March 2010 Archives

Paul English on early Kayak and how hiring impacts traction


Paul English is the co-founder of Kayak, the travel search engine. He relates how Kayak got traction, including an early deal with AOL and how they launched with 15K searches/day. Paul also gives advice on recruiting and explains how it impacts getting traction in a fundamental way.

For more interviews, visit the Traction Book site.

My Duck Duck Go reddit ad by the numbers

My reddit ad (above) has yielded the highest ROI (by far) compared to the other ad platforms I've tried, which includes Adwords, Yahoo, Bing, Facebook, Myspace, and StumbleUpon. Of course your mileage will vary depending on your ad, product, target market, etc., but my basic message to you is you shouldn't overlook reddit self-serve advertising.

My ad ran for 13 days, from 3/7 to 3/20. It cost $650, and I spent $50 per day. In total it had 1,288,378 impressions (282,732 uniques) and yielded 20,700 clicks (18,420 uniques).

That's a CTR of 1.61%, or 6.49% per unique redditor. This CTR works out to a CPC of 3.14 cents, or 3.53 cents per unique visitor, and a CPM of $1.98, or $2.30 for uniques. 

This is actually my second reddit ad. The first one looked exactly the same and was run during their beta period back in November. The first day that one came out it had a unique CTR of 13%! This rate eventually came down to about the same unique CTR% as the second ad.

Here were my takeaways from the whole experience:

  • Redditors actually try out your site. 3c per unique visitor is pretty good in and of itself, but it's all worthless unless they actually try out your site. For example, you can get 5c unique visitors from StumbleUpon (presumably in a similar demographic), but StumbleUpon visitors never would try out my sites. Reddit visitors did try out Duck Duck Go. 

  • Redditors actually comment on your site. A unique feature of reddit ads is that redditors can comment on your ad on reddit. You can turn this off if you want, but I don't recommend it. Here's why. That is my ad's comment thread, which has 656 comments! (About half are from me though.) On this thread is immensely useful feedback. I fixed bugs, got feature requests, got first impressions, etc., and perhaps most importantly, was able to engage with my users in almost real time.

  • I think it helped my other Reddit submissions. At the suggestion of one of my ad's comments, I made a Ask Me Anything submission. At another suggestion, I finally pulled the trigger on no longer storing IP addresses, which I submitted to reddit here. This last submission went to the top of the technology sub-reddit and made it to the reddit front page as well, to somewhere around #5.

    These two posts yielded more traffic than my entire ad, and I think they did so in large part because of the ad. My hunch is people recognized the site and were more likely to vote it up as a result. Two concrete examples.

    First, I doubt the privacy submission would have made it to the front page if not for the ad because doing so requires a certain upvote velocity that you don't usually get when people aren't familiar with you.

    Second, the other day @kn0thing submitted my cuil parody to reddit. On its' comment thread are two comments from redditors praising the search engine. People read these types of comments and take them seriously, and they wouldn't happen without converting reddit users into Duck Duck Go users beforehand.

  • CPM and CPC varies widely by day. As you can see from these graphs, some days were very different than others. This has to do with how reddit sells advertising. They split up the pool based on how many $ people bid for that day. So if you get a bigger piece of the pie that day, you'll get better CPM. My takeaway is to run the ad over a longer period. You can always stop it if you're getting diminishing returns.

reddit CPM graph.png
reddit CPC graph.png

  • My ad did better than others. You can tell from these graphs that my CPC was less than the average by a lot because my CPM is during the high Mar period (higher than average) but my CPC is still way less than average. I think the reason for this is two-fold.

    First, a search engine ad is a good fit for reddit ads in general. It has broad market appeal and redditters in general like trying out new technology.

    Second, I think the ad is particularly well structured. The circular duck icon draws your attention, is contrasting to site colors, and sticks out because it is a circle (as most images are square). I believe the title also has appeal.

  • People are still mentioning the ad to me. I've received a bunch of regular site feedback (through my feedback system) where people mentioned they discovered Duck Duck Go via the reddit ad. Also, at least two people I've initiated conversations with told me they've been using the search engine after seeing the reddit ad. These types of conversations have not happened with any other ads I've placed.

