Video Analytics Usage Statistics 2016

By: IPVM Team, Published on Feb 25, 2016

Our previous video usage analytics statistics (in 2014) showed that integrator use of video analytics was flat from 2011 to 2014, a clear disappointment given that video analytics continues to be voted as the next big thing.

Now, we have run the same statistics again plus added an open ended question to gather more insights into the decision making process:

"What impacts your decision to deploy or avoid video analytics?"

150 integrators responded, and the responses show three key patterns:

Skepticism is still the biggest enemy of video analytics usage
Poor performance in past keeps integrators hesitant to proactively recommend analytics
Manufacturers need to do better at marketing / positioning analytics

Statistics

*******, ******* **** ** integrators ****** ** ***** use ***** *********, ***** is * ***** **** percentage *********** ***** *********** all ****** ** ******* and *****. ******, **** 1 ** * ***** deploy ***** *********, ** shown ** *** ***** below:

*******, ** *** ******** side, ***** ** *******. While ***** *** **** to **** ******* **** to ****, * ******* jump ** ***** ** the *** **** ******* compared ** ******** ****:

Skepticism ** ********* *******

****** ********* **** **** in **** ***** ********* are *** **** ******** unless *** ******** ************ asks *** **** **** of ** ******* ** feature. **** *** * few ************ ****:

"* ***'* ***** **** are ***** *** ***** time. *** **** ********* that ****** *** ******** of ****."
"** ** ****** ** the ***** ** *** RFP"
"********* **** *** *** a ****, *** * do *** *** *** benefit"
"*** ***** ***, *********, not ************, ***** ** resources ** **** *** support *** *** *********"
"***'* ***** ** ***** well ****** ***"

******* ** *** ***** of ********* *** **** bringing ** *** *********** of ***** ***** *********, several ********* ***** ******* common ********* **** *****: cost *** ***** *****.

"**** ******** **** ********* by ******** ** ** bid *****. ***** *** unreliable ** ****** *** typical ***** ************."
"**** ** ****. ****** detection *** *** ********** usually **** *** ****."
"***** - ** *** client ******* ** *** for **."
"**** ** ***** ****** and ********** ** *** up. **** ******* ***** use **** ** ******* them. *** **** ********* they ******* **** **** to ******* **** *** to ***** ****** *** notifications."

Poor *********** ** ***** *** ***********

***** ********* **** ************* been ****** ** ******, both ** ***** ** the ***** ***** *** the ********* ********** (** the **** ** **** expensive ******* **** ******** analytics, ********* ******** ** processors, ** ********-***** ********), in ***** ** ***** and ************* **** *** the ********** ** ********, and **** *********** ** the **** ** ******** excessive ******* ** ***** alarms. ** ***** *** analytics ** ** ***** worthwhile, ***** ***** ** be * ***** **** savings ** ******* *******. We ******** ******* ********* that **** ** ******** that *** **** ********** deployments *** **** ** some **** ** ********* scenario. ** ***** ***** the ********* *** **** as * **** ** an ******** ******* *****, alerting **-**** ** ****** guards ** **** ****.

***** *** ******** ******, the ****** ***** ** be ********, **** ***** alarms **** ** * minimum ** ***** *** system ***** ******** ** ignored. ******** ** **** anti-analytics ********* ***** ***** lines ********:

"**** ** ***** ****** and ********** ** *** up. **** ******* ***** use **** ** ******* them. *** **** ********* they ******* **** **** to ******* **** *** to ***** ****** *** notifications."
"******* ** *** **** of ******** *** ***** of *** ******. ********* infrastructure *** ************ ** limitations ** ****. *********** of ********** *** ***** alarms."
"**** ***** ************ *** accuracy ** ******** ******. if *** **** ** too *********** *** ******** a *** ** ***** positives, ** ** *** recommend ***** *********, *******, do **** ********* ****** anomaly ************** ********."
"** *** ** ***** video ********, ******* ** lots ** ***** ********."
"*** *** ****** ***** Poor *********** ** ********* in *** ******* **** we *** (****)"

There *** ******* ******* ******

** *** *********** *** be ********, ****** ** product ********* ** ** ensuring *** ***** *** expectations ** *** ******** are ***** ********** ***** are ******* ******** ** integrators ********* ********* ********* and ****** ******* **** video ********* ** *******. We ******** ******* *********** responses **** **** ******* in ********** *** *****.

"************ ** *** **** cost *** ***********/********. ** see ********* ********* ********* more ********* ** ****** testing *** ****** * more ******/******** ******."
"** ****** **** *** them, ** *** *** cost ** * ****** with ***** ** ******** or * ***** ***** analytics ****** ** ********* in *** ***."
"**** **** *****'* *** exiting. ******* ******** ***** (server *****, ****** *****, parts ** **** *****, main ***** ***** ***** hours. ******** ***** ***** are ********** **** ********* or ****** ******* ******. Also ** *********** ** fence ********* *******."
"** *** ********* ** part ** *** ******* central ******* ********** ********"
"***** ** ** ***** personnel (***** ** *******) of ********* **** ** predetermined ** *** **** of ***** ********** ** intervention."
"******** **** *** ****** after ***** ** ** non-staffed *****."

