Just wondering….

Configuration problems when deploying a new build are high on the “high anxiety” list. The pressure is high to hurry up and get the build deployed. Often times, the deploy happens late at night when the deployer really just wants to go to bed.

The number one way a developer can help out in this situation is very good error messages. If an exception occurs as a result of a misconfiguration: be verbose. Be overly verbose. The sanity you save may be your own.

This error message does not cut it:

Caused by: java.io.IOException: Couldn’t initialize working directory.
at com.amplafi.core.iomanagement.FtpManagerImpl.initializeWorkingDirectory(FtpManagerImpl.java:83)
at com.amplafi.core.iomanagement.FtpManagerImpl.initializeService(FtpManagerImpl.java:50)
… 16 more
Caused by: java.io.IOException: Couldn’t create directory: ftp-working
at com.amplafi.core.iomanagement.FtpManagerImpl.initializeDirectory(FtpManagerImpl.java:61)
at com.amplafi.core.iomanagement.FtpManagerImpl.initializeWorkingDirectory(FtpManagerImpl.java:81)

The error message should have at the minimum contained the following pieces of information:

• The directory name as supplied in the configuration
• The full absolute path to that directory (not the same thing as the configuration name if the name is a relative, not absolute path)
• The reason for the problem (was the parent directory not writable, something already there, disk full, etc.)
• A suggestion on how to correct the problem

Angela Strange
BayPartners
(VC firm)

They have a fund targeted at social networking applications, not just facebook applications.

(Once again notes will be edited as time permits)

---

300 million fund
looking to fund applications that are built on f8 or other social media.

Do you really need money?
Don’t raise the money if you don’t need it.

Raise the minimum needed, trauche the money only take the money in chucks as needed. So if the money is not needed, then equity for later trauches is not given away.

if you’re going to raise then get more than just capital.

criteria
-clarity of your thoughts and ideas. ( need idea in the first 2 sentences ) if need 2 pages, then hard for her to understand reason for product.

entre/team
* strong tech prowess mainly, being able to execute.
prioir track record or indications of potential for first-time

* idea uniqueness/ differentiations/competition/depth — provocative

* business model /sustainability

most interested in apps that may start on f8 but can go outside of f8.

traction ( if already launch)

evidence of high level of engagement/ or growth rates.

for example: may be privacy apps might be

are there any applications that should wait until api matures — as/ no go for it now! Look at any productive apps that could work with f8.

as) there is a market for more utilitatarian apps — like classroom wiki

jury is out on if there is going be any market for utilitarian apps.

communication, shopping or service where you know who you are / connected.

bay partners:

* app
* amazon free hosting/data services
* post lauch services (marketting services)
* not taking board seats

for apps already launched:

1 app with 10 million users is more valuable than 10 apps with 100K users each — pretty easy to get to 100K users and then it dies off — but 10m users is much harder to do.

CTR for apps have not been very good.

earlier losers: anything too complex, so if you think it is a little complex it is probably 10x too complex.

building private social network: for boating enthusiasists.

for apps that have launched value the users $1 - $20/ active user.

convertible note: “caps” — $100K with $1m “cap” then bay partners get 10% when the next round is funded (no matter how much is valuation at that time.)

“cap” with no conversion — the next round investors will call and bay partners will say well we won’t want to participate at 10million but we will participate at $5million because other wise the initial 100K invested will be wiped out.
“equal participation (in next round)” is better than First rights.

Presenter:

Sometrics (missing card right now)

---

Claim to give better data on who the best users are.

social metrics - sometrics

more than just traffic
– impressions, uniques, visits … go deeper.

only 10 friends but converts 10 friends — they will have metrics based on action.

some limitations becuse of f8 TOS.

So application could change based on time of day. One of sometric customers didn’t know spread of sex / age distribution.

sometric will also look at location.

sometric is anonymizing and hashing and aggregate the data.

f8 touch points -where in f8 wall, feeds, email, profile wide/narrow, request invite

can’t measure unless pings over to canvas pages.

Users engagement
– Power users — certain actions indicate power users.
2 different power users : influencder - can drive application installs and drive back to canvas page.
in future sometrics will configure and define app

80% case from average user or power users —
average user or power users — who has greater influence?

applications can be able to take the application dates

active users daily/yesterday active change %.

type of users:
power users
participant
viewer (lurker)

rock you — aggressively uses notifications for first few days. but then turns it off once uptake is above a certain level.

