Dan Sweet

The Death of the Expert = Competitive Advantage

The “death of the expert” is one of the most fascinating themes I find myself going back to time and again in conversation with peers.  Let me tell you why I think this is a macro-trend that is worth taking a moment to understand.

It used to be that “expertise”, once developed, assured one of a secure place in the world.  Firms employed experts and thrived.  Those that failed to attract or retain the experts saw their competitive advantage slowly slip away.  That old equation no longer holds for the following reasons:

1.  Expertise is now more broadly attainable than ever before.
2.  Expertise is less easy for individual companies to “own”.
3.  The barriers to applying serious expertise to business challenges have never been lower.

Many companies remain unaware of these changes and, as a result, are leaving massive amounts of competitive advantage on the table.

Let’s break down the three points above in more detail.

1.  Expertise is now more broadly attainable than ever before.

If you don’t know what a MOOC is, you need to click through to Wikipedia and check it out.  Free, open, learning available to anyone with an internet connection.  The Machine Learning class taught by Stanford professor Andrew Ng is the highest profile example I have direct experience with.   If you like self paced video lectures you can consume at your own pace, check out the Khan Academy.  If you want to follow along with a traditional classroom style lecture on your own, check out MIT’s OCW and learn anything from Calculus to Linear Algebra to Python.  Take it to another level by learning to program a robotic car from Google’s Sebastian Thrun.

2. Expertise is less easy for individual companies to “own”.

As knowledge becomes more broadly distributed, it is less likely that any one company will be able to truly “own” a space by simply hiring all of the best people in a particular field.  When the top 3 winners in a open online predictive modeling competition with $100,000 in prize money are a Ecuadorian university student, a teaching assistant from Slovenia, and an actuary from Singapore, you know the world has changed.  187 teams competed over three months and have published much of their details and code here.  Oh, and did I forget to mention that all three of these winners learned machine learning from Andrew Ng’s Coursera course?  It is not possible to own the work product of all of the world’s smart people.  If your competitive advantage was built on being the company that does X really well, your days are numbered if you don’t recognize the significance of these changes and start taking advantage of the opportunities presented by this new reality.

3.   The barriers to applying serious expertise to business challenges has never been lower.  

This is the part where I plug one of my favorite startups – Kaggle.com.  Kaggle has built a platform that routinely eats experts for breakfast.  Straight chews them up and spits them out.  Top team of actuaries from the world’s second largest auto insurer – beaten in under 24 hours.  NASA – beaten by a random glaciologist from down under.  The stories go on and on.  Have a tough business challenge and want hundreds of experts from around the globe competing against each other to solve your problem?  Head over to Kaggle.com and setup a competition.  Need to keep your data private and own the work product?  Setup a private Kaggle competition and invite a few of the world’s highest ranked data scientists to crank on your problem.  Want to keep it completely in-house?  Setup an internal competition and get your R&D guys cranking on a marketing mix optimization problem.  Get your process engineers working on some consumer research.  Let your statisticians work over your social media ROI calculations.  What, they are in silos and don’t talk?  Some people see the future and are getting after it.  If you don’t see it and aren’t getting after it, you will be left behind.

 The expert is dead.  This is good news for small companies and bad news for big companies that aren’t able to adapt.  The upside for everyone is that a new source of competitive advantage exists and is still up for grabs.  The internet continues to further level the playing field.



How I shuttered a lean startup for $0.00

Three months ago I wrote a post titled:  How I launched a lean startup for $8.17.  Today I am officially shuttering Tradesthatmatter.com.  I made this decision within a couple weeks of the launch, but am just making it official now.  I’ll send an email to the three people (yes three) that have signed up for the service letting them know the product won’t be launching and pointing them here for the gory details.