Finally, I want to be clear that even without the successes of my follow-on reddit submissions, I still believe reddit ads have had the highest ROI (and still by far) for Duck Duck Go. 

Well, that's it :). In short, consider trying reddit self-serve advertising. You can test for as little as $20.

My case for Google fiber for the Philly startup community

This is a submission to Gigabit Philly for Google Fiber for Communities


What would hundreds of startup companies do with Gbit Ethernet connections? They'd do a lot of everything, e.g. crawling the Web, broadcasting video, transferring massive data sets, and lots of stuff I haven't thought of yet.

I currently have a much slower fiber connection for my business (using FIOS). I crawl about a hundred million domains a month on this network. Access to this fiber connection has certainly helped my startup get traction. But if I had cheap access to Gbit Ethernet, needless to say, bandwidth would no longer be my bottleneck.

Multiply me by several hundred. 
That is my case for Gigabit Philly in a nutshell. 

The case for startups

Startups drive economic growth because they commercialize innovation. They create jobs. They increase standards of living. Giving startups Gbit Ethernet will accelerate these processes. It's the biggest bang for the buck by far.

The case for a big city

The suburban digital town is old hat, as is the wired town in the middle of nowhere. I remember the same stories from literally fifteen years ago--let's take a place and wire up all the homes and call it the future. And what useful results have ever come out of those projects?

Urban centers, on the other hand, have been less touched by the digital revolution, and for good reason. They're a pain to deal with. There's entrenched companies, endless politics, large tax bills, old buildings, etc. I moved out of the city in part so I could avoid these things and in part so I could get things like FIOS.

And that is exactly the case for putting Gbit Ethernet in a big city. It's both challenging and needed. Many big cities are in desperate need of revitalization. There is a real future in new forms of urban planning, and digital connections will be a part of that future. 

Perhaps most importantly, though, big cities are where startup communities live. So if you believe the case for startups, then you should bring fiber to where they reside, within big cities.

The case for Philadelphia

If you can do it here, you can do it anywhere. Smaller cities should be easier to deal with because they're smaller. And the five larger cities don't need it as much.

That's right, Philadelphia ranks #6 in the list of biggest US cities. We're probably the biggest city you hear least about. And we've been losing population for the last 50 years. Put differently, we're most definitely in need of revitalization. 

To make the case more compelling--we really want it. The city of course wants it (who wouldn't?), but the startup community in particular wants it. 

We want it so bad that our biggest startup group, Philly Startup Leaders, has spent their entire cash reserves on Gigabit Philly. Other startup groups and individuals (me included) have donated so much to the cause that the initial $5000 contribution is now in the minority.

To show the city actually supports our efforts, City Councilman Bill Green and City CTO Allan Frank each personally contributed $500 of their own money. 

The case for Gigabit Philly

In short, bring Gbit Ethernet to startups in Philadelphia because we need it, want it, and it will make a real and lasting impact. 

If it were me, I'd sprinkle it all around. We're a large city, and there are startup centers throughout, e.g. co-working space, the Science Center, and the Naval Yard.

Restated: the case for Gigabit Philly is hundreds startups using this fast connection to innovate quickly and help revitalize our city.

Satish Dharmaraj on Getting Traction


Satish Dharmaraj is a partner at Redpoint Ventures and lead investor in posterous, former angel and was CEO of Zimbra, which sold to Yahoo ($350M). Satish gives the investor perspective on getting traction and explains how Zimbra got traction through an open source community, PR and product/market fit. Satish also comments on idea generation, posterous and

For more interviews, visit the Traction Book site.

Poll on Hacker News Brand Awareness

I made this poll on Hacker News a little over a week ago. For some unknown weird reason it didn't make it to the front page, but it still got enough votes for the (non-scientific) results to be interesting.

The poll asked "Which companies have you a) heard of and b) know basically what they do?"

