Presentation given as Flash Talk at Automattic Meetup in Seaside on September 2010
Presentation is given as an Ignite Talk format (20 slides x 15 seconds/slide = 5 minutes. Autopush.)
Automattic is the company I work for. The company is distributed worldwide and once a year we gather at a remote location and meet face-to-face. This year, all the employees are taking a little time during the meetup to compose and give at least one presentation for each other, talking about any subject we are passionate about.
In the e-mail requesting submissions, Matt mentioned that <a>Scott Berkun “did a very cool post and video on giving ignite talks</a>, so I modelled this talk after that.
Hope you enjoy it.
Cybersecurity Awareness Training Presentation v2024.03
Tales of Virality—Automattic
1. Tales of Virality
Terry Chay
Automattic Meetup (Ignite)
September 11, 2010, 10:00 AM
Saturday, September 11, 2010
The last two companies I worked for were Tagged and Plaxo. Both sites are based on virality.
When people ask me why I went to work at Automattic, I tell them, “I’m tired of working for
evil. I decided to give good a shot.”
2. Saturday, September 11, 2010
A former CEO once said, “Some companies are born lucky, and some companies have to make
their luck.” Companies like Hotmail, Facebook, and WordPress were “born lucky” and
companies like Plaxo, LinkedIn, Tagged, Zynga make their luck. I'll explain what this means
later.
[Friendster,Twitter]
[Yelp, RockYou, Slide…]
Source: http://www.marieclaire.com/cm/marieclaire/images/poker_hand.jpg
http://whovotedhow.com/wp-content/uploads/2009/10/card_up_sleeve.jpg
3. Mousetrap
Saturday, September 11, 2010
But first let's recall the days before the internet. In those days you first built a better
mousetrap, then people beat a path to your door, and then you used marketing to protect
that position.
Source: http://ontopinternetmarketing.com/wp-content/uploads/2010/09/mousetrap.jpg
4. Saturday, September 11, 2010
For instance, if we were to look at the advertising budgets of Apple and Microsoft, we’d find
that they spend more now on marketing now than they did back when they first started. Even
as a fraction of income!
[GM]
5. Web 1.0
1. Make Superbowl Commercial
2. ???
3. Profit!
Saturday, September 11, 2010
But that became the “Old” economy. In the DotCom boom, the marketing cart came before
the product horse. The “New” Economy basically meant that you followed the Underpants
Gnomes path to success:
[Or New New Economy]
6. Saturday, September 11, 2010
Step 1) Steal underpants (I mean make superbowl commercial), Step 2)… Step 3) Profit! It got
to the point where you were giving free shipping on 40lb bags of dogfood! (just to acquire a
customer)
Source: http://www.stefanoparis.com/stefmedia/petsdotcom/sockpuppetlove/
puppetlovehalf.jpg
http://www.youtube.com/watch?v=sICSyC9u5iI
7. Web 2.0
Saturday, September 11, 2010
For some strange reason that model failed. And VC would no longer pony up millions to put
the up a Superbowl commercial featuring a sock puppet. But what to do about marketing?
How do you get users? Well we had this concept of social networks so…
8. Viral Marketing
Saturday, September 11, 2010
Viral marketing. People often use viral marketing and “word of mouth” marketing
interchangeably. But they are not the same. For one, “word of mouth” has been around
forever. So what is the difference for reals?
[They are as different as i18n and l10n in g11n.]
Source: http://www.socialsignal.com/system/files/images/mouth-to-ear-
whisper.hallmark.jpg
9. Viral Code
function virality(&$x) {
++$x;
}
do {
virality($users);
} while (true);
Saturday, September 11, 2010
Viral marketing means the very act of users using the site brings more users to the site.
[Pause or repeat]. The best example is the first one: Hotmail.
