Consider these statistics across
social media channels for a single
day: people spend 1 billion hours
on YouTube1; on Instagram, 95
million photographs are uploaded2;
on Facebook, 60 million emojis are
used3.
The exponential growth in the
sharing of images is deeply
psychological. It is easier to
express a moment as a picture
than in words. Pictures capture
more information and require less
skills and effort than writing. The human brain processes visual
content more quickly.
According to MIT4, the human brain
can process entire images that the
eyes see within 13 milliseconds. It
explains why people find visual
content more engaging. This is one
of the key reasons why a social site
such as Instagram has become
popular quickly in terms of user
engagement5.
That it is fatal for businesses to
ignore this wealth of visual data would be stating the obvious. By
2021, the global image recognition
market is expected to touch USD
38.92 Billion6. The global video
analytics market is expected to
grow to USD 8.55 Billion by 20237.
As image analytics continues to
develop, companies will see value
in spending more dollars on this
technology. Let’s take a look at
some of the industries where
image analytics is already
delivering immense value.
Saying More With a Single Click
Consumer Packaged Goods (CPG)
companies spend billions of dollars
in sales and merchandizing
strategies, but do not reap its full
benefits due to several challenges
related to execution at the store
level. This could be due to
insufficient shelf monitoring,
unreliable store checks or limited
resources to conduct audits at
stores.
By leveraging image analytics, companies are now tackling such
challenges efficiently in the
following ways:
-
In-store operations are
effectively monitored and tracked
using shelf images to get real-time
key performance indicators
such as share of shelf, on-shelf
availability, stockouts, pricing
changes and compliance metrics
-
Field sales representatives are able to cut down time spent in
conducting manual store checks
by using image recognition
technology, and are expanding
their store coverage and tackling
more critical issues at stores
By providing real-time, accurate
in-store insights, image analytics is
becoming the game-changer in
optimizing retail execution and
recovering lost sales.
Advantage Facial Recognition
Apart from gathering insights from
social media, businesses are
applying image analytics across the
customer journey for outcomes
ranging from personalized
marketing to improving customer
experience.
For example, Singapore’s Changi
airport is replacing passport
checks with facial recognition on a
trial basis8. This is in addition to the
airport’s already existing facial
recognition for self-service check-in,
bag drop, immigration and
boarding. Facial recognition can
also be applied to trace passengers
who missed their last call to board
the flight.
The U.S. Transportation Security
Administration (TSA) is deploying
new scanners at airports across
the country9. This will allow TSA to
virtually unpack bags and ease the
queues at security checks. Such
applications will help improve the
customer experience, which has
been a concern lately for the airline
industry10.
A leading retail company is
developing a facial recognition
system to gauge customer
dissatisfaction through their
expressions and movements11.
A California restaurant chain is
implementing facial recognition
systems to activate customers’ loyalty program without the need to
swipe their cards as they approach
in-store kiosks12. The system will
also display historic meal
preferences to reduce the time to
place orders.
Facial recognition is also being
enabled to validate customer
identity at retail stores by card
providers such as Mastercard.
This is expected to reduce cart
abandonment at the payment stage
by 70 percent as it will eliminate
the challenges around one-time
passwords sent through text
messages13.
The insurance sector is using
image analytics to improve the
customer experience of motor
insurance claims. A leading
insurance provider in the U.S. has a
claims process wherein customers
can upload pictures of their
damaged vehicles14. The company
analyzes the pictures and
processes claims. This eliminates
the need for physical verifications
which is both time-consuming and
effort-intensive.
For risky assessment areas such
as rooftops, drones can help
reduce liability from workmen injuries which cost companies an
average of USD 910,000 per
incident in the U.S.15 Expedited
claims processing is particularly
important when there’s a natural
disaster. For example, insurance
companies used drones to enable
faster assessment of losses left in
the wake of hurricane Harvey16.
Drone image analytics solutions
are also helping automate the
identification of risk factors, extent
of damage and claims estimation
reporting for insurance companies.
By combining structured and
unstructured claims data with images and other data sources,
deep learning models and machine
learning algorithms are being
applied to solve image
classification problems. The results
are then used to feed underwriting
models for proactive risk
assessment, and validate property
insurance claims for rooftop
damages caused by hailstorms and
other weather-related incidents.
Such use cases have delivered
more than 95 percent accuracy in
claims assessment and delivered
USD 30 Million in annualized
savings for a global property and
casualty insurance company.
A Source of Authentic Data
Image analytics technology is
maturing fast. Like any technology,
it will go through its own set of
challenges. Privacy is a question
that needs to be addressed,
particularly with rising concerns of
customers and tightening
government regulations. However,
the business case for image
analytics is too strong to ignore.
Companies in the U.S. alone spend
USD 10 Billion on third-party audience data every year17. Third-party
data, with its limitations of
accuracy, form the basis of large
personalized marketing campaigns.
With image analytics, businesses
will no longer be solely dependent
on faceless customer data and
percentages.
In the age of image analytics, data
will be of the actual customer, not
numbers on a survey. The data will
be authentic and derived from first-hand
sources such as facial
expressions. It will be real-time,
captured as the customers
experience something, often even
before they have made the
purchase.
Image analytics indeed has the
potential to bring forth a new
generation of insights where seeing
is believing.
References:
1. https://techcrunch.com/2017/02/28/people-now-watch-1-billion-hours-of-youtube-per-day/
2. https://www.wired.co.uk/article/instagram-doubles-to-half-billion-users
3. https://www.adweek.com/digital/facebook-world-emoji-day-stats-the-emoji-movie-stickers/
4. http://news.mit.edu/2014/in-the-blink-of-an-eye-0116
5. https://www.forbes.com/sites/jaysondemers/2017/03/28/why-instagram-is-the-top-social-platform-for-engagement-and-how-to-use-it/#50a9e86636bd
6. https://www.marketsandmarkets.com/PressReleases/image-recognition.asp
7. https://www.marketsandmarkets.com/PressReleases/iva.asp
8. https://www.bbc.com/news/technology-43962881
9. https://www.wired.com/story/tsa-tests-ct-scanners-airport-security/
10. http://www.airport-world.com/news/general-news/6284-new-study-reveals-that-poor-customer-service-levels-are-damaging-airlines.html
11. https://www.retaildive.com/news/report-walmart-developing-facial-recognition-tech/447478/
12. https://www.retailwire.com/discussion/facial-recognition-software-comes-to-loyalty/?utm_source=Listrak&utm_medium=Email&utm_term=http%25
13. https://www.pymnts.com/mastercard/2018/biometric-authentication-facial-recognition/
14. https://www.allstate.com/claims/quick-foto-claim.aspx
15. https://www.propertycasualty360.com/2018/07/05/hurricane-season-drives-mass-adoption-of-insurance/?slreturn=20180709015047
16. https://www.wired.com/story/houston-recovery-drones/
17. https://www.criteo.com/insights/companies-spend-20b-data-solutions/