Presenter:
Mike Sego
Fluff-Friends

Fastest growing “pets” application on facebook.

---

(Notes will be edited as time permits)

---

Secret to success is being “cute” — emotional connection

Pushing information into newsfeed runs risk of flooding newsfeed with too much information.

First Feature added is increasing number of fluff friends. There is the point of diminishing returns. Going from 1->2 friends is interesting. 26->27 is less interesting.

Part of being cute is that each fluff friends has a story and “cute”.

gifts less interesting at 500 gifts than in beginning.

people join together and have petting groups.

100% of growth was friends seeing fluff friends on profile not newsfeed from growth.

added petting and naming fluff friend.

“Pet Me” — would say “looks like you don’t have ff… would you like to install”

He knows because when f8 api was broken and not having warning dialog pop-up… saw lower growth when f8 api was broke.

The wall is more compelling features of f8

“Don’t poke me … do something useful pet my ff” i

implemented cute verion of tracksor… because you see who petted your friend and when.

Petting is core to experience.

What ff is not:

pokemon - (cock fighting for little warriors?)
pokemon requires lots of enginerring effort.
f8 takes things that are very popular on internet and creates walled events.

gmail is losing market-share to facebook mail!

f8 groups -
f8 photo is crippled version compared to flickr.
just need to hit core features and will be successful on f8
“what are you doing?” is what f8 is all about.
so if you focus on how that question

ff is LAMP ( but Mike Sego - is java developer works on gmail )

social interaction is at the core.

added different features that increase would

As released features — add Moods 2.0 different moods for different ff so people would go to see other ff.

using transparent png (yech) css rendering is really hard because f8 rips up. so he has to hack up css for each browser.

Bubble - people would be having conversation with their ff talking to each other.

added food — released 2 a week … leveled off at 35 foods.. would cost friend.

food is gifting “costs money” — ltewright version of gifting… but found that most people feed their own ff’s not others.

ff had problem with how to earn money — How can I design the next feature to make previous feature

you have 7 seconds to return result to f8 — “Try again”

Using one-on-one hosting … had to shift from shared server to dedicated hosting

LAMP ( mem_cache apc — for mysql )

Added habitats - once again made previous features work better. Biggest feature — huge uptake.

786 different images on the shared vm server.

Using fb photo for the foods. but didn’t upload 786 photos because no way to programmatically load photos.

but this is great because fb

ff starting tacking

merchandising….
goodstorm — lower quality
zazzle — higher quality and will ship international

but only sold 400 shirts.

But need much higher quality

“I want a Liger…”

support cost — message bombing - trying to campaign on getting new features added (

was getting 300 emails a day — started to managing issues with FAQ. FAQ slowed down because link to find developer link.
make effort

facebook forum is useless.. no topics in forums.

so forums are offloaded. but 10000 registered users - 2000000 ff install but acts.

ff-users going to forum add value to the ff-ecosystem.

A lot of people wanted to fighting… (like pokemon)

did not want to do RPG-like things.

didn’t want to compete EA (Mike Sego used to work for EA on Sims)
single player - no p2p
multiplayer -async (works in f8 — no sync communication available)
multiplayer - sync ( much harder on f8 )
massively single player — you are not waiting on anyone else but what they do effects you.

fluff(Race)
number of fluffs per hours.

communities:
* self-expression
* and racing community….power lovers of ff.

—–

Choice not diverseify because that would say is he done. So chose not to.

has not looked at uninstall numbers.

super actives <5% but they really drive the

you can’t uninstall your ff (because that would be abandoning your ff — so sad )

can’t sell transfer

this middle-age woman whent around every day to pet 100 friends ff.
80% woman — early college

can’t pet a ff more than 5 secs. prevents scriptting but does not alienate hard core users.

fluff-book — is the high page view.

** only now added invites **

hit servers …

trying to stay constant relative to number of f8 users.

monetation will come when f8 makes it happen.

Held 20 Oct 2007.

This batch of notes is not mine, so it will probably not be edited too much.

