Memo to Google, Yahoo, Microsoft, & AOL: How to Turn 500M email logins into Facebook Platform & a Crapload of Revenue.

ok, so even after all these announcements about data portability, no one seems to understand what the hell i keep talking about re: using email & IM messaging data stores to improve targeting & make friend lists work better.  i admit it's a bit inside-baseball, but it's actually pretty simple.

[UPDATE: Yahoo just joined Microsoft (via Windows Live Contacts API) & Google (via Google Contacts Data API) in enabling 3rd-party programmatic access to contacts in the Yahoo mail address book accessible via API.  now all 3 just need to enable filtered access using messaging frequency & keyword relevance, and we'll finally be getting somewhere... who's gonna step up first?]

the thing is, i actually DO believe we're on the verge of something game-changing here.  i DO believe that the social gestures & interactions pioneered on Live Journal & then Facebook & now Google & MySpace are slowly going to migrate over to the rest of the web, and energize all of our normally boring websites with social cues & viral behaviors. 

HOWEVER: i'm really friggin' tired of a popup with 50 of my 1000 friends (whose names all seem to begin with 'A') or being asked to upload my address book (over 2000 contacts, many of whom are email lists or people i don't know) to use as a viral distribution method. 

so here's the skinny homeboy:

I HAVE WAAAAAAY TOO MANY DAMN "FRIENDS" IN MY SOCIAL NETWORK.  I GOTZ THOUSANDS OF DONT-GIVE-A-F*#K "CONTACTS" IN MY ADDRESS BOOK.  BUT DON'T ASSUME I'M GOING TO SPAM / UPLOAD / INVITE ALL OF THEM TO CHECK OUT YOUR CHEAP-ASS, NON-MONETIZING, PRIVACY-RAPING, BUTT-UGLY DESIGNED, RETARDED UX, PO-DUNK WEBSITE OR FACEBOOK APP.

In other words, i don't trust [most of] you.

in fact, i don't trust ANY of you -- friends or websites or apps -- any further than i can throw [a sheep at] you.

just cause i've accepted an invite to be your friend, or happened to find you on a website, or in a forum or group i've passed by in the online version of a wham-bam-thank-you-ma'am-for-the-one-night-stand, DOES NOT MEAN i want to know what you had for dinner last night, who you knocked boots with lately, or if your new social networking startup for Retired Dread Pirate Roberts or website for exterminating Rodents Of Unusual Size is worth bazillions.

on the other hand, there probably IS a small subset of crazy peeps who i DO want to share your shit with, but it's probably only ~10 people in my potential network of 1000.  and i want to figure out how to choose 10 targeted people quickly & easily out of my random network of 1000.  (and probably only 1-3 of those folks will really take a look, but those are the ones you want).

so let me state that more clearly.

PROBLEM: when i'm on a website or social network or social app, it wants me to help refer other people / help with some lame-o viral marketing scheme.  however, i'm only gonna do that for a select few who really share my context / insanity, not everyone in my network (unless my name is Robert Scoble).

what I DO NOT WANT:

• select from a god-awful list of 100 faces in your pop-up face listbox
• upload my entire address book of 2000+ contacts for you to spam the world
  
• wait for your data-loading / selection function to crash horribly, after taking forever to load

what I DO WANT:

• popup the MOST RELEVANT 5-10 peeps who meet certain key criteria
  
• use an intelligent combination of shared interests & messaging frequency to figure out who these "TOP" friends are (for the given context)
• let me select 1-3 of them to invite & checkout an awesome [video game | baby stroller | new book | really good pr0n] i just found

SOLUTION: let's see now.  WHERE? can i find an easily accessible data store with info on how frequently i contact other people, already-indexed relevant keywords for conversations i've had with those folks, and provides a built-in distribution method for sharing further information & content?

ANSWER: why, it's EMAIL! you stupid mofo.  and also your IM client, and your SMS history on your cellphone, and nowadays your photostream & Facebook wall & tweetstream & everything else related to the Centralized ME.

Now, WHO has the largest collection of email / IM data on the web?  Well i'm guessing the order looks something like this:

1. Yahoo (~400-500M email users, ~10 years of data)
2. Microsoft (~300-400M email users, ~10 years of data)
3. AOL  (~100-150M email users, ~15 years of data)
4. Google (~100-150M email users, ~5 years of data)

note for the record, MySpace & Facebook also have a shitload of this data, as do other social networks, altho like Google they're a bit newer to the game than Yahoo, Microsoft, & AOL who've been collecting all that data for over a decade.

ditto for cellphone carriers, especially in geographies where text messaging is more prevalent (= non-US, altho we're finally catching up).  except you don't really have to worry about these guys, since they're even more clueless than the big web platform dinosaurs.

so now recently, Facebook & MySpace & Google have all rolled out offerings to demonstrate off-network profile access, friends list, & other social network platform services infrastructure to enable any other 3rd-party website to provide Facebook Platform-style services & apps.   previously Microsoft, Google,  Yahoo, & AOL have created authentication & contact APIs to enable use of login infrastructure & address book services on 3rd-party sites. 

NOW i'm not sure what this is so damn hard for people to figure out, or why it's taken over a decade for Microsoft, Yahoo, or AOL to get their ass in gear, or why Google didn't do it in the last 5 years, and why Facebook Platform (which was built in barely a year?!?) blew past them all to become the first legitimate Web OS (walled garden or not). 

