterpart offerings aren't available you might not be aware of the change in quantity. Sometimes this is a visual trick because although the front of the item is the same size and shape, the depth may have changed.

Another one of the tricks that companies do is when they change the package they've usually changed the quantity is some way or another. In regard to toilet paper, if there is a raised pattern on the paper this will cause less paper per roll. Toilet paper companies increase the diameter of the tube inside of the roll (to give you less paper), they change the sheet count so you get fewer sheets per roll in different arrangements, meaning if you buy 24 rolls you may be getting fewer sheets per roll than you do if you buy three 8 packs. Also look out for another change a shorter tube where they take up to an inch and a quarter off of one end of the roll.

Soda is bad for you and this year, they've introduced a new 20 pack instead of a 24 pack. They're still charging the inflated prices for the soda to begin with but it's harder for people to make the comparison. They also don't use the same measurements for comparison from one product line to the other (2 liters is 67 US Ounces) If you're comparing a 12-pack of soda (144 ounces) vs two 2-liters (134 ounces), then you'll get one fewer can. If the price for the two liter is 11/12ths (Price x .92) the price of the 12 pack then it's a deal. I always try to break down the price per ounce for a comparison in my head, but most phones have calculators as well.

If you're looking at the two prices for the comparison, consider the amount of gas and time it takes to shop at both places. If there is very little difference in price, then it might not be worth it to do a lot of comparison shopping.

The work around, buy items all year round. If you know you have special occasions to buy for, pick up the items on clearance after the holidays and put them in a safe place. If you are running a little behind, start shopping around August for December, you'll be well ahead of the price hikes. Also don't buy the latest technology Zero-day. Companies will usually drop the prices back down around May or June, so if you can live without it, you can save up to 25% of the cost in some cases.