Skip to main content


warning: Creating default object from empty value in /home/kevin/public_html/modules/taxonomy/ on line 34.

Playing With Data for Fun and Profit

I've been meaning to share this for a while now. I was asked to talk at Howard University to their digital business class, which focuses on the use of Social Media, Mobile Apps & Platforms, Data Analytics, and Cloud Computing as strategic assets to be utilized in business.

I wasn't really sure what to do, so I decided to tell some stories about how I've used data in my businesses.

I also built a demo to let the students learn about and play with sentiment analysis in real-time. It used the Movie Review corpus from Cornell. It's a very primitive keyword based system but I thought illustrated the concept well.

Sentiment Analysis Demo

Reverse Engineering Startup Press: How and Why TechCrunch Covered My Launch

After my startup's launch was covered by TechCrunch I was asked by a lot of people how it happened. People wanted copies of my pitch to learn from. It was as if I had discovered some arcane secret. But I didn't believe that. All I had done was read a couple blog posts from other startups with copies of their pitches (Thanks Jason L. Baptiste, Vinicius Vacanti, Leo Widrich) and one journalist who shared his thoughts (Thanks Sean Blanda). So instead of saying here is the magical way to get press, I wanted to find out what journalists really thought of my pitch and how it could be done better. They receive hundreds of emails from people trying to get their attention and I wanted their advice and expertise. I also wanted to know why the author who covered me on TechCrunch chose to write about me. How did I win the press lottery? In this post, I will share their thoughts and opinions.

Let's begin with the actual pitch:

Subject: Exclusive Story Opportunity: Could Twitter replace review websites?
Hey Eldon,
We talked briefly at TC DC meetup and I showed you a quick demo of my startup: Review Signal.

We're planning to launch on September 25 and you're the first journalist I've reached out to and I'd be happy to give you guys the exclusive on our launch.

What is Review Signal?
Our goal is use the conversations on social media to build a review site and bring a new level of transparency to the (sometimes? often?) shady review industry. We've started with the web hosting industry (probably the shadiest of them all) and plan to expand after launching (we're in data collection mode for a few more niches like domain registrars). At launch, we will be the largest web hosting review site around by at least an order of magnitude (maybe two) with over 100,000 reviews.

We also have a 45 second video explaining how our system works:

You're welcome to login to our private beta (username: xxx, password: xxx at ).

If you're interested please let me know and I'd be happy to follow up with any information or questions over email or phone.

What happened after I sent this email? I got a response from Klint Finley, who Eric Eldon had forwarded my email to.

Eric forwarded this to me -- I'd love to get the scoop on this. Are you available Friday?

I will skip sharing mundane details because everything of importance was handled over the phone after the first response. I told him about what I was doing and answered some of his questions. And the article came out the day we launched.

What did the journalists think?

The opinion I was most curious to read was Klint's. He is the one who wrote about Review Signal. I was thrilled when he said he would be willing to participate in this article. Please note this is his opinion and his alone. They do not represent the opinion of TechCrunch or any other publication.

Klint Finley

For a startup launch I think pitching to the press is really not all that hard: blogs like TechCrunch and The Next Web are always looking to cover new startups. The important thing is to have something worth writing about. Offering an exclusive always help grab our eye, but it's not always necessary.

The pitches I've seen fail are 'yet another...' for spaces that have long since ceased to be exciting and don't have anything else to sex them up (well known founders, a sizeable investment, a really interesting new spin on the idea...)

If the pitch sounds too much like "It's like Foursquare, but better and built by someone you've never heard of," I don't think that many people are going to pay attention at this point. If you don't have a name or a big investment you've really got to get that differentiation in there early.

I like the "it's like X for Y" type of pitch format, but not everyone does. But telling me "We're an HTML5 mobile app framework with some really advanced features" won't tell me much. Telling me "We're like PhoneGap for building Foursquare-style apps" is going to be a bit more attention grabbing.

The subject line of an e-mail is really important. To be honest, I didn't think your subject line was very good. I probably would have missed that e-mail entirely if Eric hadn't flagged it for me to check out. But once I opened the e-mail and saw that it was about using Twitter sentiment analysis to rank web hosts, I was really interested since web hosting, sentiment analysis and data mining are interests/beats of mine.

Which brings me to the importance of finding the right person for a pitch. This can be hard. Since you think of your company mostly as a b2c company and I write mostly about b2b and dev tech, it might not have been obvious to pitch directly to me. Again, luckily Eric saw it and thought, correctly, that it might be something I'd be interested in.

