Dec 26 2011

Be sure you pick the right cellular coverage for GPS tracking!

Not all companies who track their vehicles need to worry about which cellular provider they use.  Most don’t even know they may have a choice.

But if you do work in remote areas, you should pay attention and work with your GPS provider to ensure you get the right network.

I went to see ‘Mission Impossible’ today with my son and drove into town (Flagstaff) from our house which is in EXTREMELY bad AT&T coverage.

On the way home, I drove a couple miles around in our neighborhood to try to compare the reliability between AT&T and Verizon coverage.

My vehicle has both a Verizon-based device (labeled Rubicon) and an AT&T device (labeled Rubicon PNP).

We have a couple of internal mechanisms for determining how good the coverage ACTUALLY by:

  • Viewing purported AT&T or Verizon coverage
  • Showing ACTUAL cellular “lag” on a map
  • Generating a report of ACTUAL cellular “lag” with a graph

Below is an actual AT&T coverage map showing the area where I typically lose coverage in “light blue” which is not ideal coverage.  In the “dark blue” area, you can see there are very few places where the time for the device to report through the cellular network is more than a few seconds.  In my area, you can see some “tall” pins which visually represent how long it took the device to report (which is really a measure of how far the vehicle had to move before forwarding that data once it recovers cellular coverage):

AT&T coverage vs. time it took for a device to report

AT&T coverage vs. time it took for a device to report

Note that the Verizon device (Rubicon) has a “pink” line and the AT&T device (Rubicon PNP) is in white.  Verizon shows perfect coverage throughout (not pictured).

Here is how I pulled up the “lag report” version of the 3D map within GPS Insight:

Choosing both Rubicon devices once at a time

Choosing both Rubicon devices once at a time

Choose the "Lag Report" option for a 3D History Map

Choose the "Lag Report" option for a 3D History Map

[note this functionality is internal to GPS Insight support staff and is only available to end-customers upon special request]

Here is the display of my drive WITHOUT the coverage map.  Notice that there are very few “tall” pins meaning very few “lagged” points:

Slight lag (24 seconds to 3 minutes) for remote AT&T device

Slight lag (24 seconds to 3 minutes) for remote AT&T device

Next I will run a quantitative analysis of today’s data, for just the 2 devices in this vehicle, after quickly creating a “Rubicons” vehicle group containing them both:

Create a "Rubicons" Vehicle Group containing both devices

Create a "Rubicons" Vehicle Group containing both devices

I can then run this (internal use) “Lag Report” on the “Rubicons” for today:

Cellular "Lag Report" between AT&T and Verizon

Cellular "Lag Report" between AT&T and Verizon

Note there is a “landmark” option where we could restrict the report to ONLY data within a certain area we define, such as a mine, or wherever a customer may be concerned about coverage.  Also note above, that GPS Insight still thinks I am driving, since when I pull into my bad coverage and put the car in the garage, sometimes the final ignition off event doesn’t get transmitted until I drive back into coverage the next time I leave.  This is normal behavior for poor coverage areas, and is unavoidable (unless you know enough to choose Verizon when purchasing in these cases, which is the point of this article).

Here is the output, showing 100% coverage for Verizon and 95.8% coverage for AT&T for today’s drive:

Verizon 100%, AT&T 95.8%

Verizon 100%, AT&T 95.8%

The AVERAGE time to report for Verizon is 2 seconds, whereas AT&T averages 7 seconds.  The max for Verizon is 4 seconds, and for AT&T, the max times are 51 seconds and 3 minutes, for when the report reports within 1 minute or 10 minutes (we break it out into 1, 10, 30, 30+ minute “bands”).