*** ********* ***** **** up **** *** **** perfectly:"******** ******** *** ****** manpower" **** ******** ***********, ********* can **** * ***** cost ******* ** ******** applications.

Additional ********** ************ *****

***** *** **** ****** usage ***** ** ******* around ********* ********** **** deployments, ***** *** ********** applications, **** ** ****** counting (******** ******** ** as * **** ** "Business ************"), ******* ***** recognition, *** ************ *** forensic ****** ** *****. In **** ***** *********** see ***** **** ** being "*****" ******* *** direct *********** ** *** system, ********** ** *** form ** ***** ******, is **** ******* ** detrimental ** *** ********* operation.

"***** ********* ***** *** where **** **** * years ***. * ***** analytics ** * ******** tool **** ******** **** Briefcam ** **-****** **** Agent ** *** **** viable. **** ********** *** typical ***** ****** *** receive **** **** ********* and ****** * ******** to ******* ***** ******** video *** **** **** just ************. **** ***** said, ***** ** * significant **** ********** **** many ********* *** **** customers **** **** **** to *******."
"******* **** ******** ***** are. ** ***** ****** counting ** ***."

***** ** ********* *** storage ******* ** **** a *** ***** **** we *****'* **** ** previous *********:

"******** ******. **** ********* for ****** ** ******* requirements ******** ** ****** motion ***** *********."
"** ***** ***** ***** analytics **'* **** ** our ******** ******** *** reduction *** ********* *** storage *** ********* ** the ******* ****** *****."
"******** **'* * ***** as ** ***** ******* amounts ** ******* ***** not ****** *** ******** constantly."
"** ****** ***** ********* because **** **** ** can ****** *** ******* capacity, *****-******** ******* *** storing ** *** **** video ****, ***** *********"
"******* **** * **** year *******: *** ******* records ** ** ** hours/day. ***** *********, ********* decreases ** * ** 4 *****."

***** ******* ********* ** certainly * ***** *********** for *********, ** ** a ******** ****** ************* given **** ** *** hype ********* ************* **** promoted **** *** *****. Newer *********** ******* ********* ****** ** *.*** (*.*** ** ******* ****** vs *.***) *** ********* *********' ******* in ******* *********.

Analytics ********** ** *******

**** *********** *** ********** ********* analytics, *********** ****** ****** in ***** *** ***********:

"************, **'* **** ****** customer ******, ***** *** been ***. ********, **'** begun * ****** **** into ***** ********* ******* so **** *** *******, including *** **** ******** players, *** ***** ************ are ******** **** ****** at **** ***** ***** points."
"***** ******, ** ***** it ** *** *** the ******** *** *** benefit. *** ***** **** come **** ** **** (especially *** ******** *******) that ** *****'* **** sense ** **** *** it ** *** ***** process. **** **** * few *******, *** **** try ** ***. ***** they **** **, ***** not...at ***** **** ****'* have ** *** ***** for **."
"**** ********* ********* ********* we ******* ** ** our ********. *********, ** costs ** *** ********* although *'** **** * Chinese ************ ********* ********* with *** **** *******. Adaptive ******** ********* **** to **** ******. ** see ********* ********** **** in ***** ***** *** when ** ***** ** implementation **** ***'* **** to ****** **** ***** analytics ** **** ** be **** * *********** rather **** **** ****."

Manufacturers Need ** ** ******

**** *********** ***** **** manufacturers *** *********** *** unclear ************:

"** *** ** ***** them ****** ******* ** unrealistic ******** ************ *** also ******* ** *** administrative ******** **** *** larger *********."
"*** ****** ** ****** residential *** ***** ********. Although **** ***** ********* can ** **** ** this ****** ** **** not *** *** ******** to ****** ** ** make ******** *** *****-*******."
"*'** **** **** *********** demos *** *** **** client **** *** *** been ****** *** *********. And ****** * **** not **** ******** **** I ***** ********* ** having ***** ** ** core ****** ****. ****** unrealistic ** ****** ************ most ** *** ********* have *** **** ********. If ** *** **** active *********/***** ***** ********* I ***** *** **** changing."

*******, ********* *** ***** far **** * "**********" product, ****** **** *********** have ***** ******* **** a ******* *** ** manufacturers ** ** ******** applications. ** ****** ********* to ******** ** **** challenges **** *********** *** meeting ************ ** * broad *****, *** *********** that *** ************ ****** proficient ** ******* ********* products *** *** ************ may **** ********** * head ***** *** * competitive **** **** ***** competition.

Comments (9)

Undisclosed #1

IPVMU Certified | 02/26/16 03:36pm

For the purposes of the survey, is simple 'record on motion' considered an analytic?

