Verkada Tailgate Detection Tested

By Brian Rhodes and Jermaine Wilson, Published Oct 20, 2021, 10:54am EDT

Verkada has added in tailgate detection, for no additional cost, however, IPVM testing shows many missed events and significant delays in receiving alerts.

IPVM Image

This feature caught IPVM's attention when Verkada sent mass emails alleging that it could "prevent $100,000 tailgating entry incidents".

We worked with Verkada's product and support team repeatedly over weeks and tested Verkada's feature in several scenarios, with multiple subjects at various speeds, in varying light levels, with changing intervals and distances.

In this report, we break down performance, outline the challenges, and include Verkada's response to IPVM:

  • Tailgate Detection Is Free Verkada Feature
  • Takes 7 to 11 Minutes To Alarm
  • Not Available For Outside Camera Views
  • Poor Detection When Subjects Follow Closer Than 6 Feet Apart
  • Poor Detection When Door Held At Night
  • Consistent Detection When Subjects 6 or More Feet Apart
  • Consistent Detection For Delivery Carts/Packages
  • Verkada's Setup Guidelines
  • Camera and App Setup Details
  • Pricing Compared To Alternatives

Executive *******

***** ** *** *******, Verkada's ******** ********* ****** many ********** ********** *** suffers **** *********** ******* (7 ** ** *******) in ********** ****** *** those ********* ** **** work ***** ********** *** effectiveness ** ******** **********.

************, *** ******* ***** that ******* ****** ********* tailgaters **** **** ****** 6 **** ** *** valid ****** ******** ~**% of *** ****. ********** beyond * **** ** the ***** ****** **** 100% ********.

****, ******** ******** *********** greatly. ******* ********** ******** cameras '******* **° *** 55° ******** ** *** perpendicular ****' ** *** side ** * ****. A ******* ************ ****** will ****** ** ********** differently, *** * ********* camera *** **** ** be ********* *** ** to ****. ** *** to ***** **** *** setup ** ******* ******** changed *************** ***** *** they ****** *** ****** to ** **********.

Verkada **** *********** **** *******

******* ********* ** ****'* findings, ****** **** **** will ******* *********** *** fix * ******** *** impacting *******:

****** *** *** *********** about **** ********** ** testing *** *** ********** feature. ** *** **** that **** ******* ** working ** ****** *** feature *** * ****** update. ** ** ****** our ****** ******, ** expect *********** ** ****** to * ******* ******* of ***** ***** **** time.

****** *** ******* ** know ***** *** ********* latency *** **** ***********, which ** ********** ** result **** * *** which ** *** ******* to *******.

** / **** ******* does **, ** ***** them ** ****** ** and ** **** ****** our *********.

Tailgating *********

**** ** ****** ******* point ** ****,************ ***** ********* ** deal **** ******* ********* security ***** *** ***** a ******** ******* *** system ***** ***** ** their *****.

** *** ***** *****, we *********** ********** (******* accurately *******/****** *** *****) - ***** * ************ user ****** **** * door, *** ******* ********* them ****** ******* *** open **** ****** ** recloses.

*** ****, *************: ****** ******* ***************** **********: ******* ******** Vulnerability.

Verkada ******** *********: ** ***** ***

********** *** ********** **** is ***** *********, **** specialized ******* ***** ******* thousands *** ***-** ********* hundreds *** ******. *******, Verkada **** '******** *********' analytics *** **** *** video *** ****** *********.

7 ** ** ******* ** *****

* *********** ******* **** Verkada's ***** ** *** long ****** ** **** needed ** **** * tailgating ***** *** ****** a ****.

IPVM Image****** *** ******* *******, the **** *** * tailgating ***** ** ******** was ****** ~* ****. At *****, ***** ********* took ***** ~** **** to ********.

**** ******** ** ~** to ~*.** ****** **** Verkada's ***** "*** ***** **** ******** in ******* ****** * minutes."

*** ***** ** *** right **** *******'* ********* shows *** ***** ****** appear ** *** ****:

** ***** ** ********* risk, ** ********** *** not ***** ** * tailgate ***** ***** * to ** ******* ***** it *******, ******* *** loss ***** **** ******* occurred.

