TalkTemplateNudgeAsiaHK2016vFinal

PREDICTING BEHAVIOUR
INTRODUCING ‘GAME ONTOLOGY’
PRESENTATION
AT HONG KONG GENERAL CHAMBER OF COMMERCE
MARCH 17 2016
Patrick A. McNutt FRSA
Visiting Fellow, Manchester Business School, UK
& Smurfit Business School, Dublin, Ireland.
www.patrickmcnutt.com
Follow @tuncnunc
PREMISE: PREDICTION DEPENDS
ON FINDING A
PATTERN
STEP 1. SEQUENCE
2,3,5,8,13,21,34…
FINITE ORDER [STRING]
STEP 2. DIGITAL SELF [REPEAT BEHAVIOUR]
STEP
3 CODING
BINARY 7 = 111
ALPHA-NUMERIC: WORD
= 9673
PREMISE:
APOPTOSIS [DEFINED FOR A DATA PATTERN IN A GAME]
A PROGRAMMED SEQUENCE OF CHOICE EVENTS LEADING TO THE
ELIMINATION OF INDIVIDUAL CHOICE
PATTERNS AS A FINITE ORDER [STRING]
CREATED BY DIGITAL SELF [REPEAT BEHAVIOUR]
AND BEHAVIOUR IS CODED
B
WHEN YOUR BEHAVIOUR ADAPTS
& ADOPTS
YOU ARE ENVELOPED BY A GAME
A GAME UNFOLDS WHEN THERE IS BOTH MUTUAL
INTERDEPENDENCE AND RIVAL INTERACTION
(.) PLAYS THE GAME (WINNING
STRATEGY ) OR (.) OBEYS A RULE (DOES NOT THINK)
=>
ABANDON UNILATERAL DECISION MAKING &
A RATIONAL INDIVIDUAL
AN INHERITED OR ACQUIRED GAME DNA
PREDICTION VS FORECASTING
PREDICTION
 Patterns
 Repeat Patterns….due to:
Avoid errors & Habit
Disbelief in fate
Possibility of ‘nudge-averse’
 Signals & Semi-structured
Data [tweets, newsfeeds]
=>
 Patterns and Rules
 Find the Rule.





FORECASTING
An ‘odds ratio’
What is the probability of
an event?
Republican in the White
House?
€ = 1US$?
Noise & Structured Data
[sales reports]
=>
 Patterns and Beliefs
 Influence the Belief
AT THE LEVEL OF INDIVIDUAL:
FIND THE RULE
+
INFLUENCE THE BELIEF
=
DECODING THE PATTERN
‘A BEHAVIOURAL PROCESS, A SEQUENCE OF ACTIONS AND
REACTIONS AT TIME PERIOD T WITH CONSEQUENCES AT TIME
PERIOD T+1’
LETS TAKE AN EXAMPLE: THE DAILY ROUTINE
PROBABILITY:
OBEY THE RULE.
IS IT INERTIA OR HABIT?
EXAMPLE: DRIVING ON LEFT OR RIGHT SIDE OF ROAD?
EXAMPLE: YOUR DAILY ROUTINE
Coffee stop 8.25
Texting, FB and
‘Googling’ the
weekend plans.
‘Latte to go’ NFC
payment at 832am
walk to office
At office Log in PC
845
NFC at coffee shop
825
Smartphone Wake
Up 7am
Leave the house
725am
Transit Smartcard at
745am
Check text messages
Social media
At desk: 845am Log
On PC, search web. &
texting
925 Receive text
message confirm
lunch
1230 for lunch, CC
payment 13.45
1730 Log off PC;
Returning home,
smart card 1814;
send text; at home by
1855 surf the net,
final texts, web
browsing.
Texting, surfing
‘lights out’ by 23.55
PROBABILITY:
OBEY THE RULE => DON’T THINK
=> DAILY ROUTINE
WHAT IF? POISSON SEQUENCE
OF EVENTS WITH
EXTERNAL NUDGE
[EXAMPLE: A PREFERRED COFFEE-HOUSE]
CHAIN OF EVENTS FOLLOW A PATTERN AS A RULE
OOOOO
Δ IS THE NUDGE PARAMETER
Patterns have a ‘game’ dimension - to make someone do
something they would not otherwise do