---

Presenter:
Eric Yieh
RockYou

- Facebook, is 7x more viral than bebo. because of there
- superwall, Xme, zombies/vampires,
- Facebook has: profile, profile action, feed, requests, emails, notifications, wall.
- Should use all of the facebook api, some work better than others
- Rock You also provides an API to the super wall
- Having multiple apps lets you cross advertise, and have one app use the other app.
- Buy advertising to seed various different social networks with your app.
- Viral success factors: simplicity ( does the user get it immediately, keep invites really short), novelty ( has the user seen this invite before?, be the first with your kind of app), social persuasion (reciprocity, ego boost, “your friends will like you better if you install this app”), fun ( most people go to facebook for entertainment rather than work, not too many utility applications).
- rockyou and slide are the heavy hitters in the widget space.
- Must use the social network of facebook, people who wouldn’t otherwise
- “to get an app over 1 million users you must advertise”
- “it will have to hit a bnuch of people in the same high-school before the app spreads to the whole network”
- facebook invites only go out to the direct friends of a particular user.
- Invites are now limited to 10 invites
- FBML invites are not very good, need to have a very strong call to action.
- Want to make it easy to send invites (pre-set the invite text).
- maybe can monetize without having people install an app.
- you have a canvas page that people don’t need to install an app for, this lets you monetize your canvas page. (it’s just like an html page on the web)
- Create applications that require users to communicate.
- Hard part, need to balance uesr acquisition and engagement
- if you send a person to an invite page, they may not use your app.

- which ad network to use?
- pricing
- quality of the ads
- fill rate/campaign size
- want optimization features that let you tweak which ads are displayed
- channel tuning (let you not advertise for competitors)

- rock you has an ad network as well.

- CPI 0.50, cost per install, advertising. i.e. you pay 50 cents per user who installs you
- you need to monetize your app in some other way than selling installs.

- People will need to advertise to launch your app…
- In some cases people have built a small viral app to drive people back to their “real” site.
- advertisements in the middle of the screen are much better than top flier ads

- bebo focussed in UK, dont have a dev platform yet.

- flash autoplay will be allowed.

- should tie your apps across all of the big platforms.
- news feed and mini-feed are both very effective.
- facebook doesn’t let you store user data for more than 24 hours… can only store theire user ids.
- Eric says that you need to do internal studies to figure out what kind of demographic gives you.
- monetization: peanut lab.
- clickthrough rates on facebook are low.
- as soon as a click takes you outside the site, the user closes the window because they don’t want to leave facebook.

- maybe myspace will allow unlimited invites.

- rockyou doesn’t care about their destination web site…. they care more about how many users they have on various social networks.

- Rockyou assumes that all of their users are teenagers and build for the teenage market, whether or not they are teenagers. This is because when people are on facebook, they’re not looking for productivity, they’re looking to hang out with their friends.

- Thinks that the movie studios etc. will build viral facebook apps to advertise new movies soon, instead of having a fancy website.
- “Who can do better than george bush”
- Rock you sold a resident evil branded version of zombies on facebook to sony, and sony’s going to buy again.
- rockyou has relationships with various advertising firms
- anything that distracts from the invite flow will reduce installs of an app.
- advertisers are targeting younger users on the social networks.
- Each platform is building their “own version of the internet”.
- which means that people will rebuild the internet inside each platform. So take the most popular apps from the web and rebuild them inside the platform to make money.
- Make sure to track ROI on a campaign.
- what’s the “hover over” rate.
- branded advertisers are more concerned about the messaging (the layout, the look and feel etc.) rather than necessarily the click through rate.

Presenter:

Blake Commagere
Founder, VP of Engineering & CMO
Mogad, Inc.

Create the Vampire App for Facebook platform. Also work heavily on the Causes Facebook application.

