The Seminal Web 2.0 Service

I wrote the following about YouTube the other day:

I went through a period where I wrote about YouTube a lot. It was fascinating to me to watch web video take off and the lessons we all learned from YouTube's success were important.

I am an afficianado of web services. I love using them, figuring out how they work, the neat stuff you can do with them. I learn so much from using the web. And I try to write about it as much as I can.

For some reason I woke up this morning thinking about Flickr. And it occured to me that I've learned more about how to create a great web service from using Flickr than any other service. By a long shot.

Here are 10 things I've learned from using Flickr:

1) Making online content default to public instead of private creates community

2) Every web service needs to have a profile for every user

3) Users should be encouraged to comment on other user's posts

4) Tagging content is better than foldering content and the tags should be public

5) Users should be encouraged to tag their content when it is posted to the service

6) Widgets should be used to make content available off of the service

7) Content on the service should be "bloggable" with one click

8) Engagement metrics like comments, favorites, views, can and should be used to drive discovery (the most interesting algorithm)

9) Geotagging is great but we've yet to see a great interface for geotagging

10) Machine tagging (autotagging) is the next big thing in web 2.0

I whipped off this list in about three minutes so I am sure I've missed a bunch of other stuff I've learned from Flickr.

Many people will say "Google is the seminal web 2.0 service". But I don't think so. I've learned a bunch of things from Google too, but Flickr has pushed the envelope in so many ways at such an important time ('04 and '05) that it gets my vote.

Comments

[carl rahn griffith]

agreed - i am a recent convert to flickr (thanks to your web site) but its potential is fascinating, let alone the other general business ideas it stimulates ... on a personal front i've set up a klipfolio rss reader on the pc's of my in-laws so whenever we're out and about and submit a picture to our flickr account via remote email, up it pops on their desktop. all the family can submit to this - so cool. we just need to get them that wifi and rss-enabled photo frame, now ;-)

this is why, for me, yahoo is (still) the good guy on the web - they are the first and the best in so many ways. ok, they've had their troubles and flickr was an acquisition (not in-house) but it was a discreet, modest and astute one - like with mybloglog.

[Robert Dewey]

Tagging is definitely a next technology, up there with content delivery.

It gives the users the ability to move in the direction they want, covering the niches that they demand. Imagine if eBay would have allowed tagging instead of hard-coded categories - the users would have been able to create the categories that are most important to them, leaving the not-so-liked categories non-existent.

I'm really interested in drill-down tagging as a way to sort through stuff.

[Rick]

I learned a ton from both Flickr and Delicious.

I would add to your list: Offer website content management service.

Lots of people use Flickr to host their website's photos... instead of their blogging software or their CMS, etc. Just easier.

[Noah Robinson]

If Flickr offered video uploading, YouTube might never have become YouTube.

[Malte Schloen]

Just a comment on no. 9: Geotagging really needs an easy to use interface. We have tried to build one at locr.com

[Greg]

You said:
10) Machine tagging (autotagging) is the next big thing in web 2.0

What do you mean by that? What is important about it. I believe I have to AI skills and free time to make it happen. I'd love to hear more of your vision on this.

[Raj]

Two things bother me about Flickr, though I pay for the pro account:

1. I lose complete control over the detailed stats related to my content with Flickr compared to self-hosting something like Gallery.

2. No rev share for my content.

[Jianing]

11) webservice API for 3rd party development, or more generally: move *down* the food chain.

[Dustin]

These boil down to one of the Golden Rules of interface design: give users control.

[Joe Duck]

Right on Fred! Although I strongly agree that Flickr is a brilliant 2.0 posterchild.com , is it ironic or am I wrong to think that Flickr is not really perceived as a place of "community" by many regular users like me even though I sometimes use it that way to see what other folks are up to.

[engtech]

I would have said that Google was the seminal Web 1.0 service.

[mark ghuneim]

That you can monetize social media by augmenting the experience not disrupting it.

[Peter Cranstone]

How about this for Geo-Tagging

This morning I took a picture using my iPAQ which has a GPS inside it. I uploaded the file to Windows Vista and viewed it in the gallery. Then I uploaded the same file to our server and ran it through our software. As you can see (http://www.5o9inc.com/5o9MobileMePics.html) quite a bit more information is revealed - including the exact time and location I took the photo. We're still in the process of activating the link at the bottom of the page which will take the GPS data and send it to MapQuest (or wherever you want it to go).

What you will also note is that every piece of data in that photo is now available for search engines (if they only knew that it was there). None of this is rocket science - however for those people who are interested in Geotagging in real time from a GPS and camera it's now possible.

Cheers,

Peter

[kristine]

At first, I Loved Flicker. I still use them, but not obessively as I do other social services. Here are the few things that made me fall out of love with Flickr...

1. Anyone can add you into their contact list. I haven't tried to control this, but I thought it was interesting that I didn't need to invite them first. Of course, I was added by individuals with porn content, only because I had a photo of two girls in "panties" (No, not at all porn but more lifestyle art, and no nudity.)

