Spokeo Blog | Find People through People Search

We just got on the MSN front page couple hours ago.  As expected, hundred of thousands of people are flocking to Spokeo right now, and the incoming traffic is beyond our imagination.  The site has ground to a halt, and even we administrators cannot access it at the moment.  We don’t have a solution for this because we cannot stop people from coming to the site, and buying ten more servers today will not help our immediate needs.  As a result, we will just have to wait it out until the traffic subsides to a more reasonable level.

We are sorry about this service disruption to our existing users.  We’ve always planned to have 50% more capacity available to handle unexpected traffic spikes, and due to our planning, we have been able to weather through all the PCWorld, Wall Street Journal, Sina, and other mainstream press around the world.  However, getting on the third largest website in the world (msn.com) is a little bit too much for us to handle today.

As most of you would agree, the Web 2.0 era is long over.  Web 2.0 was categorized by ad-based business models.  Everything has to be free, and advertisers would somehow pay the bills.  We entrepreneurs chased after user counts, and the ad dollars were assumed to flow in.

We started Spokeo with the standard "Google AdSense" ad model.  We thought that if we’ve built something cool that people like, we would earn enough to cover our operations and someone would buy us out.  This sounds naive, but that’s what we were dreaming, and I bet that’s what every TechCrunch reader was thinking back in 2006.

The pursuit of user counts works fine if the ad CPM (cost per thousand impressions) stays at the 2005 level.  Back then, advertisers hadn’t figured out that user-generated content converts poorly into sales, so these advertisers paid more for the benefits of the doubt.  Take Google as an example.  Google paid MySpace $900 million dollars in 2006 to take care of MySpace’s ad inventory, and all Google got in return is the lesson that user-generated content doesn’t monetize.  Advertisers aren’t dumb, and they won’t pay for ads that don’t work forever.  CPM for user-generated content took a nose dive around 2007, and all Web 2.0 sites with ad-driven business models were affected.

Spokeo cannot escape this fallout.  We ran ads for 3 months in the beginning of 2008, and we quickly realized that even if we grow the traffic by 100 times, we still couldn’t cover our basic costs.  Facing the hard reality, we made one of the biggest business decisions by turning Spokeo into a premium service.  It’s a last-ditched effort to save our business, and we didn’t know what to expect.  Fortunately, there were many people who like Spokeo to pay a little to support us, and now we should be able to weather through this economic downturn. 

Here we would like to thank all of our users for supporting us, and we would continue to introduce new and innovative products in the near future.  We got something huge scheduled for January 2009 (big projects always take time), so stay tuned!

Yesterday I had the opportunity to meet some interesting new start-ups and demo Spokeo at the Plug and Play Expo in Sunnyvale. Plug and Play’s Expo is hosted twice a year to give up and coming start-ups a chance to gain exposure for their products or services.

It was nice opportunity to get out of the office and talk about Spokeo. The overall response to our service was very positive and I gathered valuable feedback from other participants at the expo. Spokeo’s table was between two very different start-ups, one focused on user generated content and the second focused on limiting access to content. Collegewikis.com and ScreenerKey. Collegewikis enables students to share their notes online. Similar to wikipedia the notes are editable so students can add to the body of information about a class or a lecture. ScreenKey was a tool for parents to monitor the amount of time their children watched television everyday, when the alloted time for TV viewing is over the device will start to emit a audible alarm. The spirit of innovation is alive and well in the Silicon Valley.

I hear that question all the time. And I have finally made my peace with it. User growth is good. However, obsessing over user growth in a product under development is bad because it breeds unhealthy stress and leads to shortsighted decisions.

I read an interesting interview from Eric Schmidt months ago. In it, he mentioned that early on, Google deemphasized user growth. If Google was making noticeable improvements in its product month to month, why should Google go for users today? Why not grab them more easily at a later point?

I am adopting a similar mindset with Spokeo. We planned two major features today, both of which will greatly improve Spokeo in the future. I see growth as much more natural with each successive Spokeo release. The market is not closing up anytime soon, and we have no external pressure, so why not maintain a product focus? As the saying goes, the product is the marketing. As proof, our conversion rate increased 3x when we released our new skin yesterday.

Once again, user growth is still good for all the obvious reasons. We still need to get the word out. We just can’t benchmark ourselves against all the typical success stories when our product is unfinished.

There’s a proverb, "Birds of a feather flock together" (Chinese has a similar proverb, "one takes on the color of one’s company"). When you see many of your friends have started their own ventures, you’ll feel an inexplicable urge to do the same as well. These friends are more than drinking buddies; rather, we exchange ideas and experiences to help each other grow. Here I would like to introduce one of Spokeo’s friends - Midomi.

Midomi is a karaoke social network. You can sing, and others can comment on your performance. They got a lot of nice people in their community, so even if you are not pro, you won’t get trashed. My performance is here (you can add it into Spokeo to track my future performances). Warning: I like Chinese pop music so you probably can’t understand what I am singing. Also I am not exactly pro. Anyway, if you are interested to see how Midomi works, or if you find it entertaining to see what my voice is like, go check it out.

