I left Yahoo over two years ago, but prior to that I spent three years running product for Delicious. Since then I’ve remained a loyal user and supporter. To this day I keep in touch with former Delicious colleagues and consider many to be friends. And though I’ve felt that Delicious has been frustratingly slow to evolve in recent years, I’ve always wished the best for the product and the remaining team members.
original photo: afagen on Flickr
All of this has made last week’s news especially saddening and painful. First the Delicious team was hit by layoffs. Then it emerged that Yahoo is either shutting down Delicious or trying to sell it. As I write this it is still not entirely clear what the real story is, but regardless Delicious is in peril.
An online debate has already begun about various ways that Delicious might be “saved”. As someone who was on the inside for a while and who wants very much to see Delicious live on, I thought I’d chime in. For the record, what follows are opinions based on my own experiences. I have not spoken to anyone inside Yahoo about this and I do not have any special knowledge about the current situation.
Convincing Yahoo to keep investing in Delicious
This is unfortunately a non-starter. Last week much of the team was laid off and my guess is that the product is now at best staffed for “maintenance mode”. This sends a fairly clear message and it’s not something Yahoo can easily reverse because they have lost already-scarce expertise in both the product and the complicated technology stack that underpins it.
Selling Delicious to a third-party
This certainly seems like the best option for Delicious and its users, and I hope that Yahoo is able to pull it off. But it’s not a straightforward proposition.
As mentioned above, most of the team is now gone. Last week’s leak (and the subsequent fallout) also did unfortunate damage to the Delicious brand, sending panicked users to competing products.
But ultimately the real challenge here will be the technology. During my time at Delicious we rebuilt the entire infrastructure to deeply leverage a number of internal Yahoo technologies. It’s all great stuff but not exactly easy to remove or replace. Yahoo may have to license some of this technology to the buyer. I’m not sure they’ve done that before.
Open sourcing Delicious
This is a seductive concept but doesn’t make much sense. As in the case of a sale, they would need to unwind a bunch of proprietary technologies before this could happen. And open sourcing a complex product isn’t as simple as switching your GitHub repository from private to public. It involves a lot of work to clean up and document the source. For Delicious this would add up to a huge effort that would be hard to justify purely on a financial basis. Even then, it’s not clear how an open source social bookmarking system would work, given that much of its value comes from being centralized.
Donating Delicious to the Library of Congress or the Smithsonian
Now we’re getting closer. While it is folly to assume either of these institutions could take over Delicious and keep it running as a viable service, it does seem like they would be interested in preserving the Delicious corpus and making it available for research.
I love Delicious for many reasons, but chief among them is that it is the Internet’s memory storage device. In the 7+ years of its existence it has recorded the collective online journeys of millions of users during a time when the Web was evolving dramatically. Those memories are irreplaceable and have enormous value both to their owners (the users) and to society.
And so this is where we end up: Delicious may or may not have a future as a service, but regardless we can still “save it” by extracting and preserving its collective memories. There are two ways to do this:
- Yahoo could proactively release the corpus of publicly-shared bookmarks and tags. This could take the form of a mass data dump into the public domain, or it could be via an agreement with an institutional partner (much like Twitter did earlier this year).
- The Delicious user community could organize to save the data themselves via a coordinated harvesting project.
The second approach could produce valuable results but would require no shortage of cleverness in order to avoid triggering rate limiters and other abuse mitigation mechanisms. Even then it’s not clear that the entire corpus is currently accessible in this manner. The first approach would be much more direct and complete, and would likely earn back for Yahoo some of the goodwill it has recently lost.
Separately from the public data, there is the issue of personal user data. While Delicious has long had a bookmark export tool, other pieces of personal or private data are not easily exportable, notably the user’s Inbox (links shared with them by other users) and their Network (links saved by users they follow, as well as the list of those contacts — it’s basically Twitter for bookmarks). The user community has already started working on this problem: former Delicious engineering lead Josh Whiting has written a Ruby script that exports your Inbox. Ideally Yahoo should provide official tools for exporting this data.
