Startup Metrics for Pirates: AARRR! Dave McClure, Master of 500 Hats @ SXSW, March 2008 blog: http://500hats. typepad .com/ website: http://www.500hats.com/ slides: http: //slideshare .net/dmc500hats/ * reminder: National Talk Like a Pirate Day is Sept 19 th !

Panel o’ PIRATES! Dave McClure , 500 Hats (moderator) Ted Rheingold , Dogster Todd Vernon , Lijit Hiten Shah , CrazyEgg Lance Tokuda , RockYou

Dave McClure , 500 Hats (moderator)

Ted Rheingold , Dogster

Todd Vernon , Lijit

Hiten Shah , CrazyEgg

Lance Tokuda , RockYou

AARRR genda Pirate Warmups : EVERYONE! Startup Metrics : Pirate Dave A cquisition : Pirate Ted A ctivation : Pirate Todd R etention : Pirate Hiten R eferral : Pirate Lance R evenue : (sorry, this one’s all you, landlubber)

Pirate Warmups : EVERYONE!

Startup Metrics : Pirate Dave

A cquisition : Pirate Ted

A ctivation : Pirate Todd

R etention : Pirate Hiten

R eferral : Pirate Lance

R evenue : (sorry, this one’s all you, landlubber)

Pirate Warmups Ship Ahoy! Yo-ho-ho, Yo-ho-ho! AARRR ! More tips: TalkLikeAPirate.com

Ship Ahoy!

Yo-ho-ho, Yo-ho-ho!

AARRR !

More tips: TalkLikeAPirate.com

Ship Ahoy!

Yo Ho Ho! Yo Ho Ho!

AARRR!

Why You Should Be A Web 2.0 Entrepreneur Pirate! AARRR! BOOTY!

Startup Metrics for Pirates: 5 Steps to Web 2.0 Booty ! A cquisition : users come to site from various channels A ctivation : users enjoy 1 st visit: " happy ” experience R etention : users come back , visit site multiple times R eferral : users like product enough to refer others R evenue : users conduct some monetization behavior AARRR !

A cquisition : users come to site from various channels

A ctivation : users enjoy 1 st visit: " happy ” experience

R etention : users come back , visit site multiple times

R eferral : users like product enough to refer others

R evenue : users conduct some monetization behavior

Customer Lifecycle & Conversion Behavior Website.com 4. REFERRAL Emails & widgets Campaigns, Contests 5. Revenue $$$ Biz Dev Ads, Lead Gen, Subscriptions, etc 2. Activation Homepage / Landing Page Product Features 1. ACQUISITION SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Q: Business Model? “ Business Model” can be one of the following: Get Users (= Acquisition, Referral) Drive Usage (= Activation, Retention) Make Money (= Revenue * ) * ideally profitable revenue Note: *eventually* need to turn Users/Usage -> Money

“ Business Model” can be one of the following:

Get Users (= Acquisition, Referral)

Drive Usage (= Activation, Retention)

Make Money (= Revenue * )

* ideally profitable revenue

Note: *eventually* need to turn Users/Usage -> Money

Founder/CEO Q: Which Metrics? Why? A: Focus on Critical Few Actionable Metrics (if you don’t use metric to make a decision, it’s not actionable) Hypothesize Customer Lifecycle & Refine Choose ~5 Conversion steps (tip: Less = More) Delegate Each Key Metric to someone to OWN

Q: Which Metrics? Why?

A: Focus on Critical Few Actionable Metrics

(if you don’t use metric to make a decision, it’s not actionable)

Hypothesize Customer Lifecycle & Refine

Choose ~5 Conversion steps (tip: Less = More)

Delegate Each Key Metric to someone to OWN

Product / Engineering Q: What to Build? Why? A: Build Features that Increase Conversion Wireframes = Conversion Steps Measure, A/B Test, Iterate FAST (daily/weekly) Optimize for Conversion Improvement 80% on existing feature optimization 20% on new feature development

Q: What to Build? Why?

A: Build Features that Increase Conversion

Wireframes = Conversion Steps

Measure, A/B Test, Iterate FAST (daily/weekly)

Optimize for Conversion Improvement

80% on existing feature optimization

20% on new feature development

Marketing / Sales Q: What channels? Which users? Why? A: High Volume ( # ), Low Cost ( $ ), High Conversion ( % ) Design & Test Multiple Marketing Channels + Campaigns Select & Focus on Best-Performing Channels & Themes Optimize for *deeper* conversion, not just site/landing page Low-Hanging Fruit: 1) Blogs, 2) SEO/SEM, 3) Landing Pages, 4) Automated Emails

Q: What channels? Which users? Why?