Δ is the nudge parameter ..it is an individual’s change
parameter
The future is t+1 and if the change from t to t+1 can be
framed as a loss/gain it creates a change (loss) aversion
effect.
Individuals learn from observing the behaviour of others.
Individuals also learn from private information …I thinkyou think-I think loop
DAILY ROUTINE ALGORITHM
PROBABILITY: OBEY THE RULE =>
DON’T THINK =>
DAILY ROUTINE
WHAT IF? POISSON SEQUENCE OF EVENTS
WITH
BESPOKE NUDGES [SPECIFIC COFFEE-HOUSE]
CHAIN OF EVENTS: INFLUENCE OR NUDGE THE PATTERN
POISSON SEQUENCE
1.YOU AND YOUR FRIEND’S DAILY ROUTINE ARE INDEPENDENT
EVENTS.
2.YOU ARE [EXTERNALLY] ‘NUDGED’
TOWARDS COSTA.
3. YOUR FRIEND PHONES YOU FOR LUNCH TOMORROW, YOU
PERSUADE HER TO GO TO COSTA. SHE GOES TO COSTA.
4.YOU AND YOUR FRIEND’S DAILY ROUTINE ARE NOW CODEPENDENT EVENTS.
SELF AS A PLAYER IN A GAME
Information is embedded in observed behaviour
Bt during the time continuum from t to t+1. .may
be due to inertia or habit.

Bt+1 = Bt + Δ

Δ is the nudge parameter that policy makers,
management, algorithms and strategists can
influence.
 RANDOMNESS-(NUDGE)-PATTERN-(CODE)SEQUENCE-PREDICTION
PROBABILITY: OBEY THE RULE (DON’T
THINK) = BEHAVIOUR CODED WITH A NUDGE
1 = COSTA, 0 = NO COSTA (SBUX)
(14 DAYS):
110101010100011
6,5,2,4,3
BECOMES
(14 DAYS):
111111000110000
7,7,0,6,0
111110101100011010001000
7,6,5,4,3,2,1,0
NUDGE……
….
NUDGE……
…….
NUDGE
PREDICTABLE
BEHAVIOUR
DAILY ROUTINE ALGORITHM
A PATTERN EVOLVES FROM MONDAY,
TUESDAY, WEDNESDAY, THURSDAY… THEN
FUNNEL INTO FRIDAY?
CAN WE PREDICT FRIDAY?
PATTERN EVOLVES WEEK 1,
WEEK 2, WEEK 3.. THEN FUNNEL INTO WEEK 4?
CAN WE PREDICT WEEK 4?
NUDGE……
….
NUDGE……
…….
NUDGE
PREDICTABLE
BEHAVIOUR
FIND THE PATTERN:
WHY?
IT IS THE COGNITIVE KEY THAT DECODES
BEHAVIOUR
BT+1 = BT + Δ
IN THE DAILY ROUTINE WE HAVE
PRE-8.25a.m
IN BUSINESS ALSO FIND THE PATTERN
IN SONY VS MICROSOFT WE HAVE
PRE-2004
PATTERN – 2000-2006
PS2
launched
at $299
PS2 at
$199.99
PS2 at
$179.99
14 May 02 13 May 03
26 Oct 00
PS2 at
$149.99
11 May 04
100 million
PS2
shipped
Announcement
PS3 production
schedule to ship 6
million units by 31
Mar 07 at $499
PS2 at
$129.99
20 April 06
1 Nov 05
8 May 06
15 Nov 01
15 May 02 14 May 03
29 Mar 04
22 Nov 05
22 million
Xbox
shipped
Microsoft Xbox
launched at $299
Xbox at
$199
30 Oct 05
Xbox at
$179
Xbox at
$149
Xbox 360
launched
at $399
6 Feb 06
27 April 06
Xbox at
$179
Revised production
schedule for Xbox
360 to 5- 5.5 million
units by 30th June
2006
OCTOBER 2000 – MAY 2004
OBSERVABLE PATTERN EX-POST
299..299..199.99..199..179.99..179..149.99..149
OPENING SYMMETRIC MOVE…….
…………..
50% PRICE DECREASE
REPEATED IN
2010 AND 2015
Cognitive Business Strategy