Lessons learned:

  • Don’t work on things you aren’t passionate about
  • Don’t test a hypothesis if the results won’t change your approach
  • Don’t overestimate your ability to manually “fake” the output of a technical tool before building it
  • Focus on a big problem
  • LinkedIn is your friend
  • Compete.com is your other friend


SEC.gov has lots of interesting data that people are required to report.  If combed through and looked at in context, the information can be valuable.  It is highly structured data, so a technical solution is not too difficult.  My value proposition was “protect your stock portfolio and spot the best opportunities with meaningful real-time alerts”.  The initial hypothesis I wanted to test was that “people would find the context of insider activities useful and engaging.”  I wrote one post as a test of this hypothesis, shared via Twitter, got 9 clicks on the bit.ly link and 15 visits to the post.  I didn’t have much fun researching or writing it, not many people seemed to care, and I’ve always known this was a small market.  Now, on to the lessons learned.

Lesson 1:  Don’t work on things you aren’t passionate about

My primary motivation was to actually try out one of my ideas.  Map out the business model, potentially build my technical skills on a small project, maybe make some extra cash, “found a startup”, etc.  However, I don’t really care about the opportunity that I identified.  Delivering a solution doesn’t resolve a pain point that matters to me.  A great comment I recently heard from Zuck at 27:01 in this video was this:  “the companies that work are companies that people really cared about and had some vision for what they wanted to see exist in the world, not just because they wanted to start a company.”

Lesson 2:  Don’t test a hypothesis if the test results won’t change your approach

The initial hypothesis I wanted to test was that “People would find the context of insider activities useful and engaging.”  I didn’t expect my one post and couple tweets to “go viral”.  Rich people buying stocks just isn’t that interesting, even with a little extra context thrown in.  I knew from my research of the market size, competitors, traffic statistics, etc that this was not a big market.  So why test trying to get people excited and engaged about a financial utility?  If wildly successful I proceed, if no interest at all I rationalize it with “well it’s a niche product and I just have an awareness problem.”  Useless test.

Lesson 3:  Don’t overestimate your ability to manually “fake” the output of a technical tool before building it

Browsing some competitive sites, running some screens, doing a little research, etc, all takes time.  That is why building a tool to automate all this would have value.  Doing this manual process many times over to build awareness of the product would be very time consuming.  Additionally, sometimes there is nothing interesting going on in the world of publicly-disclosed insider transactions.  I thought this content would be much easier to fake than it was.

Lesson 4:  Focus on a big problem

Real-time alerts for publicly-disclosed insider trading activity is not a big problem.  Competitive sites with a basic product get between 2 and 10 thousand monthly uniques.  Run a freemium model, convert 2% at $49 a month, and you are in the $20-$100 thousand annual income range.  With taxes, some administrative overhead, etc, this quickly becomes not worth much.  I can achieve much greater upside and leverage on my incremental efforts in my day job.  This problem is too small for me to work on.  I knew this from spending a couple hours spread across Google, Compete.com, and LinkedIn.  However, my desire to “start something” led me to ignore these hard facts.

Lesson 5:  LinkedIn is your friend

So you Google up a few competitors.  What next?  Let the people tell the story.  Use LinkedIn’s advanced search and search for the competitive companies in the “Employer” field.  See who works there now, who used to work there, who the founders are, what kind of other places they have worked, what other things they have started, their education, their technical chops, etc.  The facts of a few LinkedIn profiles laid out alongside each other often tell a fairly complete story.  Questions like what kind of revenue is attainable, how hard is the technical challenge, how easy is it to recruit people to work on this idea, are all easily answerable.  In this case, the most successful competitors seem to be run by 1-2 people, coming from different industries, running what look to be lifestyle businesses.  One competitor had a team of 5-6 in place once (including a former Googler) but everyone but the founder has since moved on.  Another competitor is a retired engineer from the tel-com industry, another is a couple of young relatively non-technical guys who have done a couple other projects in parallel, another much more technical product comes from a Y-combinator originating team of 2-3 guys with no obvious public signs of going anywhere in the last couple years.

