The Science behind Viral marketing

The Science behindViral Marketing

Lessons Learned

Explosive growth!

YouTube Facebook Zynga Groupon Twitter Gilt Groupe

All have set records for growth rates

The key is Virality

The Power of Viral Marketing

When it works it is Free!! Far more effective than paid promotions and

advertising campaigns If done well can be self-sustaining + growth Unfortunately rarely achieved!

The Best Business Model

Monetization(LTV)

Cost ofCustomer

Acquisition(CoCA)

Virally acquired customers are free

Topics

The Science behind Viral Marketing How to do Viral Marketing Optimizing for Virality Making this work for your business Examples

Where my experiences come from

Tabblo Photosharing site Founded by Antonio Rodriguez, 2005 Acquired by HP within 18 months Great investment return

All credit is owed to Antonio Rodriguez

The Viral Loop

Customer first sees your app

Tries it

Decides they like it enough to invite their

friends

Creates an invitation, looks up

addresses, and sends

Friend receives

invitation

Friend decides to take a look

The Math behind Virality

Key variables: Custs(0) The initial set of customers i The number of invites sent out conv% The percentage of invites that

convert into customers

The Viral Coefficient

K The Viral Coefficient

K = no of invites x The conversion % (i x conv%)

Turns out to be a very important variable. It equals the number of new customers that each customer is able to successfully invite.

An Example

Custs(0) = 5 i = 10 conv% = 20% K = 2

Cycles 0 1 2 3 4 5

New custs added this cycle 10

20

40

80

160

Total Customers: C(c) 5

15

35

75

155

315

The Formula

New custs added in this cycle = New custs added in the previous cycle x K

Assumes that customers only send out invites once, and not continuously in every later cycle

To calculate the number of customers at any particular cycle (c):

Custs(c) = Custs(c-1) + NewCusts(c)

Sensitivity to K

See Spreadsheet model

Sensitivity to K

Viral Coeffi cient K 0.5 Cycles 0 1 2 3 4 5 6 7 8 9 10 11 12New custs added this cycle 3 1 1 0 0 0 0 0 0 0 0 0 Total Customers: C(c) 5 8 9 9 10 10 10 10 10 10 10 10 10

Viral Coeffi cient K 0.9 Cycles 0 1 2 3 4 5 6 7 8 9 10 11 12New custs added this cycle 5 4 4 3 3 3 2 2 2 2 2 1 Total Customers: C(c) 5 10 14 17 20 23 26 28 31 33 34 36 37

Viral Coeffi cient K 1.0 Cycles 0 1 2 3 4 5 6 7 8 9 10 11 12New custs added this cycle 5 5 5 5 5 5 5 5 5 5 5 5 Total Customers: C(c) 5 10 15 20 25 30 35 40 45 50 55 60 65

Viral Coeffi cient K 2.0 Cycles 0 1 2 3 4 5 6 7 8 9 10 11 12New custs added this cycle 10 20 40 80 160 320 640 1,280 2,560 5,120 10,240 20,480 Total Customers: C(c) 5 15 35 75 155 315 635 1,275 2,555 5,115 10,235 20,475 40,955

What we learned

Viral Coefficient must be > 1 to have viral growth

Viral Growth is a compounding phenomenon

Increasing the Viral Coefficient has a big impact on the rate of growth

YouTube versus Tabblo

Both started around the same time YouTube’s growth massively outstripped Tabblo

What was happening here?

Tabblo’s Viral Loop

User discovered Tabblo Wait till they take some photos Upload Share (i.e. invite others) Some portion of those think, this is great, I want to use

this myself Wait till they take some photos Upload Etc.

YouTube’s Viral Loop

User discovers YouTube Sees some hilariously funny content Decides to share that with friends Friends see hilariously funny content Decide to share that with their friends

The time to infect is far shorter!

A More Sophisticated Formula

t = timect = cycle time

Thanks to Stan Reissfor help with the forumula

What this tells us

Multiplier

Raised to the Power of …

Shortening the cycle time has a far bigger effect than changing Viral coefficient!

Some experiments

See Spreadsheet model

Some Experiments

No wonder YouTube was explosive!

Viral Coeffi cient K 1.50

Custs(t) 161 0 10 20 30 40 50 60 70 80 90 100 110 120Viral Loop Time (lt) 1 5 855 49,870 2,876,336 165,866,278 9,564,781,592 551,559,020,225 31,805,990,505,278 1,834,112,025,928,330 105,765,199,266,334,000 6,099,015,336,964,180,000 351,703,474,664,229,000,000 20,281,197,415,788,400,000,000

2 5 104 855 6,558 49,870 378,766 2,876,336 21,842,355 165,866,278 1,259,553,409 9,564,781,592 72,632,923,440 551,559,020,225 5 5 24 66 161 374 855 1,936 4,369 9,843 22,159 49,870 112,220 252,507

10 5 13 24 41 66 104 161 246 374 567 855 1,287 1,936 20 5 8 13 18 24 31 41 52 66 83 104 130 161 50 5 6 8 9 11 13 14 16 19 21 24 27 30

Time tTime - t

Lesson Learned

Reducing Viral Loop cycle time has by far and away the greatest effect on viral growth

How to do Viral Marketing

The Basics

Figure out the Viral Hook What makes people want to share content?