**** ** '**** ****' 3-minute *********, ***** **** always ****** ** *** tests, *** ****** ** time ****** ************ ** too ********* ** ** useful *** **** *****.

Not ********* *** ******* ****** *****

******* ********* ******** ********* for ******* ***, *.*., before ******* *********, ** the ********* ***** ***** shows:

IPVM Image

***** ** ******* *** test, ******* ******** ** that, ******** ** ***** marketing, ******** ********* **** not **** **** ******* scenes *** ******* *** marketing *****.

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

** ***** ** ***********, our ******* ***** ****, if ******** *** ****** closely, ********* ** ****** the ******** ** *** time, *** ** *****.

* ** ** (**%) tailgating ****** **** ****** when ******** **** **** than * **** ***** during *** *** ** night.

Daytime ***********

** ****** ****** ********** scenarios ****** *** *** night, *** ****** ********** scenarios ****** *** *** that ********* ******* ********** and ******** **** ******* during ******* ********* *****.

IPVM Image

*** **** ******* ** all **** *********:

  • * ** ** ****** when *** ******* ******* behind * ** **** feet.
  • * ** ** ****** when ******** * *-**** table **.
  • * ** ** ****** when ***** * ******** cart.
  • * ** ** ****** when ******** **** * or **** **** *****.
  • * ** ** ****** when ******** **** **** than * **** *****.
  • * ** ** ****** when *** ******* ********* to **** ******** *** the ***** **** ******** up ******** *****.

Nighttime ***********

** **** ****** ** night *** *********** ***********. In *******, ******** **** tailgate ********* **** ********, with * ******** ** the **** ***** **** open ***** ************* ******, with **% ** ***** events ******.

IPVM Image

*** **** ******* ** test *********:

  • * ** ** ****** when ******** **** *+ feet *****.
  • * ** ** ****** when *** **** ** held *** ******* ******.
  • * ** ** ****** when ******** **** **** than * **** *****.
  • * ** ** ****** when *** ******* ***** up *** ****** *********** after ********.
  • *****, ***********, *** ********** intermittently ***** ***** ********** and ******* ********* ******** that ***** ****** ********.

Verkada ****** *** *** '******* ********' ********** *********

** ***** ******* ** comment ** ********* *********** problems, **** **** **:

“** *** **** **** 100% [********** *********] *** it ** *******-******** *** should ****** ** ******** that ** ********* ** improving *** ********.”

******* ***** *** *********** shortcomings ** *********** ** is *** *** ******** applications, *** ***** ****** not ****** ******** *********** that ** ***% ********.

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

*** ***** ******, ******* the **** **** *** a ********* *** *** detected **% ** *** time.

******* ********* ** ** given *********** ** *** glass *** ***** **** are *** ******** ****** the ***, *** ***** complexity ** ****** **** the ********** *** ******** and ****** ******** *** the **********.

***** *******, * ** 10 (**%) ****** **** the **** ** ****:

Consistent ********* **** ******** * ** **** **** *****

** ******** ** ******** where ******** *** *****, Verkada's ******** ********* ******* error **** *** ******** subject ******** ** * distance ******* **** * feet. **** *** ***** in **** *** *** night **********.

***** *******, * ** 10 ****** **** * subject ******* ****** * or **** ****:

Consistent ********* *** ******** *****/********

*** ******* **** *** not ******* ******** **** carts ** ***** ******* deliveries. ******* ***** ********** typically ***** ****** *** day, *** ******* ** problems **** ******* **** evident, ** *** *** repeat ** ** *****.

***** *******, * ** 10 ********** ****** ****** when ***** * ****.

*** ***** *** ***** themselves *** *** ****** any ***** ******, *** the ****** ******** ******, not *******.

App *****

***** *** *** ******** requires ********** * ****** is *******, *** *********** the **** **** * Verkada ****** ****. *** Verkada ******* ********* ** unified, *** **** ***** and ****** ******* *** configured **** *** **** portal.

