Oh, Delicious — where did it all go so wrong?

Lots of people were happy when Delicious was rescued by YouTube (s goog) founders Chad Hurley and Steve Chen, and whisked away from the neglect it had been suffering under Yahoo(s yhoo). It was a chance for a rebirth of a small but well-liked social bookmarking service, linking up with some proven entrepreneurs who were trying to show they had a second act. It looked like it could be a case of two great tastes that taste great together.

But when the site relaunched yesterday, I noticed there were a few problems — a few elements that didn’t seem to work for me, or felt strange — and my old account had been deleted because I hadn’t gone through the transfer process.

What I didn’t expect was a high volume of comments pointing out that this was just the tip of the iceberg. In fact, the response in this comment thread was unanimous: This relaunch appears to be broken.

Here’s one commenter, Suman, explaining what went wrong:

I had a fairly bad experience with the new Delicious today. Just last week I had spent a few hours curating my saved bookmarks and organizing tags. The new Delicious doesn’t seem to know anything about it. All my effort is lost. There is no longer a bulk-edit function to redo my changes. I can no longer manage my tags – could find no option for deleting old tags. Some of my tags with special characters are now broken, I get a 404 when I try to access them. I am done. Goodbye Delicious.

Meanwhile, many seemed to have the same problem as Mindctrl, who said the transfer of accounts from the old Delicious to the new Delicious was proving problematic: “My account is gone, despite me going through the transfer process. I’ve emailed them and am awaiting a reply.”

And lots of people were angry about the changes to tagging and tag bundles. Ellen said “I’ve invested a lot of time and effort into sorting out tag bundles with hundreds of tags… but now only a small fraction of my tags remain listed, and the bundles are gone!”, and you could hear the anguish when DSP pointed out “I had over 4 thousand tags, now I have just 40. Please tell me this is only temporary!” And that’s just the start. The full comment thread is packed with people complaining about broken features, missing pages, dead feeds.

It’s not just our commenters, either. Over at ReadWriteWeb, Marshall Kirkpatrick said he wanted to like it but couldn’t. And Matt Lingard summed it up by showing the lengthy list of features that are “still in development” (many of which were entirely functional under the old design), stating simply, “if you’re this far from being ready, don’t launch.”

So what happened?

It strikes me that there were three possible reasons for this mess. Perhaps they were separate, perhaps they were linked together, perhaps there are others:

  • AVOS didn’t understand how people were using the website

The changes don’t appear to have a major impact on casual users, but how many casual, active users of Delicious were there? The visual chrome is a welcome addition for a site that’s trying to go more mainstream, but it comes at the expense of information: elements now obscured or made invisible include the tagging system (which has always been one of the site’s core strengths) and the network (the basic unit of social currency on the site). Without these, Delicious is of little use to many of the people who had stuck by it over the years.

  • AVOS didn’t get how people were using the API

Delicious had a lot of web developers and technologists as users. Many of them used the site’s APIs to pull data in and out, particularly to publish elsewhere — on blogs, news websites. Today, those things are pretty much broken — and, more to the point, there were no signals given beforehand. Nothing has been redirected or pushed elsewhere; no parallel systems seem to have been put in place to give anyone that was hooked up to the API the time — and warning — to change what they were doing. It just broke.

  • AVOS didn’t understand they were playing with a live product

This is probably the crucial element. In the web industry, we are all very used to developing sites in beta, testing things out, seeing the data that comes out. That’s the development process. Except Delicious wasn’t a new product; it was an existing one with a small but committed following. Those users who loved Delicious really loved it: they’d stuck around through years when the product was given minimal development or resources. They’ve been rewarded with deleted accounts and other problems, which has made them pretty angry.

Maybe the long-term future for Delicious lies away from that user base; but you can’t move them along simply by flipping the switch. Reworking an existing product is not the same as starting from scratch.

When you’re rebuilding or redesigning, you have a legacy to maintain. Yes, that can be a pain — but what else was AVOS buying if it wasn’t the brand and the user base of the site, and the data that they’ve put into it? It clearly wasn’t the technology, which was the first thing to get thrown out the door. When you rebuild a product, you have to remember that it needs to take into account all those people who rely on the service for all sorts of things. At the very least you give them options to fall back on, rather than simply telling them that all the stuff they’ve been using for years will be in the product again… just not yet.

My colleague Mathew Ingram thinks that the only way Delicious can prove it’s really useful in the long term is if the owners can pull quality data out of the site. He may be right, but the trouble is that if the new owners alienate everybody who stuck by it during the bad times, they might not even be able to get that far.