Content or Product/Service Initial Seeding Use communications networks to spread

Email, Twitter, Facebook, Google+, etc.

Refresh & Re-seed

Figure out the Viral Hook

Something very compelling that the user wants to share

Viral Hook - What works:

Something of Value: Applications Educational content Data Things of monetary value (Discounts and coupons)

Something entertaining Humor Games

News Inherently Viral Services

Email, Skype, etc.

Other?

Why do people share?

A paper on that here: http://marketing.wharton.upenn.edu/documents/research/Virality.pdf http://www.nytimes.com/2010/02/09/science/09tier.html (summary)

Virality driven by physiological arousal High arousal, Postive (awe) – most likely to be shared High arousal, Negative (anger or anxiety) – less likely Low arousal, de-activating (sadness) – unlikely

Content Types

Blogs Videos eBooks Applications and Services Etc.

Using engineering for marketing

Inherently Viral Applications

The most viral apps are those that are require sharing to work properly: HotMail Skype YouTube FaceBook Etc.

Initial Seeding

The bigger your initial seeding the better Identify Influencers

Reach / Influence Klout, SocMetrics and other services emerging to help

Develop the right relationship Listen, Strategize, Engage, Measure

Paid Seeding Paid search, Virool.com, etc.

Spread through Communications Networks

Early days this was email Social Networks have turbo-charged sharing

What you want: Low effort for the user doing the sharing

Refreshing and Re-Seeding

Google Search (paid and organic) Creates new seeding

New Blog posts Cause your existing readers to re-invite

Farmville, etc. Continue to invite by regular posts to your wall

Lessons Learned Optimizing for Virality

Lessons Learned

Apply my “Unblock your Sales Funnel” principles to the viral loop steps

Customer first sees your app

Tries it

Decides they like it enough to invite their

friends

Creates an invitation, looks up

addresses, and sends

Friend receives

invitation

Friend decides to take a look

CONCERNS- Hate being sold to- Find it offensive to give

name and email- Don’t want to get spam

sales emails- Worried that email

address will be given to other marketers

GET INSIDE YOUR CUSTOMER’S HEAD

UNDERSTAND WHAT MOTIVATES THEM

- Want to solve my problem- Recommendation from a friend- Education- Data/ information reports- Entertainment- Free stuff- Meeting other people like me that

have insights to share

CONCERNS

MOTIVATIONS

CREATE A SOLUTION THAT ENTICES THEM AND ADDRESS THEIR CONCERNS

- Customer testimonials address vendor risk

- Free trials address product viability and fit concerns

- Lowest price guarantees

CONCERNS

ENTICE & ADDRESS CONCERNS

Example: Groupon

Example: Living Social

Another strategy

Offer a reward both to the user and to their friend Overcomes the feeling of being sleazy Can lead to much higher success rates

Lessons Learned

Maximize the number of invites they send out Leverage tools that provide easy reach to their friends:

Twitter Facebook: post to their News Stream LinkedIn

Automate access to their address book Works well for GMail, and other on-line address

books Etc.

Lessons Learned

Look for ways to automate the whole invitation process No thinking required No work required

Don’t just invite once

Look for ways to keep repeating invites E.g. Facebook games like Farmville that post to your

News Stream

Lessons Learned

Look for other ways to get your customers to tell their friends about you

Example: ConstantContact’s email signature

Reduce the Cycle Time

Customer first sees your app

Tries it

Decides they like it enough to invite their

friends

Creates an invitation, looks up

addresses, and sends

Friend receives

invitation

Friend decides to take a look

Lessons Learned

Hybrid Viral Even if you can’t get a viral coefficient > 1

you can still enhance customer acquisitionwith some virality

Lessons Learned

The Viral Loop is an important Funnel Design each step carefully

Simplify and eliminate manual steps Test for adequate motivations and incentives

Address concerns A/B test each step

Metrics to evaluate and improve

Make it work for your Business

Creepy Effective

What can go wrong? - Branchout

Social networks are for hanging out with friends Not promoting your business

Branchout (recruiting app on Facebook) Spammed users walls with way too many entries

The Result:

Thanks to paulamarttila.posterous.com

Some Fun Examples

For a full set, visit:http://www.ignitesocialmedia.com/social-media-examples/viral-marketing-examples/

For More information

Visit my blog at www.forEntrepreneurs.com