Here were the results (% of respondents that met the above criteria):
  • Bingo card creator - 82%
  • Directed edge - 40%
  • Bluefrog gaming (previously Draftmix) - 21%
  • Duck Duck Go - 71%
  • Mibbit - 21%
  • Mixergy - 79%
  • Newscred - 8%
  • Revizr - 1%
  • Tarsnap - 59%

I chose these choices carefully.
  • Revizr - this was the top "review my startup" post (by points) that I found on And yet no one had heard of it. And the people who voted were probably more in the know than just "front page users" because they're people who review new or older stories, i.e. visit pages other than the front page on HN.
  • Bingo card creator, Mibbit, Tarsnap, Duck Duck Go - these are startups by prominent Hacker News members who also have stories about their startups occasionally on the front page. Within this group, Bingo card creator has the most straightforward name, Mibbit hasn't had a lot of recent posts, Tarsnap has a lot of comments about it and Duck Duck Go has had a decent amount of recent posts on it.
  • Directed Edge, Bluefrog Gaming -  both YC startups, i.e. associated with the program that runs HN. Bluefrog is an older one and Directed edge is newer.
  • Mixergy - Mixergy has been posted a lot lately. This seemed to definitely make a difference because its name isn't that straightforward.
  • Newscred - they just had a top story like the day before I posted this. That didn't seem to make that much of an impact in and of itself, sort of like Revizr but before an exponential decay.
OK, tl;dr :)
  • Just a few posts, even popular ones (Newscred/Revizr), did not create brand awareness.
  • Recent saturation (Mixergy) can work.
  • Persistent mentioning and community participation definitely works.
  • Having a very straightforward name (Bingo Card Creator) is helpful.
  • Being associated with the core of the community (in this case being a YC startup) contributes to brand awareness.

Justin Kan on Getting Traction


Justin Kan is a founder of, the place to watch and record live video. started out as a reality show, and before that the founders had a YC-backed calendar startup (Kiko). Justin explains how they made two major pivots and translated initial press about the show into the current business. Justin also talks about scaling and idea generation.

For more interviews, visit the Traction Book site.

Angel investing portfolio scenario planner spreadsheet


irr.jpgSeveral friends have told me angel investing is not a good way to make money. If you can do the YC/Ron Conway strategy that is one thing, and if you just want to give back that's another, but doing it at my level is a waste of money.

I do want to give back, but I also want to make money with my angel investing. So I've been thinking a lot about strategies to employ to fulfill both goals. To that end, I've been creating spreadsheets that I'd like to share with you.

First, check out the Angel investing portfolio scenario planner (Google doc spreadsheet). To get an editable copy, select 'Make a copy...' from the File menu (or Ctrl+Shift+S).

This spreadsheet shows you the internal rate of return (IRR) of a fictional set of 10 investments of $25K. Really the amount per investment doesn't matter.

The investments have 5 year time horizons, i.e. the payout happens in year 5 (row 15). And each investment outcome can fall into 5 buckets: lose everything, return money, medium, medium high and high. The payouts of these outcomes are in column B, and the % breakdown (how many investments are in each bucket) are in column C.

If you play around with column C you can allocate different amounts of investments to each bucket and see the effect on IRR (G15). If you want you can also mess with column B to change the payouts of the buckets.

The following are some of my takeaways. I'd love to know yours.

  • My ambitious goal is for a 30% IRR. The default scenario is one way to get that, which is like the 1/3, 1/3, 1/3 a lot of good investors report. In this case, it's 40% lose everything, 30% return money, 10% medium, 10% medium-high and 10% high.

  • The return money bucket really doesn't influence overall IRR that much. If you change the default to 70% lose everything, i.e. take all the return money % and move it to lose everything, it only changes the IRR by about 2%. It certainly matters psychologically and it may free up more money to invest, but in this static scenario it doesn't really move the needle.

  • If you change the outcome of the high bucket (B11) much higher, you can see what it is like to invest in the next Google. You can also see that if it is just one of your investments that's fine, i.e. the oft mentioned 1 in 10 home-run strategy.

  • If you want 30% IRR, you really need to get a high outcome in there. This leads me to again wonder whether I should only consider events where I can see a realistic path to that high exit going on what is put in front of me.