*** ***** ****** **** must ** ********* ****** ******** ********/********** (***) ** *** access ****. **** ****** may *** ****** ** used ** *********, *** it ** ********* *** tailgate *********. ******* **** us *** *** ** used ** ******* *** detection ********, *** ** scenes ** ******* ******* them ** *** ******* the ******** ********.

*** ******** ***** ***** shows *** *** ******* feature ** ***** *** configured:

Setup ****** ***********

*** ** *** ********** Detector ******** ****** *********** of *** ****** ********** ************* *************.

IPVM Image

***** *** ************ ** mount ******* '******* **° and **° ******** ** the ************* ****' * normal ************ ****** **** likely ** ********** ***********, and ******** ********* *** not **** ********.

*** *********** ** ******** cameras ** *** **** of *** **** *** a *********** ******* ******* after ** ***** ******* and *********** **** ********* performance.

******* ********* ******** *** camera ** ** ***** to *** ******* ******** accuracy:

****** *** ****** ** the **** ****** ******* results ** ******** *** likelihood ** *********** **** incoming ******. ****** ****** can ****** *** ******* to **** ********* ********** events.

Pricing ******** ** ********* *******

******* ******** ********* ******** are **** *** ** extra **** ** *** it ** *******.

** ********, ***** ********* like**** ******** ********* *********** ** ~$***, ***** dedicated ********** ******* **** as*************** ************* ***** ** $*,***+.

Camera *************

*** ******** **** ***** on ***** ******* ******* hung *******. ***** ******* told ** *** ******** pixel ******* ** ~** ppf *** *********,************* ****************** **** ** ******* camera ***** ******** **** 25' (*** **********) ** ***' (** **********)

******** *** **** *** supported. ***** ** ***** testing, ******* **** ** the ******* ***** **** previously ******** *** ******* would *** **** ***** we ******** ******** ***** it. ******* *******:

*** ******** ****** *** be *** ***** ***** than * *******. ********* for *** ********* **** -- ** **** ******* the ******* ************* ** indicate ****.

"Prevent" ** "******"

** ********* ********* - ********* "******* $100,000 **********" ********, *** *******'* *********** email ******** ******* ****** to '*******' ********** ****** the ******* ** ******** detecting **.

*** ********** ******* '*******' and '******' ** **** established ** ****** ******* use ***, **** ***********, in ******** ********. ** prevent ** ** ******** stop ** ******** **** happening. ******* ** *** capable ** ***** ****. Verkada, ** ****, ********* an *****, ***** ******* has *********.

******, ******** ** *******'* marketing *****,***** ************ *********** ***** **** "******** detection", *** **********.

Comments (13)

Shannon Davis

IPVMU Certified | 10/20/21 03:16pm

***** ****** ********. *** sure *** *** ***** use **** ******* ** the ***** ***** ** 7 ** ** ******* later *** *** ******** need ** ** *' apart ** **** *** better *********.

Agree: 4
Disagree
Informative
Unhelpful
Funny

Jermaine Wilson

IPVMU Certified | In reply to Shannon Davis | 10/20/21 03:45pm

***** ***, *******!

**** ** *****'* ******, I ****** ***** ** requested **, *** ***** you *** *** **** words!

* ***** ** **** concerned ** ** ****** on *******'* ******** ********* for *** ********, **** is *** ****.

Agree: 1
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed Manufacturer #1

10/20/21 03:39pm

"** *** **** ****.....*** it ** ******* ********..." When **** ********* *** be **** ***** *****? Mission ***********, ** ** ideal *****, ****** ** implied. **** ******** ** worth ******* ** ** works ** *** ** 60 ***** **** ********** variables. *** ** ** gonna ** ******* *******? Do *** ****** **** are ******** **** ***** if **** *** *** predictably *****? ***** **** value ***** ********** ***** of **** ******** ** would ***** ****** **** to ************ *****? ** I *** **** ****** frustrated **** *********? ****** already. ******* ** *** alone ******. *** ***** industry ** ***** ****** by ********* *** *** been *** *****. ***** needs ** ***** ** with ********.