2. Security. A very bad thing happened to me once. Somehow this "riske" photo showed up on google search under my name... Even though I did not tag this with my name. They are so good at ranking, this image, connected to my account, showed up sixth on google search, beating more than three pages of links. I was silly enough to use my full name in the email I used to sign up. I posted a complaint to the forum, and they blamed Google. Two days later, the link was broken. No explanation how, why...

3. They monitor their service. They will delete photos that are not considered appropriate for public view. They will also delete your account if you are promoting your business. I find this disturbing and very anti-web 2.0.

[Stewart Butterfield]

Thanks Fred - on behalf of everyone here, we appreciate it :)

[MidNight Mapper]

In the future what happens if I wanted to move all my Flickr content to a newer better solution? Will I get back not only what I contributed but the tags and other metadata as well? Who really "owns" or "controls" all of this tagging?

[michael koz]

uhh... thanks for this really insightful post:-(

[Rikki Tahta]

Here's a free idea for anyone who's interested in machine tagging - and if anyone wants to run with it - I'd invest a few bob as I'd love to use it

Use Wikipedia's URLs as a classification tree and the entire wikipedia as a pattern indexing base - i.e match any content eg blog / news / article to the content of wikipedia - and tag with the most relevant matched URLs

that way you are riding off a free and ever-improving rules index that is not static and you have a universally available reference source for everyone to refer to and finally a permanent hard tag in the URL .

It will allow a) automated tagging so low cost b) hard classification with an easily accessible reference file so the tags can now be used by machines and cross refernce and content becomes available to be integrated into other applications

I come from the old school of hard classification and what I miss about folk tagging is the hard database capabilities of structured trees, that allow you do more things with data - but what i like is the fluidity of folksonomy (and of course the low cost of classification)

this way by using Wikipedia as both the rule set and tree you harness collective intelligence and keep everything open source

Greg - or anyone else - feel free to run with the idea - I think it'll be good for us all

[Emil Sotirov]

At first... I was going to say that "autotagging" may not be such a big deal... because how would that be much different from search engine indexing.

And that's when the idea came to me... How about a service from Google to give me a stream of keywords from their index corresponding to the page I serve... so I can display them as "search tags" ... working as predefined Google searches.

This idea is based on something we had on Aidpage since the very beginning in 2004... where we would take user generated tags and offer them also as what we called "aidjumps" working as preset Google searches. (We had to take these "aidjumps" down because of a conflict with the AdSense terms of service... and for another unrelated reason, you won't even see tags now on Aidpage... we're working on a major upgrade).

Anyway... why not have Google's index out in the open... on my web pages... as a contextualized self-updating interface to related content... perfectly in synch with our common AdSense based interests...

A simple click on a "related" keyword (close to the main content) is 10 times (my educated guess) easier than having to come up with good search words (too much thinking) and typing them into a search box (too much work) somewhere else on the page.

Web links are THE web interface... not search boxes... which makes me think that, eventually, Google may have to offer this service to compete with the human indexing of the web as it now starts happening on a massive scale through author self-tagging and social tagging/bookmarking.

I appologize for the long comment.

[joe laz]

Emil, I'm not a technical guy, but I think it might be possible to build what you're describing using Yahoo!'s Term Extraction API...

http://developer.yahoo.com/search/content/V1/termExtraction.html

You feed that API a block of text from your site and it returns a list of popular search terms from your content. You could then display those keywords alongside the content on your site as machine tags.

[Emil Sotirov]

Joe... thanks for the suggestion. The last time (long time ago) I looked into Y's Term Extraction API, there was something that didn't quite work for our needs. But I'm sure things have evolved there... so, again, thank you for reminding me about it. I'll take a look again to see how this might work.

[alan patrick]

After yesterday's hoo-ha, I wonder if the Flickroos wish they had read your post...which I largely agree with, except that Privacy and Trust is becoming a bigger issue.

I have of course tried to explain all this here in Flickr slickr, usrs sickr :) - I have also linked to your post there as I think its a good reminder that in social media, We the Users have a lot more investment

[godz]

From what i can tell n.10 is being done already... Check out zdnet.co.uk's "Related Tags"? They seem to be using some sort of autotaging?
Do a search for "google" and a box shows up on the right with relevant tags about google, pretty cool imo!

[Emil Sotirov]

godz...

Yes, on zdnet.co.uk they have their tags probably from their own search index... and they work exactly as "aidjumps"...

The idea is that Google may offer such a free web service to anybody quite easily - as an additional discovery interface. In a way... it sounds fair - I allow Google to index my site but I want Google to give me back the results of the indexing - as tags that I can put on my pages. The whole Google index returned back to the periphery... exposed in the original context from which it was extracted... feeding back traffic to Google. Each such tag is an immediate Google search... much easier than using a search box... sending people to Google search results - what could be better for Google... or for any search engine for that matter.

The discovery experience would gain much if we combine user generated tags, local behavioral tags (footprints, etc), and tags from the likes of Google (with their machine power and global view of the web).

[Kevin Farnham]

I applied your 10 criteria to "measure" Ficlets.com for "greatness" as a Web 2.0 application. For such a young site, I think Ficlets.com stands up pretty well, though enhancements are still needed. Hopefully I applied your criteria properly.