Here is the past week, since I drove up to Flagstaff for vacation from Scottsdale, driving through notoriously bad AT&T coverage in the mountains (I know this because my iPhone is AT&T & worthless for that drive) — note that the percentage of < 1 minute reports is 100% for Verizon and only 83.7% for AT&T:

Poor AT&T Performance in mountains between Scottsdale and Flagstaff

Poor AT&T Performance in mountains between Scottsdale and Flagstaff

However, the week PRIOR to me leaving for vacation, driving around the more populated Scottsdale/Phoenix area, shows a much better 95.1% performance for AT&T vs. a 99.9% availability percentage for Verizon:

AT&T vs. Verizon coverage in Phoenix/Scottsdale

AT&T vs. Verizon coverage in Phoenix/Scottsdale

The moral of the story here is threefold:

BEFORE you purchase GPS Tracking devices for your fleet, make sure to determine A) if you have coverage issues in the areas your vehicles travel and B) whether or not you can afford to wait the minutes, hours, or sometimes overnight before the device reports in these areas.

If you may be affected, and aren’t sure, then make sure your GPS Tracking provider can provide not just multiple coverage options, but ALSO the tools to determine if there is a need for one coverage or another, as shown above.

As a point of reference, Verizon devices cost roughly $50 more than AT&T devices (due to additional modem and CDMA licensing costs).  But that’s a onetime cost and will probably amount to less than 5% over the life of the device and service.

If that $50 keeps you from experiencing occasional cellular delays due to poor (typically) AT&T coverage, then it, and finding a company which provides the right coverage (e.g GPS Insight…), is certainly worth considering.

Thanks,

Rob.

 

 

 

Tags: , , , , , , ,


Aug 16 2011

BEWARE! – Not all idling calculations are created equal!

We have been in the GPS Tracking industry for almost 7 years now. Enough to learn and FIX the limitations of GPS to ensure the highest quality data possible (e.g. 99.x% — if anyone tells you they’re 100%, well, guess what, they’re not).

So, today I found a perfect example to show the lengths to which we go to make your data 100% reliable (well, 99.x%…).

I had a 7:30 AM phone call and an 8:30 offsite meeting.

I had to get to my meeting by 7:30 so I could sit in my car and get on the call, then be there for my meeting & the breakfast prior.

So I had to idle for almost 40 minutes in order to avoid baking in my car in the hot Phoenix heat.

Here’s my stop report for 2 separate devices installed in my vehicle, both showing a ~37 minute idle stop:

GPS Insight Stop Report

GPS Insight Stop Report

One device (Rob) gets its speed data from the engine’s computer, and is more expensive because of that.  One device is less expensive but has to “interpolate” its speed from GPS Satellites traveling 9 THOUSAND miles per hour at over 12 THOUSAND miles in space.  And it’s remarkably accurate, but there is unfortunately what we call “positional GPS drift” of up to 20 feet typically.

So when the devices move 5-10 feet due to this “drift,” we interpolate a speed of 1-3 MPH typically.  But that means the device doesn’t look like it’s stationary, therefore it’s not idling.

Thankfully GPS Insight has a formula (which can be tweaked for different types of fleets, e.g. slow-moving street sweepers) which “consolidates” multiple drift points into a single idle event and position.

Our customers would never see this “inaccurate” GPS data, but here’s a picture of the REAL LOCATION REPORTS to include the drift for both the 3500 (talks to the engine for speed but not as accurate with GPS) and the 3900 (much more accurate GPS which it derives speed/distance/acceleration from):

175' of drift for the diagnostic device (we fix that)

175' of drift for the diagnostic device (we fix that)

The “drift” in the picture above is corrected over long idle stops to the “center” point which typically has the most reports.

 

175' of drift which we "correct" for diagnostic GPS device
15′ of drift for the more accurate 3900 GPS device

For the 3900, the drift is MUCH smaller — only 15′, and again, we “consolidate” that into a single 38 minute idle stop with a single “pin.”

The corrected map looks 100% accurate (well, 99.x%…):

"Fixed" stop location and idle time

"Fixed" stop locations and idle time

This shows my 2 devices in my car both stopped for ~38 minutes, and 29 feet apart (vs. the 175′ we saw above on the 3500).

And my car is 12′ long, with antennas in the front/back of the vehicle, so that’s not too bad (they show in the right locations +5′ or so each).

We consolidated the GPS drift into a single “valid” point, both in terms of position and time spent idling.