Agree: 2
Disagree
Informative
Unhelpful
Funny
Avatar

Brian Karas

In reply to Undisclosed Manufacturer #1 | 10/20/21 04:11pm

** * *** **** person ********** **** *********?

**.

***** ********* ** * very ********** ***** ** the ******** ********.

***, *** ******* ***** of *** ******** ***** is **** ** *** fairly ******** ****** ******* that *** **** ** when ***** *** >* people ** ******** ** a *****. * ** not ****** ***** ******* can ** ****, *** that ***** *** ****** products **** *** ****** point-n-shoot ********** (**: *** needing ********* *********** *** setup *********) *** ********** that ** *** **** person *** ******* ********* a ****** ******* ** some *****.

* *** >* (******* than ****) ******* ******* as **** ****, *** many ****** ****, ******* accurate ****** ** *** number ** ******* **** challenging ******* ** **** scenes. **'* **** **** challenging ** ******* *** overlapping **** *****, ** otherwise ******** ** **** way.

**** ********* *** **** offering ***** ** ****** sub-classes (**: ***/*****, ***/***/******) and ********** *********** (**: wearing *** ** *******, color(s), ***). *** ******** here ****** ******* ** real-world ***********. **** ***** it **** ** ***** "trust" ********* **** ** analytic ******** *** ****** when *** *** *** "Silver ****" *** ** tells *** **** ******** results. *** ***** ****** no ****** ****, ** did *** ** *** that ****** ** ** a ***** ** **** sedan? ** ** ***'* trust *** *******, *** we ****** ***** ** enough ** ******* * major *******?

* *** ********* *** analytics *************, ***, ** that *** ****** ** demand ***** ***** *** is ********** ** *** supply ****. ***** *** a *** ** *********, competing *** ********** *** demand, ***** ********** *** marketing/messaging ** *** * bit *** *** ***** of *** ****** ******* capabilities.

** *** ****** *** analytics ***** ** *******, but ****** **** *** wrong ************ ** ** as *** ********** ****** today. *******, ** *** video ***** ** ***** at *** ********* ***** of * * **** old *****. ** *** detect * *** ** things, ************ ** ***** conditions, *** ** ** not ****** **** *****. The ****** ** ********* is **** **** ** not *****, ** ****, and ********** ******** *** scale ** ****** ****. But, ****'** ****** *** that ***** ** ******* of ********* **** ****** for ***** **** ****** "tell ** **** ***** are **** **** **** people ** ******** ** this *****". ** ***** not ** ********* ************* want ** ****** ****** for * ******* ** enable **** ** **** in ********.

********** ********* ***** * little *** **** *** modern ******* ** *** "bag **** ******" ******** that *** * **** of ****** ***** ***** analytics ********** ** **** 10 ***** ***. **'* a ******* ******* ****** for ********* ********* ******* it ***** **** ********* that *** **** ********* value, *** **** ********* customers **** ***** ***** on. *** *** ******* is **** ** ** unlikely ** **** ******** enough ** ****** ********* for ****** ** ****** trust ** *** **** on **. ********* ************* promoting **** **** ****** as ******** ** ******, IMO, *** ********** **** they ** *** **** much ** ***** ** real *********** *****.

Agree: 6
Disagree
Informative: 5
Unhelpful
Funny

Undisclosed #4

IPVMU Certified | In reply to Brian Karas | 10/25/21 08:41am

*** ****** ** ****** ratio ***** *** ** unbalanced ** *** ****** side. ***** *** * LOT ** *********, ********* for ********** *** ******, which ********** *** *********/********* to *** * *** too *** ***** ** the ****** ******* ************.

***, ***’* *** ** Supply *** ****** -

**** ****** ********* *** demand *********, ******** ******* to **** *** ***.

Agree: 1
Disagree
Informative: 1
Unhelpful
Funny: 2
Avatar

Brian Karas

In reply to Undisclosed #4 | 10/25/21 11:40am

**** *******, **** **** sound **** ********* * would ***.

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Ross Vander Klok

IPVMU Certified | In reply to Undisclosed Manufacturer #1 | 10/21/21 01:17pm

"*********** *** ***** *******" seems ** ** *** analytics ********'* ******** *******.