Cognitive dissonance:
Why strategy does not work?
Who is to blame?
More leaps than steps
LG Prada smartphone 2006
Yoghurt shampoo
DECODING STRATEGY
Decode Bt behaviour
 Untangle the type tapestry in The Daily Routine
as consumers outsource memory and leave smart
‘breadcrumbs’ or smart ‘footprints’.
 Audience/Group or Consumer Inertia
 Audience/Group or Consumer Habit

BUSINESS STRATEGY
EXAMPLES
 If Δ is inertia:
=>
Sunk-cost investment
Nudge-averse

Strategy ‘what-if?’ is to move before a competitor
Δ = price promotion [income response].
Cognitive Business Strategy

Cognitive association:
Why doesn’t it work?
What should I change to make it work?
Less leaps more steps
Apple iPhone 2007
Faber-Castell Smart Pencil
BUSINESS STRATEGY
EXAMPLES
 If Δ is habit:
=>
Network or group effects
Situation-specific

Strategy ‘if-then’ is to facilitate consumer’s choice
Δ = stripy toothpaste [asset response]
SIGNAL/NOISE RATIO
SIGNAL/NOISE > 1 [FIND THE RULE]
CONSUMERS ARE NUDGE-AVERSE (DO NOT
LAUNCH A NEW PRODUCT)
NUDGE STRATEGY [YOU CAN INFLUENCE BELIEF]
SIGNAL/NOISE < 1
CONSUMERS ARE NOT NUDGE-AVERSE (LAUNCH
A NEW PRODUCT)
SIGNAL/NOISE < 1 IN A TIME
CONTINUUM
Engage with consumers: don’t underestimate future
gains in t+1:
When individuals are asked to choose between a piece of
fruit and an unhealthy snack for a meal tomorrow, the
choice is influenced by (i) the preference today; (ii)
observing others as a norm.
WITH SIGNALING EVERYTHING IS
CHANGING
What should I do?
With all the chaotic information available ‘what
you are’ and ‘who you are’ defines your Daily
Routine and converts it into information = a
valuable tradable asset, so
Stop and Think.
Know Your Pattern
Observe but do not judge.
PROBLEM:
HOW TO BRING A
5 FEET LONG FISHING ROD ONTO A LONDON BUS
REGULATION HEIGHT = 4 FEET
SOLUTION:
FIND A WOODEN BOX OF DIMENSIONS 3X4
INSERT ROD ACROSS THE DIAGONAL
BRING BOX ON BOARD THE BUS
WITH ROD INSIDE THE BOX
Game Ontology
At 7am the choice is between: sunglasses or an umbrella?
Game theory is observational. People filter signals and would
act or behave accordingly. But what if an external influence
could be introduced into the filtering process? The purpose is
to nudge behaviour. In 1950s there was subliminal
advertising and in 2016 there are smart nudges – cognitive
business strategy.
The pattern is a repeated action but what is the action? Is it
voluntary or habit? What is the metaphysics behind an action
or the philosophy of conduct? An answer requires the receipt
of intelligence, the appreciation of the situation, the
invocation of principles, the planning, the execution of the
action and the excuse provided for not doing an action at a
moment in time.
Poisson sequence:
Suppose the coffee house already knows by decoding
patterns that on average 60 customers walk by 12301330 each day => one customer arrives per minute. In
order to find the probability that exactly two customers
arrive in a given one-minute time interval between
1230-1330:
Let µ = 1 and x = 2:
P(2) =e-1/2!=0.3679÷2 = 0.1839
The number of times a smart card is used at 1230-1330
has an average rate of 10 times per hour then the coffee
house would like to know (i) the probability no more
than 10 times occur at the 1 hour lunchtime break and
(ii) more importantly, the probability that the number of
times will exceed 10 during the 1 hour lunch break:
P(x≤10) = Σ10i · e-10 ÷ i! = 0.58
(i = 0….10)
Thank you for listening………
‘Habit is a great
deadner’
Samuel Beckett
Waiting for Godot Act II