This is a HUGE distinction between GPS Insight and other companies who will either show you that your vehicle was someplace it really wasn’t, or far worse, show you that it wasn’t actually idling when it was.

Without doing all of the processing on “drift points” at 1-3 MPH, you wouldn’t know that the vehicle was actually idling, and you would lose a HUGE component of your potential ROI using GPS Insight.

This is fairly low-level, but I wanted to make sure the extent to which our product validates and consolidates data to make it actionable and insightful (and ACCURATE) isn’t lost.

There’s a big difference between this type of product and a typical “dots on a map” product.  You should know there are major differences OTHER than price when it comes to GPS Fleet Tracking.

Thanks,
Rob.

 

Tags: , , , , , , , , ,


Jul 19 2011

GPS Insight customer satisfaction survey

We recently hired a dedicated representative to proactively contact all our customers regularly.

She will:

  • Determine customer satisfaction
  • Take note of (free) training requests
  • Inform customers of new functionality
  • Review the account for where there is potential for more usage/ROI
  • Be the customer’s dedicated resource for things non-support-oriented

Her name is Shelly, and she’s at extension x8060 (877-477-4321 is our toll free #).

I was pleased with the initial results of our customer satisfaction:

Initial GPS Insight customer satisfaction survey results

Initial GPS Insight customer satisfaction survey results

Of 59 customers surveyed, 39% were Extremely Satisfied, 54% were Satisfied, and only 5% were Dissatisfied.

One customer stated they want to cancel their service.  When asked why, they stated that it was too expensive.

This particular customer has been threatening to quit unless we reduce their price since they became a customer in June of 2006.

Over 5 years later, we haven’t reduced (or raised) their price, and they’re still a customer.

In any case, over 93% of our customers asked are satisfied or extremely satisfied.

Maybe we should count that customer as satisfied, since they have remained a customer over 5 years.

That makes us 95%, which I think you’ll find is as good as it gets in this industry.

Even so, we now have someone to ask those 5% of customers why they aren’t satisfied, and I hope we can get that up to 100% by following up and fixing their complaints.

Rob.

Tags: , , , , , , ,


Jul 18 2011

GPS Insight makes the Inc. 500/5000 again!

Category: PURCHASING CONSIDERATIONSrdonat @ 4:34 pm

GPS Insight is happy to hear that we’ll be making the Inc. Magazine Inc. 500/5000 list again this year.

Last year, we made the Inc. 500 list of fastest growing privately held companies in the U.S. at #281.

This year I’ll be happy to make the Inc. 5000 list at a position around 1000-1500, since it’s based on growth, not income, and our 3 year growth was down from 1,076% last year to roughly 200% 3 year growth this year.  As even Google found out, you can only keep crazy huge growth up for so long before it begins to taper off.

GPS Insight makes the Inc. list again

GPS Insight makes the Inc. list again

Thanks to all our customers, employees, vendors, partners, and families for their help making GPS Insight the success it has become.

Rob.

[P.S., don't tell the folks at Inc. Magazine that we let this out before they told us we could...]

Tags: , , , , , , , , ,


Oct 26 2010

Why would you name your GPS device the “POO 7″ ?

Category: competition,HARDWARE TYPES,Humorrdonat @ 12:10 pm

We get a lot of inquiries from Chinese manufacturers (all of our hardware is assembled here in the United States by the way).

This one made me laugh — a lot.

poo 7

poo 7

Why in the world would you name your device the Poo 7 (or any numbered variant of the word Poo)?

I wonder how well it works…  True to its name?  I’m guessing something got lost in translation.

We won’t be selling it anytime soon…

Rob.

Tags: , , , , , ,


Oct 26 2010

Ask us to add features to GPS Insight — sometimes it happens in less than a week!

A customer recently asked us:

From: <deleted>
Sent: Thursday, October 21, 2010 11:56 AM
To: support
Subject: Location list

Is it possible to have the time the vehicle stopped at a location rather than how long it has been stopped be displayed on the vehicle location list?  For example 2:35 p.m. rather than stopped for 10 minutes.