---

(Notes are raw and unedited, editing will occur as I get the chance!)

most people are using f8 for communications not business.

internet generations are much smaller than regular generations

english majors can do html
internet dating is casual to the kids today.
First porno, was probably digital.
70% are 18-35 f8 users

what people want has not changed in 5000 years but the manifestation

for example, don’t assume users are asexual.

for example, the wall — fundemental desire is validation - to feel important.

for example, friend count when friend count meaningful. (i.e. not 30K friends but 35 friends v. 50 friends)

social aspect of wall-to-wall communication : “john, lets have coffee” the public aspect of advertising that john and bob increased wall count. And john got validation by bob ask

wall is ordered by date with most recent — implecit ranking because Mike is cooler than Karl because Karl’s post is older (below) Mike’s wall post.

flirting happens with posts on wall posts

if a female posts on males’s wall — if a male other user posts above her. she is far less like to post again than if another female is posting.

for example in real -life if have female friend that is talking to another male. a male is likely to go up and talk to female just because ‘competition’

(from audience : f8 user) -
if she just met a guy in class - she would go to his facebook wall to see what kind of competition she has.

over 30% of profile views on linkedin profiles are to self — and this is just a resume.

having a girlfriend is important on f8

30% more like to get use apps if they have pimped their profile.

sales people try to create social obligation — that is why they send xmas cards.

blake — was lead developer on “Causes” (”Yellow bracelet” campaign was a way to self-identify with Lance armstrong) tried to do application with “Care badges” was unsuccessful in myspace but myspace shutdown widget spreading)

causes only exists in f8 — people do care about charities — “I donated $20 to a charity” 20-something don’t have time that they did in college and dont have money either to donate. But donating to causes is self-expression.

Users are “selfish”

(vampires - kicking around idea of creating that did nothing but spread — but not have any other value — i.e. but convert the act of inviting — make the advantage of inviting a core feature — makes it more viral)

anti-example : the quote applications .. no real benefit to inviting someone.

f8 poke — is not flirt — its a hey what’s up. I think someone is flirting with me because they are poking more frequency. “Do I poke them back right away, or do I wait — a day to say that I am cool and hip and I have lots of people poke back.) This is a friend validation. If they don’t get poked back, it says something.

If I post on your wall and you don’t– “oh you are private message’d from now on.”

voyear-istic measureable to see how flacky you are.

f8 photos are inferior to flickr but tagging in f8 automatically notifies your friends.

people look for their own photos first. and then the attractive people. People want to see themselfs

message has 2 barriers — the sender has to decide to send the app. The actual barrier is to the recieptant to accpt. You want them to feel cool.

People do things to benefit themselves first, their family/friends and lastly their community.

“Useful” apps run the risk of being boring.

f8 is not designed to get work done.

Causes has grown… but needed bizdev team and engineers — did a lot of work with companies. established relationship with just-give online charity and guidestar - to validate non-profits.

application, engagement, monetization. ( iterate with users get immediate feedback ) — right now

Inspired this 31 hour 4 minute run coast-to-coast. The youtube videos and the previous record run.

and …

The checklist for the 31 hour 4 minute run, which hit a top-speed of 160mph (256 kph):

• Twenty packs of Nat Sherman Classic Light cigarettes
• Breath mints,
• Glucose and guarana,
• Visine and riboflavin,
• Gatorade and Red Bull,
• mail-order porta-pissoir bags of quick-hardening gel,
• Randolph highway patrol sunglasses,
• 20-gallon reserve fuel tank,
• Tasco 8 x 40 binoculars fitted with a Kenyon KS-2 gyro stabilizer,
• military spec Steiner 7 x 50 binoculars,
• Hummer H1-style bumper-mounted L-3 Raytheon NightDriver thermal camera and LCD dashboard screens,
• front-and-rear-mounted sensors for a Valentine One radar/laser detector,
• flush bumper-mount Blinder M40 laser jammers,
• redundant Garmin StreetPilot 2650 GPS units,
• preprogrammed Uniden police radio scanners,
• ceiling-mount Uniden CB radio with high-gain whip antenna.
• E39 BMW M5,
• copilot,
• $150,000,
• chase plane…

Some interesting data points:

• Direct Sales: Someone shows up at customer site
• Tele-sales: person-to-person calling
• Direct Marketing
• Word of Mouth

Method	Cost per sales person	Minimum Sales per year	Minimum price (200 selling days / year)
DirectSales	$300-400K	$1 Million	$5000
Tele-sales	$80K	$250K	$1250

Basically means this means that if you need to have person involved in every sale, the minimum cost of your product has to be over $1000. If it isn’t, advertising and word of mouth better carry the day!