Some of the pitches I've gotten recently that I wrote about came directly to me and mentioned articles I'd written before. Sometimes I see this stuff and it's totally left-field and vague, like "I see that you once wrote something about 'cloud,' I too am in the cloud business." But sometimes I'll luck out and someone will e-mail me something along the lines of "you once wrote that it would be interesting to see a company doing X... well, that's what we're doing!"

So, in summary: do something awesome, make sure you explain your differentiation, put that in the subject line of your e-mail, and send it to the right person. Easier said than done.

Klint also pointed out that “this would be just one journalist's preference.” So I also decided to get a few more opinions and approached other journalists I've talked to in the past.

Ville Vesterinen and Miikke Kukkosuo, Arctic Startup

Was not too bad at all. Nice job! My critic:

1) It's a bit too long. Journos might just skip if it's too many lines to read.

2) I'd answer your headline by 'How:'

I have pretty similar thoughts as Ville.

Headline could be slightly sharper. 'How' is good. Or maybe even something edgier if you can think of something else.

I got a bit lost in the explanation, wasn't too easy to follow it. Depends a lot on the reader I guess how it's received - very big sites like TC might get so many mails that they move on quickly, a bit smaller ones probably would try to make sense of the message even if it takes a bit more time.

I think it would be enough to state the problem&market you're tackling, and why you will rock. Make it easy to read and understand, go straight to the point.
For example:
The review industry is shady. We will change that using social media conversations to bring a new level of transparency. At launch, we will be the largest web hosting review site around by at least an order of magnitude (maybe two) with over 100,000 reviews. Then we'll expand to more niches.

Carl Pierre, InTheCapital

I would take out the Could Twitter replace review websites part and just write exclusive story opportunity, maybe include your name in the subject line too so they remember who you are.

Um...the first part is cool, I would probably avoid mentioning the demo part, just say "we talked briefly at TC DC meetup and you gave me some good points on my new startup, Review Signal."

Then you should probably just say that you talked about providing an exclusive for your launch, and wanted to share the information with them before you go live.

Probably avoid the part that says, What is Review Signal, I would probably skip to something like...

Here is a quick refresher as to what we do:

- bullet point
- bullet point

Trust me, nobody likes slogging through paragraphs for info, especially as a journalist who gets bombarded with pitches constantly. Keep it succinct, short, and with key information in bullet points.

Again, this won't guarantee that you'll get a story, but it should hopefully significantly improve your chances.

From talking to the journalists, the basic recommendation seems to be:

  • Do something interesting
  • Explain why you're different
  • The subject line is your one chance to communicate why someone should care
  • Target the right journalist
  • Explain your idea clearly
  • Keep it short

I have to thank all these journalists for taking the time and participating. I had a great time talking to them and getting feedback about the pitch. I learned a great deal and hopefully this helps others too.

If you have any other tips please feel free to share them in the comments.

Review Signal Launch: Stats, Failures, Successes and Lessons Learned

It's been two weeks since Review Signal launched. The initial excitement of reaching it's first major milestone is over and I finally have a moment to think about how it went.

The Stats:

3,041 Visits
10,163 Pageviews
3.34 Pages per Visit
Average Visit Duration: 2m 16s
Bounce Rate: 5.52%
Visitors from TechCrunch: 727

Failures and Short Comings:

Social Sites - Reddit, HackerNews
It got very little traction in these communities which was disappointing because I thought these were the audiences most likely to appreciate Review Signal. Reddit was a gamble that it would be picked up, it's a huge community and it's hard to balance legitimacy and self-promotion. HackerNews I thought would be more appreciative of the problem I was trying to solve. I got a couple hundred hits and a few comments (15, many of which were my own replies). The commenters brought up some good issues, for example, the hosting recommendation doesn't work well for the HackerNews crowd. There are so many options for hosting that making a good recommendation for advanced users is incredibly hard to do with a simple form.

Contact an Expert Broke
It worked when I tested it, but somewhere between testing it and launch day it stopped working. Not sure how that happened, but it did. Some people will never hear back from me because it didn't save and send their requests for help properly. I am truly sorry about that.

I didn't prepare enough for launch day. I had been working towards this day for months and prepared an exhaustive list of things I wanted to accomplish on day 1. I completed maybe 15% of that list. I hit the major ones, but I missed a lot of low hanging fruit which could have helped make a bigger impact. The more human side of it was I got overwhelmed and I vastly underestimated how much I could do on the fly.

Utilizing Other People
I built Review Signal alone, but on launch day I had two friends take the day off and volunteer to help me do things. I say things because I don't think I effectively used their skills. Let me be clear, Zack and Danny are both fantastic people and what happened says nothing about their talents and abilities. I am incredibly lucky to have such amazing friends. I wasn't prepared enough to work with other people on launch day. I had a few vague ideas about what they could do to help but very few concrete things for them to actually do. I definitely didn't have instructions or information to make it easy for them. More planning and better communication would have made a big difference.

