last modified: 2023-05-15

EMLyon logo corp

   

1. From segmentation to personalization

Segmentation helps refine the picture from a mass of data to meaningful subgroups of data points.

Why not go down to extreme segmentation: segments the size of an individual?

  • Major websites do it (Amazon, Yahoo!, Netflix, etc.)

  • Ads providers do it (Facebook)

  • News feed do it (Prismatic, Pulse)

Advantages: pinpoint accuracy and relevance Inconvenient: operational complexity

amazon personalization
Figure 1. How is an Amazon page (old version!) personalized

   

2. Beyond behavior: tracking individual bodies

The relation between connected objects and personalization
Figure 2. The relation between connected objects and personalization

   

A list of bodily aspects being measured with examples:

Table 1. Location
Bodily Measurement Device Company Product Location
gmap location history

   

Table 2. Movement
Bodily Measurement Device Company Product Movement
fuelband

   

Table 3. Gestures
Bodily Measurement Device Company Product Gestures
myo

   

Table 4. Weight, heart rate
Bodily Measurement Device Company Product Weight, heart rate
withings

   

Table 5. Sleep
Bodily Measurement Device Company Product Sleep
aura

   

Table 6. Fingerprint
Bodily Measurement Device Company Product Fingerprint
fingerprint

   

Table 7. Facial recognition
Bodily Measurement Device Company Product Facial recognition
facial recognition

   

Table 8. Emotions
Bodily Measurement Device Company Product Emotions
crowdsight

   

   

Table 9. Behavior in public places
Bodily Measurement Device Company Product Behavior in public areas

Multiple devices

AGT International

Pedestrian traffic

Cameras

agt

   

A description of how AGT monitors large audiences in public events (click on the pic for the full document):

   

Video showing how Placemeter monitors pedestrian traffic:

   

3. The case of Nicholas Felton: constant data monitoring

a. The Feltron reports

nicholas felton3
Figure 4. Nicholas Felton

   

Nicholas Felton is a designer and data artist who produced printed annual reports from 2005 to 2014.

These reports synthesize his bodily data and social life, which he measures constantly during the year. This practice (pushed to the extreme in this case) belongs to the quantified self movement.

   

b. Not just Feltron

Insurance companies are interested in boosting individual health, using connected objects as monitoring devices

autodesk

Companies are looking to provide a 360 degree solution to health and well being through constant monitoring:

   

Monitoring on health is also a B2B market to achieve "corporate welfare". See Nokia’s brochure on the topic of health services.

4. Issues, limits

These technologies open a vast number of issues: from data privacy to the redefinition of well-being, and the grey boundary between monitoring and surveillance. A full session of this series is devoted to discussing these issues.

For the moment, let us just repeat cautionary remarks already mentioned in a different session:

a. "personalization" has been blamed for reinforcing "bubbles" or "tribes" views of the world

(paying version of the paper, free version here).

Content personalization is also blamed for favoring political polarization via an "echo chamber effect": social media tend to show me content I already agree with (paying version of the paper here, free version here).

b. Personalizing the customer relationship, even when effective, is not inherently a good thing.

It has been shown that the Coca-Cola #ShareaCoke campaign is effective at making more children choose a soda with a label to their name, over a healthy drink (paying version of the study here, free version not available).

c. Does personalization always need technology?

Companies rated with the customer service do personalization differently: with humans.

See how Zappos offers a great service to their customers:

   

or see (in French) how Trainline makes its customers happy.

The end

Find references for this lesson, and other lessons, here.

round portrait mini 150 This course is made by Clement Levallois.

Discover my other courses in data / tech for business: https://www.clementlevallois.net

Or get in touch via Twitter: @seinecle

site
    stats