Thank you,

Lori

We added this to the product here:

New "stop time" column option in Location dashlet

New "stop time" column option in Location dashlet

After making this change to the dashlet, it shows up like this (note that Joe turned from yellow to red since his stop is now > 1 hour):

New "stop time" column option in Location dashlet

New "stop time" column option in Location dashlet

It took less than 5 days from the time she asked for it October 21, until the time we had that feature done today:

From: <deleted>
Sent: Tuesday, October 26, 2010 10:20 AM
To: Adam Varner
Subject: RE: Location list

THANK YOU!!!  You guys are awesome.


From: Adam Varner [mailto:adam.varner@gpsinsight.com]
Sent: Tuesday, October 26, 2010 9:55 AM
To: <deleted>
Cc: support
Subject: RE: Location list

Lori,

Your request has been completed and you can now select ‘Stop Time’ in the menu of the Location dashlet.

————————————————————-

So, when you have a good idea or need the product changed just a bit for your needs — please ask.  We have a very configurable product to begin with, but our dashboard architecture allows us to add new capabilities and features/options almost instantly.  And if you are thinking about purchasing GPS Insight vs. a competitive product, please know that there is about a 0% chance that the other provider can do this for you like we do.  It’s what makes us different in the space & is a big distinguishing factor.  Try us out with a pilot & think of something for us to modify  — we do it all the time & are happy to take your thoughtful requests for enhancements.

Thanks & enjoy the new capabilities we bring you several times a week.

Rob.


Tags: , , , , ,


Aug 18 2010

GPS Insight makes the Inc. 500 List at #281

We are very happy to make the prestigious 2010 Inc. 500 list of fastest growing privately held companies in America.  Here is the link to our Company Listing.

We made the list at #281 with 3 year growth of 1,075.7% growth and 2009 revenues of $7.7 million.

GPS Insight included in this year's Inc. 500 List

GPS Insight included in this year's Inc. 500 List

My “words of wisdom” were somehow chosen as #1 on that page:

* Rob Donat will not be held responsible if you take these words of advice and drown, literally or figuratively...

* Rob Donat will not be held responsible if you take these words of advice and drown, literally or figuratively...

And, drum roll please, here is the citation:

GPS Insight in the Inc. 500 at #281

GPS Insight in the Inc. 500 at #281

We’re in good company:

Past Inc. 500 Honorees

Past Inc. 500 Honorees

Huge thanks to everyone who has helped us make this list, and more importantly, grow and thrive in this miserable economy, as we have — self-funded and privately held.

Thanks to our Customers, Employees, Partners, Families, and Friends.

Now on to trying to stay on this list for next year… (we may slip to the Inc. 5000 — 1075% 3 year growth is hard to sustain — just ask Google.)

Thanks!

Rob.

Tags: , , , , ,


Jul 25 2010

Our Odometer Readings are as accurate as you can get! (not very exciting but very important to get right…)

Category: competition,New Features,Odometer Accuracyrdonat @ 1:26 pm

That seems like I’m stating the obvious, but GPS does not equal accurate odometer readings.

Many of our competitors don’t even give you that information — just the # of miles driven.

Here at GPS Insight, we have recently released a VERY ACCURATE odometer calibration mechanism to ensure that no matter what type of vehicle or GPS Tracking device you use, your odometer readings will be about 99.8% accurate.  This is a big improvement from the typical 98% “best” we see with competitors.

If your vehicle drives 10,000 miles a quarter, we may be off 20 miles.  They may be off 200!  Even so, that’s not going to cause an engine blowout, but if you are billing or charging departments based on accurate mileage, GPS Insight will help you avoid the inevitable questions once GPS mileages are scrutinized.

A year and a half ago, we allowed our customers to enter multiple historical corrections, which is much better than the single “offset” which most companies provide.

Now we have improved this in 2 ways:

  • Offsets (corrections) are “spread out” historically to avoid spikes in mileage
  • Corrections are used going forward as a calibration to more accurately depict odometer readings so odometer corrections are much less necessary & can be done once or twice a year, vs. once or twice a month.