What went well:

TechCrunch covered the launch.
That was really exciting and sent a lot of very high quality traffic. The bounce rate was under 2% and people were staying for over 2 minutes. I thought the article did a fantastic job explaining what Review Signal was and the challenges it faces. Some might say it's a vanity metric, the number of people who reached out to me because of the article was incredible. It opened a lot of doors that I am sure would still be closed if I hadn't gotten covered in TC.

Secondary Press

The TechCrunch article also got me covered in The Web Hosting Industry Review (TheWHIR). One of the largest, if not the largest, web hosting magazine. I also had a couple other articles written by smaller startup blogs.

Nginx +

Once I knew I was getting TechCrunch coverage I got worried about how much traffic my server could handle. I setup Nginx as a reverse proxy and cache and it performed like a champ. I had people complimenting how fast the site loaded. As far as I can tell the server never blinked. It peaked at 60 active users according to Google Analytics. I tested my site constantly with which allowed me to test up to 300 concurrent users for free. The site was struggling under Apache with that load, but once I got Nginx in front of Apache, all my concerns faded away. I couldn't generate enough concurrent users to see where Nginx would actually start to slow down.

Lessons Learned and how I would do it next time

The most time consuming thing on launch day was crafting messages. Coming up with post titles, writing emails, IM'ing friends, and all sorts of other messages. Almost all of that could have been done before hand. I could have drafted emails, I could have written a few templates for IMs and message boards. I had the stories I wanted to use crafted but they weren't ready to simply copy+paste into messages. Each message also required some degree of personalization because nobody likes spam emails. Next time, I will have everything 1 click away from sending for launch day.

Test Everything. Again. And Again.

I should have done more thorough testing. I especially needed to make sure that the contact points with customers functioned properly. I got a lot of emails reaching out to me on launch day, but I missed quite a few hosting recommendations. The hosting recommendations probably had the highest potential value of any visitor to my site, and I lost all of them. Fail.

Paid Press Releases

I used PRWeb and their analytics tell me I got 15508 impressions and 221 reads. I have no idea what or where those impression numbers come from or how they are calculated. PRWeb sent 83 visitors which had the lowest stats of any referrer in terms of bounce rate, pages/visit and visit duration. But, that got syndicated across a bunch of websites like Yahoo. If the goal is strictly to get press, it was a waste of money. From an SEO standpoint, I don't know and it's incredibly hard to measure the value of the release.

If you have any ideas, questions or feedback I would be happy to hear it. You might also enjoy my previous project's launch story: Gift Lizard Launch: Stats, Failures, Successes and Lessons Learned.

Creativity and Kids

One thing I constantly find amazing is the solutions kids come up with for technology problems they encounter. The tools they pick can often seem baffling to those of us who are steeped in tech knowledge. They don't seem to limit their thinking to picking the exact tool designed for the problem at hand. They often pick tools which seem terribly suited for the problem but make it work.

What brought about this blog post? A quiz game created in powerpoint with a full score tracking system. How? Hardcoded with links on various pages (think a giant list of goto statements or a create your own story book).

Some people might think it's just ignorant behavior, others might call it innovation. I think those of us very comfortable and familiar with technology, programmers especially, easily dismiss options which see ill suited or don't even recognize them as options at all.

How does this tie into entrepreneurship? I often realize I have these blinders on when watching kids (and sometimes adults) get things done using various programs/apps. Do I really understand the problem people are trying to solve and the way they think about solving the problem? Most of the time, no, I haven't a clue is what I've realized.

Thought Experiment: Google exposed a vulnerability in Bing's search ranking feature, could it be exploited?

Everyone in the tech world has heard about the latest spat between Google and Bing (Microsoft).
Matt Cutts has talked a lot about it.

The core issue is Bing is using clickstream data from either IE (Suggested Sites) and/or their
Bing Toolbar. They clearly are OK copying Google's results if users click them.

If I throw on my blackhat thinking cap, we can begin the thought experiment.

Big Questions:

  1. How does clickstream work: is it only looking at URLs or analyzing actual content?
  2. Does it only work for Google searches?

How does clickstream work: is it only looking at URLs or analyzing actual content?

This is an important distinction if we are looking to manipulate results for our benefit.

If Bing only looks at the stream of URLs, in theory, you might be able to inject any URL into there.
It may not have to show up on that Google search result at all. If that worked on long-tail, it
may also help with more competitive keywords, fooling Bing into thinking it was getting more clicks
on Google and sending a positive ranking signal.

If they are double checking content, this type of manipulation may not work when it tries to 'steal'
the information from the Google search result. Someone would have to test this to find out.