But now that everyone has FINALLY woken up, the next step is to put all this shit together & make it useful for everyone else OUTSIDE the walled garden.  In short, that means:

1. Use 3rd-party acct infrastructure to provide "no login rqd" single sign-on
2. Use profile data for landing page customization & campaign targeting
3. Use email data for friend lists, filtered by msg frequency & keywords
4. Use all of the above to enable viral distribution on the non-social web
5. Use payment info (after MSFT/GOOG buy AMZN/EBAY) to buy stuff

Platform vendors -- at the very least Google & Microsoft, hopefully Yahoo & AOL too -- should be offering these enhanced services for any 3rd-party sites that utilize the above infrastructure.  Perhaps it will take further consolidation to make this happen.  My guess is one camp eventually is Microsoft + Yahoo + Facebook + Amazon, and another is Google + AOL + MySpace + eBay, but those parts could fit together slightly differently.  Regardless, once you combine portals + search + social networks + messaging + commerce, you get a very very interesting & quite lucrative combination of assets.

Those assets SHOULD be utilized to create a Web Operating System which in turn drives the implicit web that accelerates user acquisition, improves activation & retention, energizes referral, and eventually drives substantial increases in revenue for a whole new wave of web 3.0 startups.

Of course i've already covered all of that in Startup Metrics 101.

| View | Upload your own

Comments

[Dan Khan]

Hi Dave,

Man this is the most refreshing and insightful tech read I've had for a long time!

Reminds me of some of the work I saw in the initial xobni outlook plugin beta in the way they aggregated and extracted info from email in similar ways.

[SEO]

this is a good read email isn't over but like everything it has a life span it will morph into something diffrent kids of the future will laugh about email

[Jan Horna]

Dave,

the great post! But still, I believe we can go much deeper in these thoughts. We just need to go to the roots which is the structure (format) of data. The internet turns from a rich application medium into a rich-data place = becomes semantic.

Imagine you send a message and the related data flowing through mail servers have already known exactly how it is going to be displayed in your email/web based client according to your personal preferences.

We just need a better prioritizing system for electronic communication.

[Jason Devitt]

Dave,

We've built this for cell phone data. I'd love to show you. (We're still in private beta.)

http://skydeck.com

Jason

[Joseph Hunkins]

Dave you data sharing boot knocker!

This is brilliant and to the point. I'm inclined to think ubiquitous and interconnected website social widgets, perhaps seeded with out email data (!) will usher in the new era you be talking about, Willis. I'm guessing phase one will be lots of pestering and weeding out of those of our "friends" who are ... not. (yikes - that part could take Scoble 10 years!)

[Rob Goldman]

Some interesting thoughts in there.

My take is that the "short-list" algorithms will be hard to write and will often fail given the great diversity of services and wide variations in email behavior.

My bet is that the open access will be used to add social context to existing off-network sites way before they will be used to prioritize invites.

[Dave Hodson]

If you add IM into your stats:

>1. Yahoo (~400-500M email users
>2. Microsoft (~300-400M email users
>3. AOL (~100-150M email users
>4. Google (~100-150M email users

Positions 1 & 2 flip (and AOL possibly moves to #2 with AIM)

[dave mcclure]

whether or not it's email or IM, the same logic & same platforms apply. the leading IM providers are also the leading email providers (Yahoo, Microsoft, AOL). both types of services provide the same targeting opportunities.

and while my # of contacts may be outside the mainstream, the same problems apply with just 100-200 contacts (which is well within mainstream).

whether or not you think email is "over" (i disagree vehemently), the ability to use email and/or other messaging systems to target & filters users remains quite valuable.

and yet, we still await a service offering to provide this functionality.

hopefully someone wakes up & smells the coffee soon...

[email? 2000+ friends? how are these realistic problems?]

I would tend to disagree with the whole email thing you're pushing. I don't think I've emailed any friends of mine in a long time. It's all just IMs.

Most "kids" (13-24) these days don't bother with email anymore with an exception for professional use. It's too formal for friends.

So who has the most information about friend connections? Meebo acquisition anybody?:)

That being said, many of the issues you've brought up don't apply to the majority of people. I think that you're not exactly the mainstream audience. Most Facebook users do not have 2000+ contacts, most of which they do not actually know. In fact, quite a number of Facebook users actually will not accept friend requests from strangers (knowing that if they do, their personal photos, contact info, and address information will be revealed, and that they will be spammed by every action that user ever takes...).

I guess that's why many niche companies that focus on solving these problems can't get mass market appeal. It just isn't really a problem to more than 99.9999% of users (that's right, right? ~130,000,000 users, and probably less than 10,000 silicon valley techie types with 2000+ non-friend connections...) or perhaps a more quantitative estimate would be that the average number of friends a facebook user has is ~150, so a person with 2000+ would be ... 10x+ standard deviations from the mean? Statistically speaking, that's like getting hit by lightning while winning the lottery.

I do agree about viral distribution on the non-social web. Friend Connect would have helped with that, but already the boys aren't playing nice.

[Nathan Ketsdever]

Great post. I'm looking forward to the next generation of social media aggregators like Friend Feed to allow you to focus on:

a) Intensity of connection to individuals in your social graph. So you can hear the personal stuff and tweets by them and focus on the content area(s) they like that you also happen overlap with. (and you don't have to hear personal tweets you could consider noise, spam, etc--especially those you have less of a connection/affinity/relationship with)

b) Intensity of interest in category (aka tag) According to last weeks digital townhall, Digg is moving in the direction of intensity in their subcategories--so that you could customize to only get certain feeds if they had a particular # of diggs (ie you only like world news a little--so you only get those posts with 100 diggs in that subcategory in your feed). I hope others follow this model--before its implemented at Digg, which could be 6-12 months off.

[Dave Ambrose]

Dave, this is ridiculously insightful and refreshing to read with all of the current fanfare.

Kudos.