In conclusion, releasing the public corpus is the right thing to do for Delicious, for Yahoo, and for the Internet. If a sale proves difficult — or even if it succeeds — I hope Yahoo will take this path and I would strongly encourage them to do so.
Today Google launched Google Instant, and make no mistake: this is big. It’s far more than just a new fancy interface. It’s a fundamental change to a user interaction model that’s been largely unchallenged for years. It also represents a significant financial and technical commitment on Google’s part (all those new searches and suggestions aren’t free). Last but not least, it clearly demonstrates that Google still has both the means and motivation to challenge a status quo they themselves helped create.
By all accounts this is a bold and brave innovation. Which is why it may surprise you to learn that Google Instant is actually five years old. Yahoo built it back in 2005.
Yes, Yahoo. In 2005.
I can see you’re confused. I don’t blame you. Let’s back up.
Back in 2005, prior to joining the del.icio.us team, I was a product manager on Yahoo’s search team. I led the (truly awesome) team that built and shipped Yahoo’s first AJAX search tools. Our biggest project and crowning achievement was LiveSearch: a fully interactive search UI that produced auto-complete search suggestions and displayed instant search results — in real-time as you typed.
It was slick. It was extremely useful. It was a potential game-changer. And it was almost exactly the same product as Google Instant.
There were differences of course. The search suggestions were displayed on the left side of the results, instead of above them like in Google’s implementation. This was to avoid the distraction of shifting the result set down the screen as the user typed. The relevance of Yahoo’s results at the time were likely lower than what Google is capable of displaying today, but they were still quite good. And the layout clearly needed further optimization, since the default Yahoo SERP at the time was top-heavy with ads and in-network content “shortcuts” (some of which were quite useful, others not so much).
All in all, my assessment is that Google Instant is a superior implementation, but LiveSearch was close. Damn close. It’s something Yahoo should be proud of, and it’s a shame more people don’t know about it.
There is basically no trace left of LiveSearch these days, but here’s an article about it from Google Blogoscoped. And below is a screenshot, taken from the same article:
UPDATE: and here is a YouTube video of LiveSearch in action!
So what the frak happened? If this innovative product was built nearly 5 years ago (it actually shipped in early 2006), why hasn’t anyone heard of it?
The answer is (mostly) in the screenshot. Yahoo would not let us ship LiveSearch on yahoo.com or as a part of Yahoo’s search engine. Instead we were only allowed to launch it on AllTheWeb, a smaller, lower-traffic search engine that Yahoo had acquired years earlier and largely left to atrophy. (In comparison, Google just launched it on google.com. Boom.)
You have to remember that, at the time, Yahoo’s search business was doing just well enough that there was very little institutional appetite for product risk. As a result, “big” or disruptive ideas were too often left to whither on the vine. By focusing on the local maximum, Yahoo unwittingly traded innovation for incremental optimization.
LiveSearch was thus relegated to a tiny test bucket of users who didn’t actually use Yahoo’s search product (or any modern search engine). Usage data from this flawed test was used to internally evaluate its success in comparison to the model it was actually trying to disrupt. Lacking high-level support for its larger vision and starved for resources, LiveSearch was understandably put out to pasture.
To be fair, this all happened several years ago and user attitudes have changed since then. An interaction paradigm that today seems brilliant was back then perhaps not as immediately accessible to users. I also look back and feel that I could have more effectively “sold” the vision to Yahoo’s management and fought for it. Ah, the benefits of hindsight.
But still, here we are, five years later. And what is essentially the same product is launching to most of the world’s online audience in one fell swoop. Clearly an opportunity was lost somewhere along the way.
While Yahoo’s reputation as an innovator has suffered in recent years, I can tell you first-hand that there was no shortage of amazing ideas inside those walls, some of them way cooler and more disruptive than LiveSearch. Nor was there any shortage of smart, driven, and creative people on the payroll. I worked with a bunch of them.
Instead, what was lacking was the ability to bust past that local maximum. Google may have challenges of its own these days, but today it clearly demonstrated that it still has the organizational courage to challenge its own preconceptions. As I read the coverage and play with Google Instant, I can’t help but wonder how things might have gone if Yahoo had shared this trait five years ago. It could have been pretty damn great.