  • The time horizon really does matter a lot. I added rows 16 & 17 to show the difference between a 5 year time horizon vs a 4 and 3 year one. VC-backed companies have significantly later exits on average. These scenarios make me wonder if I should really be looking at companies (at least in part) that could make it without VC.

Garry Tan on Getting Traction


Garry Tan is a co-founder at posterous, "the dead simple place to post everything." Posterous was part of Y Combinator in the summer of 2008, and immediately got traction after being launched on Techcrunch.  Garry explains how product development, idea generation, YC, previous experience and execution contributed to their success. He also gives advice on press and customer support.

For more interviews, visit the Traction Book site.

Hack Hack Go



I want to make Duck Duck Go a better search engine for programmers like me. If you're a programmer, I'd appreciate your feedback and ideas.

Duck Duck Go is intended to be a general purpose search engine and that isn't going to change. Our user base certainly reflects this purpose, i.e. is quite varied on every metric I've tried to measure.

Yet there are certain search niches like casual research where Duck Duck Go really excels. I'd like programming to be one of those areas.

To that end, here's what I've got so far.

  • A general search engine. The good news here is I know a lot of programmers who use it as their primary search engine. It works and (at least some) people really like it. I'm always willing to add new features whose absence are preventing people from switching. Currently on that list are some maps and images.

  • Zero-click Info. There are red boxes above links on some searches with info you can get without clicking, i.e. on-site. We have a lot of info that is specific to programming topics. Of course we have Wikipedia, e.g. Dijkstra's algorithm. But I've also added software sources, i.e. github, freshmeat,, versiontracker, and sourceforge.

  • Category pages. I've mined sources to create to useful topic lists for browsing/learning, e.g. Search Algorithms.

  • Disambiguation pages. I've created pages to help you isolate programming topics in common query terms, e.g. cookie links to HTTP cookie, which has results more geared toward that meaning. There are also programming specific disambiguation pages, e.g. nearest neighbor.

  • Crowd-sourced links. I also mine links from crowd-sources sites, e.g. coroutine.

  • Wikipedia paragraphs. I've deep-indexed Wikipedia at the paragraph level. You don't have to match a topic nearly exactly anymore to get some Zero-click Info, e.g. python switch statement. This is way more than a regular search index, as it is sub-section/section/title aware and uses some NLP for relevancy. I hope to make that matching algorithm even more sophisticated over time.

  • Bang. There are a few hundred !x shortcuts that can be used, e.g. !cpan Net::DNS

Here's what I'm thinking of doing.

  • O'Reilly Paragraphs. I think it would be awesome if I could index all O'Reilly books at the paragraph level, like I've done for Wikipedia. This content is well-written, encyclopedia-like, is largely in paragraph form, and has surrounding contextual information (section titles, etc.) that will make the relevance matching excellent. Problem is, I don't know anyone at O'Reilly. I think it's a win-win because it can link right to their Safari product or individual book pages. And I don't think it canabalise Safari because you're getting people in a very different context (when searching). Anyway, I thought I'd start by writing them an email. I did that and haven't heard back yet.

  • More topic sources. I'm going to add man/info pages, so you can type in a command and get a description. I could also do packages for distributions/languages in a similar manner if people think that would be useful to them. I've explored indexing these at the paragraph level, but the content doesn't seem to work well for that purpose. Other, more general sources, may be incidentally useful to programmers like Amazon product descriptions. I'd love your thoughts here.

  • Bang documentation. The current bang commands aren't documented. I'll document them as well as add more that are useful to programmers. Any you want?

  • Zero-click Info by IM. I'm thinking of making a chatbot that will respond to you via IM with Zero-click Info (and links). So you send it a search query and we'll send you back a description along with a few links. Would you use that?

  • API integration. I wrote the Perl binding for Wolfram Alpha. I'm exploring ways to use it to integrate good WA content. I'm open to using other APIs, but I'd strongly prefer to get dumps instead so I can ensure speed. Another one I'd like to integrate for programmers is (previously

That's where I'm at right now. If you're a programmer, my questions for you are:

  1. Do you find the above compelling?

  2. Do you have any particular feedback/ideas?
Feel free to comment below, on HN, on reddit, or email me directly.