Post on 18-Jan-2017
1
How to make your product go viralDimitry Rotstein
August, 2016
2
1
11/
0
kkNtN
Tt
Nuclear chain reaction
Where: N – number of atoms T – reaction cycle duration n – number of fission parts p – hit probability
pnk
3
1
11/
0
kkNtN
Tt
Viral chain reaction
pnk
Where: N – number of users T – viral cycle duration n – avg. number of shares p – adoption probability
4
k>1
k=1k<1
Significance of the k-factor
Ntotal
t
N(t)
totalNtNnptk 10
5
k>1
k=1
k<1
Ntotal
N(t)
t
totalNtNnptk 10
Significance of the k-factor
6
Nmax /Ntotal
k
Optimally, 1.5 ≤ k ≤ 2.5
1.0
0.5
1 2 3
7
dN/dt
t
Day-to-day growth
TtkNtn /0
k>1
k=1
k<1
Viral loop/protocol
8
4. Persuasion (L→A)
1. Motivation
3. Contact
2. Selection (TA,Ch)
9
Viral loop/protocol (realistic)1. Motivation 2. Selection (TA,Ch)
3. Contact 4. Persuasion (L→A)
10
PCBobSM PPNPPk
Viral protocol quantified
Where:
PX – probability of passing stage X of the protocol
NBob – number of Bobs selected by average Alice
11
Sharing creates/increases value for youM1. Natural L
? מיניין
12
Sharing/on-boarding gets you a knick-knackM2. Synthetic L
1+1 Group discount Friend brings a friend Good luck chain letters
13
M2. Synthetic - example
You are pressured to share (NOT use)M3. Coercion
14
L
E
15
M3. Coercion - examples L
E
Bad luck chain letters “Maximum repost” “No soup for you” Public shaming Soft peer pressure
16
Share to help someone else (including the product itself)M4. Altruism L
E
17
Share to say “Look at what I got/know”, elevate social statusM5. Vanity L
E
“How good are you in X” quiz
18
Share product/content just because it’s amazing or exciting
M6. Wow-factor
Copyright © New Line Productions Inc.
L
E
19
Share product/content just because it’s amusing/entertainingM7. Fun-factor L
E
“What kind of X are you” quiz
20
Share warm, fuzzy feelings (adorable, touching, nostalgic, etc.)M8. Happy L
E
“80s cartoons” quiz
21
Share “bad” feelings (fear, sadness, anger, disgust, envy, etc.)M9. Unhappy E
Copyright © Disney Enterprises Inc./Pixar
22
Share just by using the product (perhaps without realizing it)M10. Parasitic
Copyright © Universal City Studios Inc.
U
23
Save time and get your email on the go with the Yahoo Mail appGet the beautifully designed, lighting fast, and easy-to-use Yahoo Mail today. Now you can access all your inboxes (Gmail, Outlook, AOL and more) in one place. Never delete an email again with 1000GB of free cloud storage.[Learn more] [Try it now]
The message was checked by ESET NOD32 Antivirus.http://www.eset.com
Share just by using the product (perhaps without realizing it)M10. Parasitic - examples U
“Share if you agree”
24
Anti-motivator
S V U C N W F H A P
Fear
25
x
xxaMM
Motivation formula
• Motivators are not created equal• Motivators are independent of each other
PUHFWVACSNx ,,,,,,,,,
26
“Word of Mouth” study,Ernst Dichter (1966)
= W
= V
= A
= W + F + A
27
Get the word out about causes or brands (84%) = A
Grow and nourish our relationships (78%) = A + W + F + H
Self-fulfillment (69%) = A + V
Define ourselves to others (68%) = V
Bring valuable and entertaining content to others (49%) = W + F
Survey of 2500 heavy sharers,“New York Times” (c. 2010)
28
F = 30-40%W = 25-30%H = 10-15%U = 5-10%
Analysis of 100 million articles,BuzzSumo & OKDork (2014)
29
Mx=[0,1] Description ax
Natural Create|increase product value for the sharer 4|2
Synthetic Reward sharing by artificial knick-knacks 2
Coercion Force to share via hard|soft pressure 3|1.5