Agree: 2
Disagree
Informative
Unhelpful
Funny

John Honovich

IPVM | 10/20/21 08:00pm

**** **** ***** *** Jermaine!

******* ***** ** ******* the ******* *** *********** teams ** (*) *** allow ********/******** ** **** until **** ****** **** they **** *** (*) significantly ******** ******* / engineering ********.

** **** **** ***** behind *** ******, ***** I ******* ******* *** answering *********, ********* *** product ** ** *** the ****, *** ******** to ******* *** *******, this ** *** ********** for * ********** ******* (which ** ***).

******* *** ********* ************* with ****** ***** *** way. ** *** *****, we ***** *** *** any ********** **********, *** then **** *** ******* it ******* ******, ******* us ***** ******** ****. Verkada **** ** **** pushed *** ** ******* and **** **** *** had * ****** *****.

*** ******* ******* ****** we **** ***** **** are *** ***** ****** but *** ********** ********* they *** ************ *** are ******* *** ******** that *** *** ***** to *******.

** **** * ***** quantitatively, ******** ***** ******* has **** **** ** the ****** ** *********** than **** **** ********* and **** ********* ****** in ***** ** *** outstripping **** ** ***********:

IPVM Image

Agree: 1
Disagree
Informative: 2
Unhelpful
Funny

Undisclosed Integrator #2

In reply to John Honovich | 10/20/21 08:47pm

****, ** **** ***** persuing ***** ******** **** seem ** **** * decent ****** ** ******** for ******* ***********/******* ** we'll **** ** *** if **** ******** ******. They ***** *** *** of ******* ** ***** of ******** **** ***** to ******* *** **'* a *****.

Agree
Disagree
Informative: 1
Unhelpful
Funny
Avatar

Jacob Hengel

YourSix | 10/21/21 04:02am

*** ******* **** * light ** *** ***** for ** ****? * minutes ******'* **** ** acceptable *** *** ** this **** * ************? A ****** **** ********* in *** ***** *+ minutes ** **** ** alert?

Agree: 4
Disagree
Informative
Unhelpful
Funny

Undisclosed Integrator #2

In reply to Jacob Hengel | 10/21/21 02:00pm

** * **** ** guess **'* ******* *** the ****** ** *********** when ** ***** ** linking ****** ****** **** the ***** *********.

Agree: 1
Disagree
Informative
Unhelpful
Funny

Undisclosed #3

In reply to Undisclosed Integrator #2 | 10/21/21 02:14pm

* ***'* ***** ***** is ****** *** *** technical ****** ** ** why ** ***** ~* minutes ** **** ** alert. *******, * ***** he ** ******** *** how ********** ******* ** for ******** ****** ***** ever **** ********* **** only ***** **** ***** the ******* ******.

*.*. ** ** ********** useless ** ********* *****...

******** * **** **** to * ***** ** shit *****'* **** ** any **** * ***** of ****.

IPVM Image

Agree: 3
Disagree
Informative
Unhelpful
Funny: 3
Avatar

Paul Anderson

10/25/21 08:03am

****** *** *** ******, very ***********. * ***** with **** ********, ********** around *** ****** ** time ** ******** ** alarm. **** ** *** analytics ****** ****, * mins ** ************ ** i ****** ** ******* Tailgating. ** * *** reviewing *** **** ** tailgating ***** **** *** installed, * ***** **** the **** ***** ** risk!! *** *** ***** that ****** *** ** 3 ****? **** ****** could **** **** **** in **** ****? *** alone * **** ** longer!

Agree: 1
Disagree
Informative
Unhelpful
Funny
Read this IPVM report for free.

This article is part of IPVM's 7,330 reports and 971 tests and is only available to subscribers. To get a one-time preview of our work, enter your work email to access the full article.

Already a subscriber? Login here | Join now
Loading Related Reports