Midomi is more than a social network. It is also a voice search engine. Just sing a song like "Happy Birthday" into Midomi, and it can correctly find that song in seconds. The accuracy is quite astounding, so you should definitely check out that unique feature.

How Midomi’s search works is that it can recognize "pitch" (as opposed to speech). It does not attempt to interpret what people are saying; however, by focusing on pitch recognition, their engine can detect the melody better. Obviously I am not an expert on this subject. Just try it out, and you’ll know what I mean.

Lastly, we are going to upgrade our system tomorrow. You’ll experience a short downtime, but you’ll appreciate the new changes.

If you happened to check out our company page today, you’d notice that Spokeo got a great marketing boost in Mital. She is very lively, a fireball of energy, a human Mount Vesuvius. She is exactly what we are not, so it’s a great fit, and we’re super excited that she is with us. A very small aside: HARRISON YOU BROKE THE ABOUT PAGE AGAIN.

Taking a step back, I now see that we have hired the same way for all three very different positions. It has taken me a while to internalize our hiring philosophy to the point where I can actually verbalize it.

Our hiring philosophy errs towards false negatives. By definition, a startup is an underdog for its first few years. What we lack in resources, reach, and an established user base we must make up in A players. Our leverage resides solely in the people we have.

We also only hire when we are in pain. When in pain, we know exactly what we are hiring for. When in pain, we know we are acquiring a resource that will be efficiently utilized. The downside is that our new hires hit the ground running. And they run fast.

Fundamentally, a clash exists between false negatives and hiring in pain. The first quality tells you to take your time shopping, while the second creates a situation where you really don’t want to take your time. Alas, the joys of a startup.

Ten rules is a wonderful post for startups written by a great web entrepreneur. I’ve read it many times and even taped it to my wall, but I am still not happy with our performance on several rules. I only present the rules here. Please refer back to the original post for descriptions of each rule.

1: Be Narrow
Fail.

Our main message is ‘Track Your Friends’ (which will be more apparent in future releases). At first, it seemed narrow enough, but now some questions need answers. Track all your friends and grant continued account access to Spokeo? Track only your public friends?

As a startup, we also can’t support every service under the sun. Track friends across the major social networks? Track friends across every [blog or photo or video or music] network in [some category]?

2: Be Different
Pass.

Oh, yes. Very different. Makes life difficult at first.

3: Be Casual
Pass.

4: Be Picky
Pass.

We have increased our pickiness across the board.

Pickiness is a lesson that cannot be learned in a general sense. Trained as engineers, we knew from the beginning how to be picky in engineering and technical recruiting. However, engineers are probably least picky when it comes to features and abiding by rule 1. Engineers tend to want to build a lot of stuff.

5: Be User-Centric
Pass.

We completely obsess over UI. In fact, our skin is getting optimized even further as I type.

6: Be Self-Centered
Pass.

We are finally all users of our own software. I, myself, didn’t find Spokeo attractive enough until a month ago. That was a big confidence booster for me.

7: Be Greedy
Not sure I agree with this rule.

8: Be Tiny
Pass.

Rule 4 enforces this rule.

9: Be Agile
Pass.

10: Be Balanced
Fail

Some of us are more balanced than others. I tend to have spurts where I live work, eat work, pester co-workers about work after work, and even dream work! They tend to happen when Spokeo enters a critical phase.

Overall Grade
We passed 7 rules, failed 2 rules, and punted on 1 rule. Normally that would be a 70%, but since rule 1 is so important, I give Spokeo a 50%.

Slow creeping growth… Does that even sound like a good thing? Yes! After months of reworking the idea, tweaking the usage flow, and streamlining the interface, we have advanced out of stagnation. During our monthly review today, we officially declared that we are indeed seeing some pickup.

Stagnation (stage cleared)
Slow Creeping Growth (current stage)

And due to that, we all had a few beers at 4pm and a lovely dinner afterwards.

Must have been funny seeing a table of guys drinking to ‘Slow Creeping Growth.’

Cheers.

This past week I visited some family in Asia. For lunch one day, we went for a meal of spicy noodles and pork dumplings at a small family restaurant. When we were done eating I raised my hand to ask for the check. A second later, a waitress said “980 NT” (which comes out to $30) and approached with 20 NT change in hand, correctly anticipating a payment of 1000 NT (there is no tipping there).

Now, that is what I call great service. No need to flag down the waitress for the check, wait for her to calculate the check, and then wait again for her to bring the change. The difference was completely in the anticipation.

At Spokeo, anticipating the problem is how we operate. We provide a demo to explain the concept. Then we try to preempt difficulties with tooltips. If support is required, we try our darned hardest to resolve it in 1 email.

I seem to be making a lot of analogies these days. In a previous post I compared Spokeo to a car/boat/sled. What’s next?

Once in a while I come across an awesome user. Sometimes this user helps us nail down a feature set with his feedback. Sometimes this user writes an FAQ section more clearly that we originally wrote it. This time, flameape was able to find two bugs that had escaped our testing team and thousands of our users.

Very good work, indeed. And thanks for everyone’s patience and understanding!