Failed Retail Video Analytics Startup Analyzed

By Brian Karas, Published Dec 19, 2017, 11:35am EST

Most manufacturers in this industry try hard to cover up failure.

But an industry startup backed by Silicon Valley's most prestigious accelerator has provided an in-depth account of the causes of their demise. Inside this note, we review and analyze them, pointing out other important causes, and companies struggling in retail video analytics.

Prayas *********

****** *********, *** ****** *******, was ******** ** ***** a ****** ********* ********. The ******* *** **** of ************** ****** **** ******* batch. ****** ***** ******* is ***** ******, *** company ******* ** * different *********** *** ****** their ****** ********* ******** in June ****.

**-******* ******* ********** [**** ** longer *********] ********* *** ******* *** the ******* ******, ** * **** detailed *** *********** **** post.

Lessons *******

******* ****** **** ** the **** **** *** particularly ******** ** ******* challenges ** ****** *********:

** *** ****, ***** looking ** *** ****, someone *** ** ****** out **** ** ** and **** ********* *** solution. **** **** *** analytics ******* **** ******, and ********* ******** ** adoption. 

**** ** ******* *** of *** ******* ********** for ****** ********* ********. While **** *** ******* metrics ** ******** ****** the *****, ** ***/****** profile ** ********, *** retailer ***** ***** ** decide **** ** ** with **** ****, ***** is *** ****** *******. Further, ** *** ** difficult *** *** ********* products ** ***** ***********, as ********* ***** ** retailers **** ****** ********* target ******* *** ***** layouts. *** ******-**** ***** browsing ** ****** *** 10 ******* * **** sign ** * *** one? ** **, **** do *** ** ** encourage ** ********** ****?

**** ******* **** ********* protection-style *********, ***** *** action **** *** ********* is ******* **** *****-***: determine ** *** ****** is *** ********** ** be ** *** ********, and ** **, ******* to *** **** ** leave *** ******** *** guards, ******, ***.

******* **** *** ******** model *** **** ****** didn’t ****** *** **** like **** *******, *** so ***** *** ** imperfect **********. ***** ********* weren’t **** ** ***** this ****, ** **** them * *** ** figure *** *** ** use **. ** **** couldn’t ****** ** ***, the **** *** *******.

****** ********* ********* *** often ********* **** **** customers *** *** ********** have, ***** *** ******** early ********, *** ** the **** **** ****** the ******* **** *** first ***** *****: ********* do *** **** ********* and ******* ** ***** to *** ** **** data. ** *** **** can *** ** **** in **** *** **** increases *******, ** ***** value.

*** ********, *** ** the ***** ****** ** should’ve **** ***** **** been ** **** ** an ****** ** *** field ** ******** ******* and *** ****** ************. We *****’** ****** ** this ****** ***** *** plans ** *** *** the ***** **** ***** cameras, *** *********** ** they **** ********* ** not. ******* ** **** panicked *** ******* *** a **** ****, *********** between ******* ******* **** it ***** **** *** hopeless **** ** ******’*.

*** ******** ** ****** lacked ******** ******** *********, or **** *********** **** the ******** ********. ***** model *** ***** ** pulling ***** **** ******** surveillance *******, *** **** doing *****-***** ******** ** the ***** ** ******** their ****/*******. *******, **** were *** **** ********* it *** ******** ** retrieve *** ***** **** required, *** **** ********* unaware **** ***** *** a **** ****** ** systems ********, ***** **** unique ***'* *** **********, which **** ** *********** costs.

******* ********* ******** **** infrastructure ******. ** *** case, ** *** ******* with *** ****** ******* and ****** ****** **** typically **** *** * 1 **** ********** ** power ***** *** ******.

******** ********* *** **** a ******** ****** *** many *****-***** ***** ******** in *** ********** *****. Though **** ** ********* bandwidth *** ** ******** better ** **** **** when **** ***** ******** in ****, ** ***** remains * ******** ****** for ***** ********* ** scale *******, ** ******* customers **** **** **** a *** ******* *** location.

** **** ********** ****** by ***** ***** ** progress. ** ****** ******* that ** **** **** close ** ******* *** deal... Three **** ***** * was ******* ******* ****** that **’** **** ***** to * ******* ****** contract **** ****. ***** years ***** — ***** no ********!

**** ** * ******* often **** ** ******** in *** ******, ********* customers *** ** ********* say "***", *** ****** to ***** ********, *** they **** *** ***** not ******* ** **** a **** "**" ******. This ******* ** ********* anticipating ******** **** **** never *****, *** ******** resources ** ******** **** opportunities, **** ****** *** would-be ******** ***** ******* and ************. *******, ****** customers ** ********** **** a ********** *** ***** extremely ****-*********, ********* ** spend ***** ** ********** that **** *** ******** improve *****, ****** ******, or ********* **** * direct ******-**** ******.