This was surprisingly a HUGE amount of work, which explains why none of the other providers we’ve seen have ever bothered to fix this fundamental flaw in GPS tracking devices.

Here’s why they’re not accurate, by the way:

GPS Tracking devices typically pull mileage from “GPS Interpolation” — since they know where a vehicle is at any point, they can compute the mileage between points.  Our GPS-based odometer calculations take place 4 times a second, but they are still slightly off — typically 1-3% lower than reality.  This is also because we are adding to mileage when a vehicle is in reverse, even though the odometer isn’t incrementing.

Some devices, such as our LD-3500, pull odometer readings as a function of data from the engine’s computer — but not the odometer reading itself.  Again, this is typically 1-2% off, overstating mileage.

There are all sorts of nightmarish billing and leasing problems which we’ve heard about from our customers when odometer readings aren’t 100% (or at least 99.8%) within GPS Insight, so we made these changes.  After a few weeks since the newest, now “calibrated” adjustment, my vehicle is within 1 mile of accuracy, which is partly due to the rounding on this report.  I probably won’t need to adjust the odometer again all year as a result of these changes — here is what GPS Insight thinks my odometer is for each device, when my actual odometer reading is 35,482.  By the way, some people claim GPS is MORE accurate, since tire pressure & size, as well as tire slippage can throw off an odometer.  Good luck proving that one though.  The reality is everyone goes by the odometer reading & we need to ensure we’re as close to that as possible.

off by 1 mile with 2 different types of device installed

off by 1 mile with 2 different types of device installed

Show historical odometer correction history:

Historical odometer corrections

Historical odometer corrections

Here is the new interface for viewing and editing historical odometer corrections, along with a really cool new graphing mechanism we’ll be using within the site going forward for other things:

New Odometer Graphing/Editing Interface

New Odometer Graphing/Editing Interface

So, to summarize, just know that GPS Insight is working very hard to ensure your data is as accurate as humanly possible, given technical limitations which exist with GPS Tracking devices.  By the way, the good news is that Heavy Duty Vehicles (J-1939/1708) using our 3500-HD have always had 100% accurate odometer readings since that is the only device available which gets real odometer readings from the engine.

Enough about Odometer Readings — I’m very glad this project is over — very important but not the most exciting thing in the world…

On to more interesting things now (like user-definable categories & attributes for vehicles, drivers, landmarks, users, stops, and trips)!

Rob.

Tags: , , , ,


Apr 20 2010

Why I hate Microsoft (or why resetting your Ford Sync GPS is a bad idea)

Obviously I am pretty dependent on GPS for just about everything when it comes to getting around [also for money...].

I have 3 GPS Navigation devices with me when I drive my car (factory installed, Garmin, & my iPhone) so I never print directions any more.

So, on the way out to Long Beach a couple weeks ago, my family and I took our 3 month old car with Ford’s/Microsoft’s “Sync” system (with GPS Navigation).  No directions, just our factory installed GPS navigation unit.

Just outside of town it crashed on us.  I figured out how to do a factory reset and get it back up & running.

Here’s the point:  AFTER RESETTING IT, I FORGOT TO TELL IT “FASTEST” & NOT “SHORTEST” when optimizing our route.  DOH!

We took the 10 out to LA area, but then it took me through some really sketchy, slow-moving areas.

On the way home, I thought about it, & realized I needed to change my GPS setting to FASTEST from SHORTEST.

I was curious how much longer it took me time-wise to get there than to get back, so I ran a 3D history report & saw very quickly that it was twice as long (60 vs. 30 minutes):

Run a week long history for our trip to/from Long Beach

Run a week long history for our trip to/from Long Beach

Slow, direct route vs. Fast, indirect route

Slow, direct route vs. Fast, indirect route

All I had to do is look at the 2 points where the route deviates then converges again, and compare times and distances:

Where I took a GPS dictated "dumb turn"

Where I took a GPS dictated "dumb turn"

The times/mileages are:

Going there: 17:47 & 4571.6 miles to 18:53 & 4602.5 miles

Coming back: 10:26 4618.8 miles & 10:58 & 4656.3 miles

Doing the quick math, it took 31 miles & 66 minutes there the “short” way, & 37.5 miles & 32 minutes (half as long) the “long way.”