A: High Volume ( # ), Low Cost ( $ ), High Conversion ( % )

Design & Test Multiple Marketing Channels + Campaigns

Select & Focus on Best-Performing Channels & Themes

Optimize for *deeper* conversion, not just site/landing page

Low-Hanging Fruit:

1) Blogs, 2) SEO/SEM, 3) Landing Pages, 4) Automated Emails

Customer Lifecycle / Conversion Behavior Website.com Marketing Channels: largest-volume (#) lowest-cost ($) best-performing (%) SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Marketing Channels:

largest-volume (#)

lowest-cost ($)

best-performing (%)

Customer Lifecycle / Conversion Behavior Website.com Activation Criteria: 10-30+ seconds 2-3+ page views 3-5+ clicks 1 key feature usage do LOTS of landing page tests & A/B tests – make lots of dumb guesses & iterate FAST 2. Activation Homepage / Landing Page Product Features 1. ACQUISITION SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Activation Criteria:

10-30+ seconds

2-3+ page views

3-5+ clicks

1 key feature usage

Customer Lifecycle / Conversion Behavior Website.com Automated emails are simple, easy retention feature (but don’t overdo it) lifecycle emails @ +3, +7, +30d status emails weekly/monthly event-based emails as they occur but: make it easy to unsubscribe 2. Activation Homepage / Landing Page Product Features 1. ACQUISITION SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Automated emails are simple, easy retention feature (but don’t overdo it)

lifecycle emails @ +3, +7, +30d

status emails weekly/monthly

event-based emails as they occur

but: make it easy to unsubscribe

Customer Lifecycle / Conversion Behavior Website.com Encourage users to refer * after * they have “ happy” user experience; avg score >= 8 out of 10 2. Activation Homepage / Landing Page Product Features 4. REFERRAL Emails & widgets Campaigns, Contests 1. ACQUISITION SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Customer Lifecycle / Conversion Behavior Website.com This is the part * you * have to figure out… I don’t know jack about your business 5. Revenue $$$ Biz Dev Ads, Lead Gen, Subscriptions, etc 2. Activation Homepage / Landing Page Product Features 4. REFERRAL Emails & widgets Campaigns, Contests 1. ACQUISITION SEO SEM Apps & Widgets Affiliates Email PR Biz Dev Campaigns, Contests Direct, Tel, TV Social Networks Blogs Domains

Example Conversion Metrics (note: *not* actuals … your mileage may vary.) Category Conversion Status Conv % Est. Value Acquisition Visit Site (or landing page, or external widget) 100% $.01 Acquisition Doesn't Abandon (views 2+ pages, stays 10+ sec, 2+ clicks) 70% $.05 Activation Happy 1 st Visit (views X pages, stays Y sec, Z clicks) 30% $.25 Activation Email/Blog/RSS/Widget/Acct Signup (anything that could lead to repeat visit) 5% $2 Retention Length of Session / # of Clicks (length/intensity of engaged visit, >180s) 10% $1 Retention Email Open/ RSS view -> Click/Repeat Visit (3+ visits in first 30 days) 3% $5 Referral Refer 1+ users who visit site 2% $1 Referral Refer 10+ users who activate 0.2% $10 Revenue User generates minimum revenue 2% $5 Revenue User generates break-even revenue 1% $25

M AARRR keting Plan Marketing Plan = F(Customer Acquisition Channels, Campaign Themes) 3 Important Factors = Volume ( # ), Cost ( $ ), Conversion ( % ) Measure conversion to target customer actions Test audience segments, campaigns, Call-To-Action ( CTAs ) Match Channel Costs to Revenue Potential Est. Conversion + Revenue Potential Avg Txn Value ( ATV ), Ann. Rev. Per User ( ARPU ), Cust. Lifetime Value ( CLV ) Design channels that cost <20-50% of target ATV, ARPU, or CLV Consider Costs & Resource Tradeoffs Actual expenses Marketing time & resources Product/Engineering time & resources Timing of expense vs. revenue, profit vs. cashflow

Marketing Plan = F(Customer Acquisition Channels, Campaign Themes)

3 Important Factors = Volume ( # ), Cost ( $ ), Conversion ( % )

Measure conversion to target customer actions

Test audience segments, campaigns, Call-To-Action ( CTAs )

Match Channel Costs to Revenue Potential

Est. Conversion + Revenue Potential

Avg Txn Value ( ATV ), Ann. Rev. Per User ( ARPU ), Cust. Lifetime Value ( CLV )

Design channels that cost <20-50% of target ATV, ARPU, or CLV

Consider Costs & Resource Tradeoffs

Actual expenses

Marketing time & resources

Product/Engineering time & resources

Timing of expense vs. revenue, profit vs. cashflow

Example Marketing Channels disclaimer: estimates of vol, cost/user, time & effort are subjective – actual costs are dependent on your specific business

AARRR genda Acquisition : Pirate Ted, Ship Dogster/Catster

Acquisition : Pirate Ted, Ship Dogster/Catster

Marketing Initiative Feb ‘07 Goal: 2,000 New Registrants a Day Marketed sites via AdSense and BlogAds 1st - Linked ads to our homepages 2nd - made a general landing page with intro info and big link to register 3rd - made dynamic landing page with relevant text and registration form.