Deserve ******* *** ********* *** *********

****** ********* *******, *** knowing **** ** **** onto ********* **** ** uncommon, *** **** ** the ******** ********, *** in **** ****** ** well. ******* ********** ******** respect *** *********** *** knowing **** ** **** on *** ****** ********* else. ****** ********* ******* into "*********", ***** **** pivoted **** *********** [**** ** longer *********] ** ****** ****, which ** ***** ********* active (****** *** ******* to ** **** ******** or ****** *********).

Retail ********* *********** ******

****** ** ********* *** the **** ******* ** find ****** ********* ** be * *********** ******. ***** ******* ******** ** **** ** a ****** ********* ******** to ***** ******* ** Silicon ******, *** *** ***** quietly ******* ** * broader-scale ******** ********* *******, marketing ** ***** *********, such ** **********, *** offering ******** **** *********** search, ** **** ** various ***** ** ********* for ******* ******** ******* purposes. ***********, * ****** analytics ********** *** **** years, *** **** ~$* million ** *********** ******** ** ****, *** ** *** extent **** ***** ******** will ******* ***, **** depend ******* ** **** strong ****** ******** ** retail ** *******. ** 2015, ****** ********* *********** ********* sold ** ******** [**** no ****** *********], * ********** small ***** ********.

**** ****** ********* ********* have ******* ****** ********* applications **********, ******* ********** that ******** ******, *** willingness ** *** *** solutions, ** *** ******** to ******* **** ** perimeter ********** ** *********-******** search ***********.

Comments (9)

Undisclosed #1

12/19/17 11:39pm

"However, retail customers in particular have a reputation for being extremely cost-conscious, unwilling to spend money on technology that does not directly improve sales, reduce losses, or otherwise have a direct bottom-line impact."

Being 'extremely cost-conscious' (because of the thin margins/high competition in retail) tends to also make these customers 'extremely vaporware-conscious' as well.

I love how the dude insinuates that retail customers are apparently just too dumb to figure out the benefit/value of the data produced by his product:

"problem with our business model was that people didn’t really use data like this already, and so there was no imperfect substitute. Since customers weren’t used to using this data, it took them a lot to figure out how to use it. If they couldn’t figure it out, the data was useless."

Say what?  Whose fault is it that the customer doesn't understand the value of your product again?  The customers?

yeah... sure.

 

Avatar

John Bredehoft

Bredemarket
In reply to Undisclosed #1 | 12/21/17 06:53pm

I didn't interpret that statement as implying that the customers are dumb. I took it to mean that the vendor realized that he didn't educate the customer on how to convert the data into something actionable. 

Undisclosed #1

In reply to John Bredehoft | 12/21/17 09:28pm

My perception (and position) is that the data produced is not 'actionable' to begin with.  i.e. retail analytics are of limited actual value - and the market has spoken.

Hence, my interpretation concludes that Mr. Maheshwari also knows this and is left with nothing other than casting doubt on the customers ability to understand the benefits of the data his product produces in order to avoid the cognitive dissonance of the fact that his product adds no actual value to retail customers.

Undisclosed #2

IPVMU Certified | 12/20/17 12:36am

The founders of Prayas lacked security industry expertise, or even familiarity with the products involved. Their model was based on pulling video from existing surveillance cameras, and then doing cloud-based analysis of the video to generate their data/reports.

Putting the “cart” before the horse...

Marty Calhoun

IPVMU Certified | 12/20/17 11:30am

Openly admitting failure, and knowing when to move onto something else is uncommon, not just in the security industry, but in most others as well. Pranshu Maheshwari deserves respect and recognition for knowing when to move on and pursue something else

Very well said....

Undisclosed #3

12/20/17 01:49pm

Many analytics companies fail to realize they are data rich and information poor.  Video content analytics is not about video, it is about the content.

Undisclosed #1

In reply to Undisclosed #3 | 12/20/17 02:07pm

"Video content analytics is not about video, it is about the content."

that is exactly right.

Retail analytics may provide data (content).... but if capturing this data provides no additional value to the customer, then why would they buy the product that produces this (useless to them) data?

Personally, I have long believed that the data that retail analytics produce is (primarily) data that is already known by retail establishments.  hence the struggle that providers of this data have historically experienced - and is probably the main reason that these providers have largely been unable to convince retail to buy their 'solutions'. 

 

Undisclosed Manufacturer #4

12/21/17 03:28am

I could write a book about this subject, (and maybe I will when I retire), but the single line from the blog that should be etched in stone was:

When we started working on the retail analytics idea, we didn’t even really have conversations with customers —

The "if we build it they will come" mantra may play well in Silicon Valley but it doesn't in Wisconsin. If you've asked no questions, then how could you expect to deliver relevant answers?

Undisclosed #3

In reply to Undisclosed Manufacturer #4 | 12/21/17 01:39pm

Hope is not a strategy. I find that engineering led start-ups to have this mindset more so than sales led start-ups.  Tools don't build houses, carpenters do and salespeople sell them.

Read this IPVM report for free.

This article is part of IPVM's 6,817 reports, 914 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.

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