So to save 6.5 miles, I wasted 34 minutes of my life, praying we didn’t get car-jacked.  At least if we did, I would know where the car went…

Anyway, I thought of this the other day & was curious just how much extra time it took us because of that one GPS setting on my (Factory Installed — not GPS Insight…) navigation device.

Because I track that vehicle, it took me about a minute to figure it out using GPS Insight.

Oddly enough, while I was writing this, my new Microsoft Windows 7 box crashed Google Earth as well.  It knew I was badmouthing Microsoft.  Sooner or later, all things Microsoft eventually crash.

I’m really glad we don’t run our systems on Microsoft products.

I just checked and our two “primary” servers which our customers rely upon (with lots of auxiliary and backup servers, of course) have been up for two years to two years & 3 months:

GPS Insight servers run for years without incident

GPS Insight servers run for years without incident

I’m glad most of our competitors run Microsoft though…

Rob.

Tags: , , , , ,


Dec 10 2009

Which of your vehicles has been to the crack house?

I used to use this as a far-fetched example when talking about the benefits of retroactive landmark reports:

“Let’s say you catch one of your drivers buying crack at a crack house — don’t you want to know which others may have visited there in the past year or more?”

Well, in Detroit, they actually found a city employee’s vehicle at an actual crack house.

It’s detailed in this Automotive Fleet article.

So, now that there is a real life example of this, how would you use GPS Insight to easily determine the other vehicles which have visited that same crack house?

Here’s how:

First, find the exact location by looking at that vehicle/date/time and create a landmark with the convenient link from that point (we’ll pretend my house is a crack house).

First, run a 3D history map for that day (pretend yesterday):

Use GPS Tracking to find out who's buying crack with your vehicles

Use GPS Tracking to find out who's buying crack with your vehicles

I’ll pick the “crack stop” at my house (really me coming home from taking the kids to Buffalo Wild Wings, a different kind of crack) and blur the street names in case anyone wants to come see for themselves — then I click on “Landmark: Create from Point”:

Pick a stop & create a landmark around it

Pick a stop & create a landmark around it

Now I choose a Polygon landmark, change the color to green (why not?), and outline the areas a vehicle might PARK IN (not my house, which is a common mistake — you want landmarks to be where people park, not where the actual landmark is!).

I call it “Crack House.”

Now when I refresh my menu to pick up the new landmark under “Reports: Landmarks” I can run a 1 month at a time landmark report (note clicking the month name selects the entire month):

Run a GPS Tracking landmark report on a crack house in GPS Insight

Run a GPS Tracking landmark report on a crack house in GPS Insight

Other than my wife & I, no other vehicles show up in December, so I go to November and see that a few other vehicles have been tracked in that exact area. Note the “Passing through” option which is checked — this means the visit will show up even if the ignition is not turned off while there (e.g. a drive by drug buy — my guess is crackheads like to idle too).

There was too much activity for my vehicle (with 3-4 devices), my wife’s & the company Scion (3 devices), so I created an “all but robs” group and ran the report against that:

Quickly create a vehicle group in GPS Insight

Quickly create a vehicle group in GPS Insight

So Elliot and Ryan were in the crack house zone in November:

Elliot & Ryan at the crack house?

Elliot & Ryan at the crack house?

You get the idea. Obviously this is just a simulation — Elliot was dropping off a credit card we had forgotten at a restaurant the night before, and Ryan was dropping my wife off after her car needed service.

But what is important here is GPS Insight allows you to go BACK in time to check for landmark activity.

Several competitors do NOT (including two of the largest/oldest ones in our space). They will only allow you to report on landmark activity in landmarks which you created BEFORE the activity took place.

That means you would need to know all the crack house locations in advance! I hope our customers don’t have that information handy.

Although, I’ve often said you would have to be smoking crack to go with another solution…

Rob.

Tags: , , , ,


Next Page »