Altruism Help others solve acute|ambient problem 3|1.5
Vanity Enable bragging 2
Wow Amaze|impress by functionality or UX 3|1.5
Funny Amuse|entertain by product or content 3|1.5
Happy Induce fuzzy feeling (touching, nostalgic, etc.) 2
Unhappy Induce anger, envy, disgust, fear, sadness, etc. 1
Parasitic Automate sharing
Motivators summary8
30
Objective Vote (M=A+v+w+U=6.5)
Mar 2015
Election
Tota
l vis
itors
http://objectivevote.org.il/?page=platforms&lang=he
31
More examples
8.5 (N+A+w)
4.0 (N)
3.0 (A)3.5 (n+f)
7.0 (N+A)
32
M
PM
Motivation probability (PM)
4 100
1
33
Motivation probability - tips
Spell out motivation explicitly
AwarenessMPM
4tan 1
Remind to share at suitable time
Run A/B tests
4 < M < 10
34
A/B testing
<?php if(rand()%2): ?> <div>Invite your friends to have fun</div> <script> var ab=1; </script><?php else: ?> <div>Share your awesome results with your friends</div> <script> var ab=0; </script><?php endif; ?>
Example: vary motivator awareness (for some game)
Note: A/B testing must always be randomized
35
Selection probability and size (PSNBob)
ChTAMPS M – MotivationTA – Target Audience understandingCh – Marketing Channels awareness
Increase M Define target audience explicitly List channels (social networks, forums, email, phone, face-to-face, etc.) Run A/B tests
36
Selection probability – channel listing Share/like bar Share/like buttons Email invites Custom/embeded URL
37
Contact probability (PC)
TotalTotal
LeadsC N
NNN
FilterMP 1
M – MotivationFilter – Probability of being blocked
Increase M Use filter-proof services, e.g. MailChimp Prefer channels with less filtering Tighten the Target Audience OR Address a very common problem
38
Persuasion probability (PP)2MessageTrustMPP
Increase M Optimize Target Audience for Trust Optimize Message Supply the Message explicitly/directly Run A/B tests
Message = Simple+Memorable+Interesting
39
<head> ... <meta property='og:url' content='url' > <meta property='og:type' content='website'> <meta property='og:title' content='line1' > <meta property='og:description' content='line2' > <meta property='og:image' content='img_url'> <meta property='fb:app_id' content='app_id' ></head>
Supply the message via Facebook Preview
Then test/refresh values using Open Graph Debugger: https://developers.facebook.com/tools/debug/og/object/
Metadata for the Facebook Crawler:
40
Bad example (“Objective Vote”)http://objectivevote.org.il/?page=platforms&lang=he
41
Good example (“Objective Conflict”)http://objectivevote.org.il/?page=conflict&lang=he
42
Measuring k-factor analytically
tTkNkNn Tt
lnlnlnln 0/
0
ln(n)
t
ln(k)/T
TtkNtn /0
43
ln(n)
t
A real world example (“Objective Vote”)
Problems: Different channels Need to know ‘T’ Need to track individual referrals (MS)
44
Measuring k-factor directly
Ask for referrals: How did you learn about us? 1. Google 2. Bing 3. Facebook 4. Other: ___________
Identify referrals by unique coupon ID
Embed referral key into URL…
45
URL-embedded referral key
<div class='fb-like' data-share='true' data-width='350' data-show-faces='true' data-ref='num' data-href='url-to-like[?id=xxx]' style='height:80px;margin-top:10px'> </div>
FB ‘like’ button with referral key:
Disadvantage: no social validation (disconnected “likes”)Solution: random polling
See more at: https://developers.facebook.com/
46
Everything put together
LUCKMessageTrustNN
NN
FilterChTAAwarenessMMk
TotalTotal
Leads
2
213 14tan
47
Contact me on:
Dimitry Rotstein