Marketed sites via AdSense and BlogAds

1st - Linked ads to our homepages

2nd - made a general landing page with intro info and big link to register

3rd - made dynamic landing page with relevant text and registration form.

Meeee…. AARRR !

LANDING PAGE Daily Candy

LADING PAGE Dog Info

A Registrant is Not Always a Member We started getting to 2,000 registrants a day. BUT only 60% were activating.

We started getting to 2,000 registrants a day.

BUT only 60% were activating.

REGISTRATION PAGE -Activate

Activants Aren’t Always Members Either Of those new members only 8% were making a pet page vs. 80% for people who found our sites naturally. Within 3 months we halted all ad spending.

Of those new members only 8% were making a pet page vs. 80% for people who found our sites naturally.

Within 3 months we halted all ad spending.

AARRR !…(F)

HOMEPAGE - Current

CRAZY EGG Referring sites

CRAZY EGG Searchers

HOMEPAGE - Current

Tracking Tools Google Analytics (Bounce rate, Filters, Goals/funnels) CrazyEgg (Click tracking from all types of visitors) Common SEO tools (Keyword density, Spider Simulator, Search engine ranking tool) Internal tools (anything we can think could be interesting, we track)

Google Analytics (Bounce rate, Filters, Goals/funnels)

CrazyEgg (Click tracking from all types of visitors)

Common SEO tools (Keyword density, Spider Simulator, Search engine ranking tool)

Internal tools (anything we can think could be interesting, we track)

AARRR genda Activation : Pirate Todd, Ship Lijit

Activation : Pirate Todd, Ship Lijit

Old Signup Process Enter email, user name Enter blog URL Enter account names for other user generated content (18) Enter URL’s of other blogs in your network Do the one click Typepad Install. Signup Time ~ 15-20 min Conversion Rate ~ 0.1%

Enter email, user name

Enter blog URL

Enter account names for other user generated content (18)

Enter URL’s of other blogs in your network

Do the one click Typepad Install.

New Process – Automate almost everything Use automation to remove over 95% of required tasks. Make the process entertaining and a little mysterious Create a situation where we reward the user at each step, winning one more click. Don’t require any typing until the last step. Signup Time 15m -> 2m Conversion 0.1% -> 4%

Use automation to remove over 95% of required tasks.

Make the process entertaining and a little mysterious

Create a situation where we reward the user at each step, winning one more click.

Don’t require any typing until the last step.

Signup Time 15m -> 2m

Conversion 0.1% -> 4%

Change #1: Used MBL Widget to discover Blog URL User hits site and blog URL is automatically entered into input field using JS call to MBL Widget (running on our homepage)

User hits site and blog URL is automatically entered into input field using JS call to MBL Widget (running on our homepage)

Change #2: Use Blog URL to inspect blog Lijit Networks, Inc. Extract commonly used account name(s) from Widgets Read Blogroll – Use for default social network

Extract commonly used account name(s) from Widgets

Read Blogroll – Use for default social network

Change #3 : Find other User Generated Content Using discovered Acct name(s), find other content AJAX Screen entertains prospect during discovery

Using discovered Acct name(s), find other content

AJAX Screen entertains prospect during discovery

Validation Signup Time 15m -> ~2m Conversion 0.1% -> ~4% (40x) Users Comments “ Easy to install” “ Took me only a few minutes” “ You’re freaking me out, how did you get my blog URL.” http://ouriel.typepad.com/myblog/2007/08/execution-is--1.html

Signup Time 15m -> ~2m

Conversion 0.1% -> ~4% (40x)

Users Comments

“ Easy to install”

“ Took me only a few minutes”

“ You’re freaking me out, how did you get my blog URL.”