Lesson 6:  Compete.com is your other friend

What kind of traffic are people doing?  Are they growing?  On what kind of trajectory?  What type of keywords are they owning?  What are their sources of traffic?  This can be good for finding other competitors you didn’t know existed as well.  Look at the range of pricing models in the market, look at the traffic they do, look at the industry conversion rates if using a freemium model and turn that into some revenue projections.  Do those projections support your cost structure?  What will you do differently than those competitors?  Will you convert better, charge more, get more users?  What makes you think so?   Sounds like a good area for some user testing.

Let’s see if I can learn from these “lessons learned” the next time around.   Please leave a comment below with any key startup lessons you’ve learned.

Recent talk on data visualization, psychology, design principles

Simple Portfolio 2011: Update with 6 weeks to go

Since May 3, 2011:
DOW up 3.4%
S&P500 up 4.4%
Nasdaq up 10.0%

This portfolio has returned 17.0% since May 3, 2011. Given the Wall Street approach of measuring relative returns on a one year basis, I feel good about closing in on a 7-13% market beat.

Here is the current portfolio, purchased in May 2011, originally posted here:

AAPL 14% (Huge margins, huge revenue growth, tons of cash.)
GOOG 22% (“Less than free” looks likely to crush most competitors.)
GTU 23% (US policymakers will continue printing money = gold higher.)
FCX 20% (The developing world will continue developing.)
IBB 22% (All those smart scientists will eventually do something cool.)

Not sure what to think about the 29% hit my FCX position has taken. I still believe in the developing world continuing to develop but haven’t enjoyed the loss.

I bailed on the gold trade (GTU) at 1,815 Back in August of 2011 as it seemed a bit frothy. I avoided quite a bit of pain but that money has been sitting in cash since then. Good thing it is just my money so I’m allowed to do that.

Key lessons learned so far: Don’t overthink it. Once you make a good plan, let it work for you.

Minimum incentive to open a new credit account?

I wrote this recently in reply to a question posted on Quora.com.  If you haven’t checked out Quora yet, go check it out.  Quora is a great example of a business model based on Clay Shirky’s concept of cognitive surplus.  I like the execution because they make it easy to share your knowledge and also make it social enough that it motivates people to contribute more.

Q: If a merchant offers you a discount when you apply for their credit card, how much should it be worth to go for it?

Short answer: minimum $100 offer value
Long answer:
3 filters to use
1 – What is your current credit score?
2 – Are you getting a major loan (house, car, boat, etc) anytime soon?
3 – Are you making a larger purchase from this vendor in the future?

If the answer to 1 is “I don’t know” then just say no to all of these offers in general.  Go to a site like creditkarma.com and check your credit for free without a “hard” inquiry.  “Hard” and “soft” exist.  Google it if you don’t know the difference.  If your score is 700+ then proceed to question 2.

If the answer to 2 is “yes”, then you should also just walk away.  $100 or whatever isn’t worth a few extra hundred in interest on that new loan because you dropped your score just under 720 with some dumb “offer” you “took advantage of”.  If the answer to 2 is “no” then proceed to question 2.

Many of these offers are one-time offers only good at that store on the initial purchase when the card is opened.  This is where question 3 comes in.  For example, 5 years back I bought a furniture set at Macy’s signed up for a card and got 20% off.  $2-300 in my pocket.  Yesterday I asked about a coupon that was valid only for Macy’s cardholders.  My account had gone inactive but, no problem sir we’ll open it right back up for you and you’ll save $14.  No thanks, I am thinking about some new furniture down the road.  Probability I buy furniture from Macy’s again * expected value of future deal > $14 so this is a pass for me.  You do the calculations for your scenario.

If its not a one-time deal with a vendor, you have decent credit, and no plans for major new loans in the works, then go to town.  It will impact your credit if you do a number of these.  For that reason, I don’t do these for less than $100 as plenty of them come along.