Does it only work for Google searches?

Clearly, they've decided Google is a valid source, but what else is? Could artificially creating
traffic patterns from important and/or relevant sites increase search rankings? This still depends
on how they gather data, whether it's only looking at the clickstream or actually analyzing content.


If I were doing some SEO for a super competitive niche, I might be exploring how could this type of
signal be adjusted for my own benefit. Google could manipulate it, why couldn't you?

The Blackmail/Extortion Business Model

The Blackmail/Extortion Business Model


A business that has the potential to negatively influence perceptions of
another company but negotiates the information disseminated for some type of fee.

Examples (Allegedly, with sources):

Businesses want positive reviews in one of the web's definitive guides to local
businesses. Yelp allegedly takes advantage of this (source )
by changing what reviews do or don't show based on whether you pay or not.

Advertisers use ComScore as the definitive source for traffic statistics.
ComScore is accused of under-counting companies that do not pay their fee ($10,000!)
to get their tracking code. (source)

Better Business Bureau
Businesses get rated and consumers make complaints. The BBB makes money
from businesses subscribed to their service. It has been accused that the grading
system favors paying businesses (source).

Why it Matters:

These organizations have a reputation with the audience purchasing a service or product.
They take payment from businesses providing that service or product in order to
present information about those businesses. Exerting their power over businesses
listed in their system can increase their profits. This business model sustains
itself as long as lawsuits and legislation allow. It is one of the strongest
lock-ins available. Consumers keep using it because companies keep promoting it
because those that don't, often fail.

The Hobbyist Business Model

The Hobbyist Business Model


A business that grows out of a hobby or passion that did not start out with a commercial intent.


“The comic began in September 2005 when Munroe decided to scan doodles from his school notebooks and put them on his webpage. Eventually the comic
was changed into a stand-alone website, where Munroe started selling T-shirts based on the comic. He currently "works on the comic full time," making
xkcd a self-sufficient webcomic.” ( xkcd is one of the most popular
web comics and is listed as one of the self-sufficient web comics.

HuskyStarcraft / HDStarcraft
These two starcraft casters (someone who records gameplay and commentates what's going on) have built a legion of fans across the globe.
HuskyStarcraft has 378,000 subscribers and HDStarcraft has 322,000. Their views are also some of the highest of any channel in the gaming category.
They are both behind IGN and EA but ahead of GameSpot and Ubisoft. They have hosted tournaments, get invited to events and presumably make a pretty
penny off YouTube partner program with their 100 million plus views each.

A hobbyist's MUD started in college that had 1 million users before becoming a business. Today, RuneScape one of the more popular MMORPGs with ~$60 million
in revenue during March 2008 - March 2009 and ~$28 million profit. (

A twitter account that has spawned a book deal and a tv show. An unemployed guy decided his dad's hilarious comments needed to be shared with the
world. The world loved it.

Why It Matters:

People are being given the tools to self-publish, self-broadcast, self-develop all sorts of content. Comics, games, music, videos, books, news and any
other medium where an individual is given the tools to create *anything* can make a huge impact. There are thousands of people out there sharing,
creating art, shipping their passion across all sorts of mediums. The web has given these artists the means to share their work with the world easily.
The web has also allowed people to learn almost anything they wish to learn about. They can download any piece of software for free (legally through
open source and illegally through pirating) giving themselves the tools to required to create their art.

Today, any passionate individual has the tools to be as powerful as any corporation with regards to connecting with an audience. Their passionate
projects and art compete for free because they don't see it as a job. They are simply creating or doing what they love. Most will never be recognized,
but a few will be superstars. The superstars are presented with all sorts of monetization options simply because of their success and popularity. Any
business that looks at a passionate hobby turned business and thinks their business model made the business successful is fooling themselves. This type
of success cannot be replicated in a business plan. It is a lot of talent and passion combined with good fortune.

Seven Sites in Seven Days - My Personal Challenge

I feel like I need to challenge myself. Since I moved last month, I have been doing a lot of little things here and there (see: consulting, freelancing). My own projects have stagnated and I haven't been nearly as productive as I would like.

I have a lot of ideas for different websites I own but have had a very hard time committing to one. I enjoy projects, building a first implementation and testing a concept.

In an effort to try and get creative and find a project or two to really work on, I decided why not build a lot of websites/projects and see what happens?

Seven Sites in Seven Days Challenge:
1. I will try and build 7 websites in 7 days.
2. There is no limitation/minimum requirement on what each will be.
3. Some will probably suck (likely most/all of them).
4. I may get help from freelancers (paid).
5. I may use existing resources - old code, domains, servers, etc.

Syndicate content