AARRR genda Retention : Pirate Hiten, Ship CrazyEgg

Retention : Pirate Hiten, Ship CrazyEgg

Retention

Before

Before

After

After

Results Before After

AARRR genda Referral : Pirate Lance, Ship RockYou

Referral : Pirate Lance, Ship RockYou

AAR R R! R for Referral

“ To engage the world through social applications” The RockYou Mission

“ To engage the world through social applications”

Source: eBay Investor Presentation, RockYou eBay starts Q2 98, PayPal starts Q1’00, Yahoo! starts Q1’95, AOL starts Q1’92, Facebook starts Q4’04, and RockYou Starts Q4’05 Facebook data represents active users, which was disclosed on 12/05 and 12/06. Undisclosed active user data is extrapolated by applying an average active user penetration to global Unique Visitors (per comScore Media Metrix) (2) Registered Users (MM) RockYou Growth: Fastest Ever. Facebook platform launch

Source: eBay Investor Presentation, RockYou

eBay starts Q2 98, PayPal starts Q1’00, Yahoo! starts Q1’95, AOL starts Q1’92, Facebook starts Q4’04, and RockYou Starts Q4’05

Facebook data represents active users, which was disclosed on 12/05 and 12/06. Undisclosed active user data is extrapolated by applying an average active user penetration to global Unique Visitors (per comScore Media Metrix)

Referral traffic is key for viral growth Primary sources of referral traffic Install flow Engagement flow

Primary sources of referral traffic

Install flow

Engagement flow

Install flow example for Bumper Sticker

Engagement flow example for Friends for Sale

Metrics for viral loop User Call to action to invite friends x = invited friends Accept? No Yes y% = accept rate x * y > 1 gives you viral growth!

Typical viral numbers Install flow x = 5 (friends invited on average) y = 22% (acceptance rate for invites) Viral factor = 5 * 0.22 = 1.1 VIRAL!!! Engagement flow x = 20 (friends invited on average) y = 6% (acceptance rate for invites) Viral factor = 20 * 0.6 = 1.2 VIRAL!!!

Install flow

x = 5 (friends invited on average)

y = 22% (acceptance rate for invites)

Viral factor = 5 * 0.22 = 1.1 VIRAL!!!

Engagement flow

x = 20 (friends invited on average)

y = 6% (acceptance rate for invites)

Viral factor = 20 * 0.6 = 1.2 VIRAL!!!

Use multiple flows and channels Install flow x = 5 y = 10% Viral factor = 5 * 0.1 = 0.5 Engagement flow, invite channel x = 3 (invites) y = 10% (acceptance rate for invites) Viral factor = 3 * 0.1 = 0.3 Engagement flow, notification channel x = 6 (notifications) y = 5% (acceptance rate for notifications) Viral factor = 6 * 0.05 = 0.3 0.5 + 0.3 + 0.3 = 1.1 VIRAL!!!

Install flow

x = 5

y = 10%

Viral factor = 5 * 0.1 = 0.5

Engagement flow, invite channel

x = 3 (invites)

y = 10% (acceptance rate for invites)

Viral factor = 3 * 0.1 = 0.3

Engagement flow, notification channel

x = 6 (notifications)

y = 5% (acceptance rate for notifications)

Viral factor = 6 * 0.05 = 0.3

Important tips Install flow should generate most of your referrals everyone installing will use that flow so make it good user test different calls to actions if all else fails, offer $10,000 for invites =) Engagement flows supplement virality not everyone will engage, generally weaker than install flow consider having the primary flow invite new friends user test different calls to actions if all else fails, offer $20,000! Decay invite rate decreases as you saturate acceptance rate decreases as you saturate engagement tends to decay continuous improvement required!

Install flow should generate most of your referrals

everyone installing will use that flow so make it good

user test different calls to actions

if all else fails, offer $10,000 for invites =)

Engagement flows supplement virality

not everyone will engage, generally weaker than install flow

consider having the primary flow invite new friends

user test different calls to actions

if all else fails, offer $20,000!

Decay

invite rate decreases as you saturate

acceptance rate decreases as you saturate

engagement tends to decay

continuous improvement required!

AARRR! Any Questions, Ye Scurvy Dogs? Additional References: “ Influence: The Psychology of Persuasion ”, Robert Cialdini (book) “ Putting the Fun in Functional ”, Amy Jo Kim (etech 2006 preso) “ Futuristic Play ”, Andrew Chen (blog) “ Don’t Make Me Think ”, Steve Krug (book) “ A Theory of Fun ”, Raph Koster (book, website)

Additional References:

“ Influence: The Psychology of Persuasion ”, Robert Cialdini (book)

“ Putting the Fun in Functional ”, Amy Jo Kim (etech 2006 preso)

“ Futuristic Play ”, Andrew Chen (blog)

“ Don’t Make Me Think ”, Steve Krug (book)

“ A Theory of Fun ”, Raph Koster (book, website)