Sagy Amit

02/26/16 06:35pm

The tragedy of Video Analytics mainstream adoption is the fact that every single company that had real potential with quality products, and high investment in R&D, ended up being acquired and stripped out of its creativity.

ioimage went down those lines first with DVTEL acquisition, hopefully FLIR will revive that brand and invest heavily in more R&D.

Then Avigilon took the edge recording out of VideoIQ because it doesn't fit their agenda of selling VMS and storage. The iCVR was a great product and with a little investment in more camera models and smaller size could have been an industry winner.

Curious to see what comes next but I am certainly not holding my breath. Seems like the commodification of surveillance cameras will also put the final nail in the Video Analytics coffin as the lines with "motion detection" get blurrier by the day.

Undisclosed #1

IPVMU Certified | In reply to Sagy Amit | 02/26/16 07:56pm

Then Avigilon took the edge recording out of VideoIQ because it doesn't fit their agenda of selling VMS and storage. The iCVR was a great product and with a little investment in more camera models and smaller size could have been an industry winner.

As a point of public interest, maybe a special exemption could be granted allowing B.K. to respond? :)

John Honovich

IPVM | In reply to Undisclosed #1 | 02/26/16 09:38pm

Oh brother...

Fwiw, I was never a fan of the iCVR approach (too expensive, too bulky). I like what Avigilon did by making the VideoIQ analytics available in all new H4 cameras and at no additional cost.

Basically, I don't see a large market that wants high end analytics and long term on-board storage on the same device.

Sagy Amit

In reply to John Honovich | 02/26/16 10:06pm

Exactly my point John. As far as I know, there are more iCVR's channels installed around the world than any other analytics. Brian can correct me if I am wrong..

Point is that DESPITE the bulkiness, and the unnecessary large hard-drives it was still a best seller.

with smaller form factor, smaller SSD and a freeware VMS like what used to be called VIEW, I think we would have seen a better adoption percentage for analytics.

Brian Karas

IPVM | In reply to Sagy Amit | 02/26/16 11:01pm

The tragedy of Video Analytics mainstream adoption is the fact that every single company that had real potential with quality products, and high investment in R&D, ended up being acquired and stripped out of its creativity.

Without even needing to get into specific companies, I think the larger issue is that analytics is a feature, not a product.

There haven't been any "analytics" companies that have managed to be hugely successful because analytics on their own don't really do much. There's the obvious part about needing a camera to run on, but I think there are less obvious road-blocks as well.

Analytics really need a strong UI/UX to be valuable. In the security world this is essentially the VMS. If you can't percolate an alert up to an operator, and give that operator efficient tools to handle the alert and do things (audio talk down, toggle relays, move a PTZ), the analytics will have reduced value.

Even with ONVIF there still isn't a fully ratified, fully tested, fully functional API for:

Configuring rules/alarms
Receiving events
Automating responses (on Rule X do Action Y)
Sending feedback back into the system (tweak rule sensitivity, etc.)

This was even more true back when you and I were trying to push this stuff out into the world half a decade ago.

VideoIQ attempted to solve this by rolling an all-in-one solution, including a VMS that was tightly integrated. This worked really well for people that weren't married to a VMS, and less well for people that already had a VMS or needed features not in the free VMS.

IOI attempted to solve this by pursuing VMS integration partners and trying to be as flexible as possible, but that is neither easy or cheap.

But the result is that its kludgy and costly to make an analytics "product" that can be marketed as used as widely as a camera. We tend to see close partnerships with some VMS's and not others. Things don't work the same in every interface, features are not universally available, etc.

IMO it's a good thing that some of the most viable analytics companies have gotten acquired by larger camera/VMS manufacturers because I think that is the most probable path to success.

Brian Karas

IPVM | 02/26/16 10:40pm

As far as I know, there are more iCVR's channels installed around the world than any other analytics. Brian can correct me if I am wrong.

I don't see how that would be possible. Just given the sheer number of analytics companies, there's a lot of them out there.

I would wager there's probably more VideoIQ/iCVR/Rialto devices still in active use 6 months after deployment than any other analytics, as a ratio of channels shipped:channels active.

Undisclosed #1

IPVMU Certified | In reply to Brian Karas | 02/26/16 10:51pm

Correct me if I'm wrong but the iCVR analytics were never just thrown in 'free', true? If Sagy intends "more iCVR channels installed around the world than any other paid-for analytic", that's possible, no?

Undisclosed #2

03/02/16 03:14am

I would lobby for analytics, but Clients often don't use it. I am a consultant, so I use analytics sparingly. There are some changes in the risk environments that makes certain analytics viable and saves the client on cost. In schools for example, virtual trip wire -- particularly when coupled with other rules such as motion, direction, etc -- are a viable solution to detect and display fence climbers and bad guys accessing rooftops. During the night, you would rely on the interior IDS to provide security.

I would be interested to find out what metrics integrators use for analytics (i.e. 2 hours for each camera? Some percent of overall bid amount?). Preferably, what labor hours do you expect to spend on VA (man hours is easier to translate across the country)?

Read this IPVM report for free.

This article is part of IPVM's 6,534 reports, 880 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.