July 31, 2008

Well, the verdict is in: Cuil has a great publicity team, but operational execution & brand damage control are clearly not strong points. 50 million queries in 48hrs is pretty impressive.

Since they raised $33M in capital, I can see them having a second shot at making Cuil cool.

So unlike the relentless bashing, I’ll go ahead and list 10 improvements that can bring this search engine back into the reckoning.

  1. If you are going to crank up the richness of these search results, you are inevitably going to crank up the complexity. Accept that as a fact and counter it by offering flexible layouts. Different prisms, same result-set, same query. Stuffing the magazine layout down everybody’s throat is not how you differentiate or breed loyalty. 2 columns vs. 3 columns is just not enough.

  2. The image matching is [NSFW] eccentric. People get some weird matches. Try a search for Mars… and for Mars.com it associates the movie DVD cover for Red Planet with it. Fix this temporarily by offering a zero thumbnail layout.

  3. Cuil seems to pride itself on giving people different” results from the big G. I would like Cuil to attempt being the first successful asynchronous search engine. Let users have the option to sign-up, post queries, and forget about them until Cuil is ready with the richest possible dataset. Email back the user when ready. Yes, that changes the search experience significantly - but i don’t think it is too much of a stretch to ask users to put in a multitude of research terms and get back a basket full of super-rich and super-matched search results. Cuil is in a hurry and I don’t know why.

  4. Get hooked into a Content Delivery Network like Akamai for your images. That should help you scale better. I assume the eccentric image association is done through some inference algorithm that decides to associate a particular image thumbnail resource from the cuil inventory to the result itself. On the bright side, it does sometimes admit defeat and give up ;)

  5. Is it too hard to have a Did you mean xxxxxxxx?” query correction feature a la G.

  6. Cuil’s crawler NEEDS to recognize and index recent events quickly. The fact that IBM plans to buy iLog still hasn’t registered with the engine. Without this capability, the index cannot build user confidence. I don’t want an engine that is oblivious to changes in the infosphere.

  7. Now that the Cuil launch validated the Internet’s thirst for an even better search experience, get one of the founders back on the VC beat and raise some more funding. Saying Wow, that was intense.” should be reserved to a more private activity. I would have said, thanks for demonstrating your thirst & intent, now let’s go get some more funding, dilute our share and focus on building the best search engine in the world because people want us to.”

  8. Rebrand. Cuil is short, but weird. It is designed to cause spelling errors with the double vowel in the middle. The etymology of the word is vague as well. And most importantly, users now have a predominantly neutral-to-negative association. I think the best word to describe the fall-out of this launch is indifference”. So, please put the ego aside, think of a new word, and give yourselves a real 2nd shot at this.

  9. Consider giving Cuil (or whatever you end up calling it) a social component. Yes, you don’t record any search activity etc. etc. etc. But seriously, one hell of a way to make a more RELEVANT search engine would be to monitor people’s queries (not WHO they are or WHERE they are) and have the indexer figure out why so many requests are suddenly coming in for this subject. The indexer should then make improving those result-sets and fulfil the infosphere’s immediate and dynamic demands. There will always be a subset of people who want and get the scoop before it is the scoop. But for the majority, Cuil’s new found awareness will satisfy their curiosity and foster trust.

  10. 50m queries is a good amount of data to assess how Cuil can improve internally. Hold your people accountable in a positive way vs. having a victim complex. Beyond what we, the users, can see - I am sure there were bugs, operational bottlenecks & unexpected patterns. Get it done and put on a better show next time.

Go n-eiri an t-adh leat.




Previous post Videogame Nostalgia When it came to gaming, my experiences began with the Atari 2600. River Raid was a favorite. Boxing was fun too. And Pitfall. There was also this Next post Ode to Freddie Flintoff To get my Cricket fix, I usually purchase a web feed from Willow TV. They are a good setup, they’ve stablized their act over the last year and the