10. Saturday, September 11, 2010
When people use Hotmail to send an e-mail… at the bottom of the e-mail was a signature:
“Want free e-mail? Try Hotmail.” It grew like hotcakes, bled money, and sold to Microsoft for
just shy of half a billion dollars back when that amount was worth something…
Source: http://farm2.static.flickr.com/1290/4690702547_d9d5000e48_m.jpg
11. Hotmail (and Youtube)
Strategery
1. Make a Viral Product
2. …
3. Profit???
Saturday, September 11, 2010
…and Microsoft would have paid far more for it. Okay! this doesn’t exactly solve your
Underpants Gnomes problem (unless you are Hotmail) but it does solve the marketing one.
How does it work? The key is to realize…
12. Saturday, September 11, 2010
…unlike other forms of marketing, every step in viral marketing can be measured. The page
views are measurable. If the user signups are measurable.The contacts they import are
measurable. Even if the user views the e-mail is measurable!
13. Viral Coefficient (v)
v<1 v≥1
Saturday, September 11, 2010
When you crunch the math you get two curves: Either user growth looks like the first curve,
or it looks like the second. The difference is the viral coefficient. Remember the born lucky
thing in the beginning? Maybe you build your website your coefficient is greater than 1—
that's being born lucky!
[If a user brings in 1 or more other users your growth looks like the first curve, else you do
less and it looks like the second curve. Remember the born lucky thing in the beginning?
Maybe when you build your website your user growth looks like the first graph.. That's being
born lucky The equation looks like this……you have to look to other avenues to sustain
growth… And the solutions is new users = e^kt. If k is positive…]
14. This Smiley can make
the difference… :-)
Saturday, September 11, 2010
As for making your own luck…For instance, Tagged found that if they added a smiley face to
the end of the subject line of the e-mail, 20% more users click would actually bother to read
the e-mail. How do we know that? We used a web bug. I did mention that I worked for evil
right?
SOurce: http://images.fanpop.com/images/image_uploads/A-Bug-s-Life-a-bugs-
life-626985_1280_1024.jpg
15. :-(
Saturday, September 11, 2010
!!!CUT 4
But at LinkedIn they found the exact opposite effect. 20% less users!. Imagine if your viral
coefficient was .99. Then the smiley face may put you over…or keep you from going viral.
[How do you know the difference? You test them!]
[You use AB-testing]
16. Saturday, September 11, 2010
Maybe before you averaging 100 signups a day. After tuning you have a 300 signup day, then
1000! then Barry is pissed with you because your code didn't scale and the servers went
down. This is called the “hockey stick” and it matches the the born lucky curve exactly..
Source: http://www.hockeydogs.com/ProductImages/sticks/CCM%20V4.0%20Stick.jpg
17. Key
Ruler
Test
Key = Measure + Test
Saturday, September 11, 2010
The key is not virality but measuring and testing—and that can be applied anywhere. Want to
map out the cost of site speed? Deliberately slow down the site for some users and measure
their activity. Use your imagination but remember they aren’t viral problems (the viral curve
doesn’t apply)…
[So why mention virality? Want to compare two CDNs? Put a javascript to return your site the
page completion times and bin users into A-B.]
18. Saturday, September 11, 2010
Now don’t get too deluded with statistical power. You can’t measure thinks like "word of
mouth". Once Tagged applied a Zynga trick for virality. It put our viral coefficient above 1
again. And also got us this…
[adding an X factor to your faith in numbers … Against the objections of many in the
company it was implemented.]
19. Saturday, September 11, 2010
Sure for two days before it was turned off we had a million and a half signups, But this article
was the most e-mailed article on Time.com for the entire week. That damage is
immeasurable… literally!
[—that’s the entire city of San Francisco!]
Source: http://www.time.com/time/business/article/0,8599,1903810,00.html
20. Saturday, September 11, 2010
Oh yeah… the company got sued by the state of New York…Twice! So what’s the moral of
this story? It's way better to be born lucky… (even if mean reversion means that your founder
is unlucky at cards.) Thank you.
Source: http://ma.tt/